appln_id
int64 7.41k
576M
| appln_filing_date
stringlengths 10
10
| docdb_family_id
int64 3.49M
82.2M
| granted
stringclasses 2
values | appln_abstract
stringlengths 120
10k
| appln_abstract_lg
stringclasses 1
value | appln_title
stringlengths 4
567
| applt_coun
stringlengths 4
289
| invt_coun
stringlengths 8
347
| cpc
stringlengths 11
1.62k
| ipc
sequence | __index_level_0__
int64 25
167k
|
|---|---|---|---|---|---|---|---|---|---|---|---|
481,201,605 | 2016-08-29 | 57,272,821 | Y | The utility model relates to a can mark static force 's pulse manifestation sensor and pulse manifestation collection system, this pulse manifestation sensor includes: the static force sensor for measure diagnosis by feeling the pulse in -process pulse manifestation sensor and wait to detect the position applied pressure to the human body, developments force transducer for developments power when measurement diagnosis by feeling the pulse in -process pulse is beated, wherein, dynamic force transducer includes the casing to and set up sensor body and the flexible conductive element on the casing, and flexible conductive element sets up with this physical contact of sensor, and response contact surface and human body that flexible conductive element can make dynamic force transducer wait to detect that fully laminate in the position and conduct pressure to the sensor body uniformly, in order to realize dynamic force measurement, developments force transducer's casing links to each other with static force transducer for with pressure conduction to static force sensor, with the measurement that realizes static force. The utility model discloses can all -round collection pulse manifestation information, guarantee simultaneously that high sensitivity developments power detects the measurement with accurate static force. | en | Can mark static force 's pulse manifestation sensor and pulse manifestation collection system | 58229617_ | 59393049_ | A61B 5/02,A61B 5/4854,A61B 5/6801,A61B 5/6825,A61B 5/683,A61B 5/6843,A61B 5/7203 | [
"A61B 5/02",
"A61B 5/00"
] | 110,131 |
17,062,004 | 1994-01-11 | 26,230,029 | N | Compound corresponding to the general formula (I) <IMAGE> in which X represents one or a number of atoms or groups chosen from hydrogen, halogens or alkyl, trifluoromethyl, alcoxy, alkylthio, methylsulphonyl, cyano, aminocarbonyl and carboxy groups, Y represents one or a number of atoms or groups chosen from hydrogen, halogens or alkyl, trifluoromethyl, methoxy and trifluoromethoxy groups, R1 represents hydrogen or an alkyl, phenylmethyl, 2-phenylethyl, acetyl or alkoxycarbonyl group, R2 and R3 each represent hydrogen or an alkyl (optionally substituted), prop-2-enyl, prop-2-ynyl, phenyl, 1-(phenylmethyl)piperid-4-yl or 1-[(cyclohexen-1-yl)methyl]piperid-4-yl group, or else R2 and R3 form, with the nitrogen atom which carries them, an optionally substituted heterocycle. They possess hypnotic, anxiolytic and anticonvulsant properties and, consequently, can be used for the treatment of ailments related to disorders of GABAergic transmission, such as anxiety, sleep disorders, epilepsy, spasticity, muscular contractions, cognitive disorders, disorders of weaning with respect to alcolism, and the like. It can be used in premedication and general anaesthetics for inducing and/or maintaining anaesthesia, or as local anaesthetics, optionally in combination with other anaesthetics and/or muscle-relaxants and/or analgesics. | en | 9H-Imidazo 1,2-a benzimidazole derivatives with GABA activity. | 324664_FR | 3271267_FR,3369302_FR,3369300_FR,3369301_FR,3369299_FR | A61P 25/08,A61P 25/20,A61P 43/00,C07D 487/04 | [
"A61P 25/20",
"C07D 233/68",
"A61P 43/00",
"A61P 25/08",
"A61K 31/415",
"C07D 235/00",
"C07D 487/04",
"C07D 519/00"
] | 18,555 |
46,951,084 | 1981-10-20 | 22,739,251 | N | Non-emissive, flame-retardant coating composition comprising a flame-retardant polymer; a solvent selected from water and/or a liquid, curable, non-flame promoting polyhalogenated solvent containing air-dryable functionality; and, optionally, a non-emissive, non-flame promoting surface-active agent selected from a cationic, non-emissive, non-flame promoting, curable polyhalogenated compound containing an air-dryable group and a cationic surface-active group, an anionic, non-emissive, non-flame promoting, curable polyhalogenated compound containing an air-dryable group and an anionic surface-active group; and a nonionic, non-emissive, non-flame promoting polyhalogenated compound containing a nonionic surfaceactive group. Also disclosed is a method for reducing the emissivity of a polyamine/polyamide hardener for an epoxy resin wherein volatile, mobile, low molecular weight amine components in the hardener are reacted preferentially with a trapping agent selected from an epoxy compound and an alpha-, beta-ethylenically unsaturated carbonyl compound. Further disclosed is a method for reducing the tendancy to crystallize and to retard dispersion in water of a highly halogenated polyepoxide resin by forming an adduct between at least a fraction of said halogenated polyepoxide resin and a long chain aliphatic compound. | en | NON-EMISSIVE,FLAME-RETARDANT COATING COMPOSITIONS | 15591276_US | 37469523_US,37469524_US,37469522_US | C08K 5/0008 | [
"C08K 5/00"
] | 32,957 |
484,207,164 | 2017-03-24 | 59,900,436 | N | The objective of the present invention is to identify objectively and quantitatively a time filter related to vision, non-invasively employing an oculomotor reflex. In the present invention, an initial image having a uniformly constant brightness, a first stimulation pattern image having a mean brightness the same as the initial image, and a second stimulation pattern image which induces apparent movement together with the first stimulation pattern image, are presented sequentially, while eye movement is measured during a fixed period while the second stimulation pattern image is being presented and is stored in association with the duration for which the first stimulating pattern image used during the measurement was presented. At this time, a setting of the duration for which the first stimulating pattern image used in each trial is presented is changed. Parameter values of an oculomotor reflex kinetic energy model are optimized in order to minimize the difference between a measured waveform specified in accordance with changes in a sight line direction, calculated for each presentation duration, and a simulation result calculated by inputting the first stimulating pattern image and the second stimulating pattern image into the kinetic energy model, and a time filter unique to an inspection target is calculated. | en | VISUAL FILTER IDENTIFICATION METHOD AND DEVICE | 67870365_JP | 69669091_JP,66927923_JP | A61B 3/0008,A61B 3/0091,A61B 3/024,A61B 3/028,A61B 3/113,A61B 10/00,G06V 10/145,G06V 40/193 | [
"A61B 10/00",
"A61B 3/113"
] | 112,049 |
2,456,244 | 1999-05-07 | 25,645,777 | Y | The multifocal pattern visual evoked potential (VEP) has been adapted to detect visual field loss as a form of objective perimetry, with specific application to glaucoma. Multifocal pattern stimuli are cortically scaled and presented using a standard system. In order to effectively record the visual field responses, a unique array of bipolar occipital straddle electrode positions are employed, overlying the occipital visual cortex. Multi-channel VEP recording is then performed using these electrode positions. The responses are combined from the channels to produce a representation of the visual up to 25 DEG of eccentricity. Analysis of results to detect early field defects is enhanced by asymmetry analysis between fellow eyes, and by a scaling algorithm to compensate for inter-individual variations. VEP objective perimetry corresponds well with subjective Humphrey visual field defects, showing loss of signal in the areas of visual field loss. Further, it shows reductions in areas of the visual field in glaucoma suspects where subjective testing is still within normal limits. This technique represents the first practical application of the multifocal pattern VEP to objective detection of visual field defects in glaucoma, and may have the potential for detecting disease at an earlier stage than conventional testing. | en | ELECTROPHYSIOLOGICAL VISUAL FIELD MEASUREMENT | 14643417_AU,14643416_AU,5556106_AU | 14643416_AU,14643417_AU | A61B 3/024,A61B 5/378 | [
"A61B 5/0484",
"A61B 3/024",
"A61B 3/10"
] | 2,794 |
352,600,354 | 2010-05-26 | 43,308,790 | Y | A sleepiness assessment value P1 is set to be an average heart rate HR (S130). A sleepiness assessment value P2 is set to be a standard deviation σ of a respiration interval RespI for one minute (S150). A sleepiness assessment value P3 is set to be a value that is obtained by dividing, by one minute, an integrated value of a square of an average heartbeat interval RRIavg in a case where a variation in which a heartbeat interval RRI increases occurs (in a case where an RRI count value Xi is a value of one) (S190). A sleepiness assessment value P4 is set to a value that is obtained by averaging a variation ratio in a case where a variation in which the respiration interval RespI increases occurs (in a case where an RRI count value Yi is a value of one) (S210). A sleepiness assessment value P is set to be a value obtained by summing values that are obtained by multiplying the individual sleepiness assessment values P1 to P4 by weight coefficients a to d, each of which is set for a corresponding one of the sleepiness assessment values P1 to P4 (S220). When the sleepiness assessment value P is smaller than a value of zero, it is determined that a person is highly alert (S240), and, when the sleepiness assessment value P is equal to or larger than the value of zero, it is determined that the person is less alert (S250). | en | Sleepiness assessment apparatus | 49974885_JP,11174504_JP,49961167_JP,50058075_JP,50011669_JP | 49961167_JP,50011669_JP,50058075_JP,49974885_JP | A61B 5/02405,A61B 5/0245,A61B 5/0816,A61B 5/18,A61B 5/6887,B60K 28/02,B60K 28/066,B60W2040/0827,B60W2540/22,B60Y2400/902 | [
"A61B 5/08",
"A61B 5/0205",
"A61B 5/00"
] | 72,873 |
494,668,468 | 2017-11-14 | 60,107,420 | N | [Problem] To provide a signal data discrimination device that is robust with respect to an unessential variation such as a light source variation and is capable of performing discrimination with high accuracy with a smaller amount of learning data than ever before. [Solution] A signal data discrimination device comprises: a signal data input unit for inputting signal data to be discriminated; a feature map generation unit which generates a feature map by extracting a feature quantity with respect to the signal data using a learning unit that has performed learning in advance on the basis of a plurality of pieces of sample data attached with a normal or abnormal teacher signal; a distance map generation unit which, using a plurality of pieces of data of the feature maps generated on the basis of the plurality of sample data attached with a normal teacher signal and the feature map of the signal data, generates a distance map by taking a feature map difference between combinations of the pieces of sample data and the signal data; a distance value computation unit which determines a distance value between the signal data and the sample data from the distance map; and a signal data discrimination unit which performs discrimination as to whether the signal data is normal or abnormal on the basis of the distance value. | en | LEARNED MODEL GENERATION METHOD, LEARNED MODEL GENERATION DEVICE, SIGNAL DATA DISCRIMINATION METHOD, SIGNAL DATA DISCRIMINATION DEVICE, AND SIGNAL DATA DISCRIMINATION PROGRAM | 64114798_JP | 67827608_JP | G06N 3/0454,G06N 3/08,G06T 7/00 | [
"G06N 99/00",
"G06T 7/00"
] | 117,856 |
529,065,676 | 2018-07-31 | 65,476,329 | N | The purpose of the present invention is to recognize from images a dynamic object to be recognized. Through use of a spatio-temporal image data generation unit (2), an image recognition device (1) performs a Hilbert scan of frame image data (6) constituting moving-image data (4) and generates unidimensional spatial image data (7), and furthermore arrays the unidimensional spatial image data (7) ina time direction and generates two-dimensional spatio-temporal image data (8) that holds spatial information and temporal information. In this manner, the image recognition device 1 converts the moving-image data (4) into the two-dimensional spatio-temporal image data (8) while holding the spatial information and temporal information. Next, through use of a CNN unit, the image recognition device(1) performs a convolution process in which a two-dimensional filter is used on the spatio-temporal image data (8) and recognizes from the image the behavior of a pedestrian who is the object to be recognized. In this manner, the image recognition device (1) can perform behavioral recognition of a pedestrian that requires a three-dimensional analysis consisting of two dimensions for space and onedimension for time, and estimate the state of the pedestrian by means of the image recognition process of a two-dimensional image by CNN. | en | IMAGE DATA GENERATION DEVICE, IMAGE RECOGNITION DEVICE, IMAGE DATA GENERATION PROGRAM, AND IMAGE RECOGNITION PROGRAM | 59250908_,61525712_ | 74078073_,75256891_,64215137_,72657756_ | G06K 9/6256,G06K 9/6273,G06V 10/10,G06V 10/82,G06V 20/56,G06V 20/58,G06V 40/103,G06V 40/20 | [
"G06N 3/04",
"G06K 9/00",
"G06K 9/62",
"G06N 3/08"
] | 138,966 |
529,801,360 | 2019-02-22 | 69,812,302 | N | The present invention relates to an attribute generation contention network and a clothing matching generation method based on the network, and relates to the field of generation model and clothing matching. Firstly, a matching clothing data set is established with respect to problems related to generation of mutually matching clothing images based on an existing clothing image. Matching clothing image information, text information, and corresponding attribute information are included therein. Secondly, an attribute generation contention network is designed, which includes a generator, a matching discriminator, and an attribute discriminator. Subsequently, contention training is conducted in relation thereto and a generator network parameter is learned and obtained. Lastly, mutually matching clothing images are generated by new clothing image input to the trained generator. The present invention provides a reference basis for a user′s costume coordination by generating a clothing image matching an image input by the attribute generation contention network designed through training. The present invention matches people′s normal matching rules on advanced semantic attributes and the application potential of the present invention is huge in fields such as clothing recommendation, clothing search, and fashion design. | en | AN ATTRIBUTE GENERATION CONTENTION NETWORK AND A CLOTHING MATCHING GENERATION METHOD BASED ON THE NETWORK | 76218546_CN | 75474412_,75332043_,76130002_ | G06K 9/6256,G06K 9/6268,G06N 3/08,G06Q 30/0631,G06Q 30/0643,G06T 11/00 | [
"G06Q 30/06",
"G06N 3/08"
] | 139,248 |
533,612,872 | 2020-03-18 | 71,182,474 | Y | The invention discloses a hesitant iterative calculation method and device for multi-modal machine learning target recognition, and the method comprises the steps: training a plurality of different machine learning models; calculating a mis-association tensor; calculating a model label accuracy matrix and a multi-modal conclusion matrix; calculating a first hesitating conclusion vector, a second hesitating conclusion vector and a third hesitating conclusion vector; recalculating the weights of the plurality of machine learning models, and calculating hesitance vectors; iteratively solving a second hesitation conclusion vector, a third hesitation conclusion vector and a hesitation degree vector until a termination condition is met; and obtaining an attention coefficient vector, and obtaining a label conclusion of the to-be-identified body according to the attention coefficient vector, the third hesitation conclusion vector and the hesitation degree vector. According to the method, multi-modal recognition result fusion is realized by using a hesitation iterative calculation method, the interpretability of a plurality of machine learning methods participating in the recognition process is enhanced, and the hesitation degree of the recognition result is obtained on the basis of the accuracy of the recognition result. | en | Hesitant iterative calculation method and device for multi-modal machine learning target recognition | 63257575_ | 65704066_,60188768_ | G06K 9/6201,G06K 9/6262,G06N 20/00 | [
"G06K 9/62",
"G06N 20/00"
] | 142,095 |
4,936,173 | 2006-07-12 | 35,033,649 | N | The present invention provides a neurostimulator system for alleviating cerebellar tremor, in particular, associated with multiple sclerosis, comprising a programmable electrical pulse generator; and at least one electrode connected with the programmable electrical pulse generator; and wherein the programmable electrical pulse generator is programmed to generate electrical signals with the following parameters: a current magnitude of not above 3 mA, in particular, not above 1 mA, a stimulation signal on-time to signal off-time ratio in the range of about 2:1 to about 1:2, and signal on-times and off-times in the range of about 10 seconds to about 5 minutes, a signal frequency below 30 Hz, and a pulse width within the interval of 50 .mu.s to 500 .mu.s. The invention also provides a method for manufacturing a neurostimulator system intended for alleviating cerebellar tremor associated with multiple sclerosis, comprising providing a programmable electrical pulse generator; programming the programmable electrical pulse generator to generate electrical signals and providing at least one electrode connected with the programmable electrical pulse generator for applying the electrical signals generated by the electrical pulse generator to a cranial nerve, in particular, the vagus nerve, for controlling cerebellar tremor. | en | VAGUS NERVE STIMULATION BY ELECTRICAL SIGNALS FOR CONTROLLING CEREBELLAR TREMOR | 10417896_US | 16889244_IT,16889243_IT | A61N 1/36053,A61N 1/36067,A61N 1/36135,A61N 1/36171 | [
"A61N 1/36"
] | 9,384 |
562,557,283 | 2021-08-04 | 78,836,880 | N | The invention relates to a medical entity recognition method based on multi-granularity text embedding, and belongs to the technical field of information extraction and knowledge graph construction. The medical entity recognition method comprises the steps: construction of multi-granularity text embedding: constructing the multi-granularity text embedding through a pre-training language model, the multi-granularity text embedding comprising character embedding, word embedding, part-of-speech embedding, substring embedding and phrase embedding; mode weight generation: generating mode weights of all characters in the Chinese sentence according to the medical term composition mode; node embedding representation learning: performing node embedding representation learning by using a graph attention network and a mode enhanced attention mechanism; output of a medical text entity recognition result: generating an entity category label of the medical text by adopting a conditional random field, and outputting the medical entity recognition result. According to the method, the problems of insufficient utilization of graph representation information and single embedding granularity of text distributed representation in medical entity recognition are solved, so that the performance of medical entity recognition is improved. | en | Medical entity recognition method based on multi-granularity text embedding | 58644572_ | 60896404_,84083121_,60518607_,80327085_,66996338_,66432955_,60658019_ | G06F 16/35,G06F 40/242,G06F 40/289,G06F 40/295,G06N 3/0454 | [
"G06F 40/242",
"G06F 16/35",
"G06N 3/04",
"G06F 40/295",
"G06F 40/289"
] | 161,141 |
39,574,846 | 2007-03-06 | 38,474,932 | N | The cry of an object infant is acquired in procedure 1, and the cry is converted into time-series data on a sound feature parameter by sound feature analysis (procedure 2). From the time-series data, an affect sound model held in advance in a database (X), and segment occurrence information held in advance in a database (Y), the affect giving the highest general similarity is determined, and it is judged that it is the affect of the infant (procedure 3). The affect sound model held in the database (X) is made by converting the cry of an infant having a predetermined affect into time-series data on a sound feature parameter by sound feature analysis, and by subjecting the sound feature parameter of each affect to appearance distribution learning processing (procedure X1 to X3). The segment occurrence information held in the database(Y) is created by dividing the cry of the infant having a predetermined affect into sound feature segments according to the sound feature, labeling them to convert them into sound feature segment-series data, and subjecting them to occurrence order learning processing for each affect (procedure Y1, Y2). The invention provides a method for accurately judging the affect of an infant on the basis of the voice uttered by the infant and a device and computer program for executing the method. | en | INFANT EMOTION JUDGING METHOD, AND DEVICE AND PROGRAM THEREFOR | 32384872_JP,11230039_JP,32561156_JP | 32384872_JP,32561156_JP | G10L 17/26 | [
"G10L 11/00",
"G10L 17/26",
"G10L 15/00",
"G10L 15/10",
"G10L 17/00"
] | 25,890 |
445,036,213 | 2014-12-17 | 49,955,845 | N | The present disclosure generally relates to a method and device for encoding an image, characterized in that it comprises: obtaining (10) a luminance component (L) and each colour component (C(i)) of the image, said components belonging to a perceptual space having a metric whose values are representative of the differences between the visual perceptions of two points of said perceptual space, the metric being defined such that a perceptual threshold ([Delta]E0) exists below which a human being is not able to perceive a visual difference between two colours of the perceptual space; obtaining a quantized luminance component of the image by quantizing (11) the luminance component (L) of the image; obtaining at least one transformed color component by transforming (12) each colour component (C(i)) of the image according to a transform (T) which depends on an upper bound of the metric ([Delta]E) which enables a control of the visual losses in a displayed decoded version of the image; obtaining at least one quantized colour component by quantizing (13) each transformed colour component (C(i)<SP>T</SP>) of the image in order that a metric which is calculated between the image and a decoded version of the image, stays below the upper bound ([Delta]E); and encoding (14) each quantized component of the image (LQ, C(i)TQ). | en | Method and device for encoding a high-dynamic range image | 5330922_FR | 50436539_FR,41079913_FR,12511766_FR | H04N 19/124,H04N 19/154,H04N 19/164,H04N 19/176,H04N 19/186,H04N 19/33,H04N 19/36 | [
"H04N 19/186",
"H04N 19/169",
"H04N 19/20"
] | 96,180 |
563,234,452 | 2020-02-26 | 72,338,670 | N | It is possible to more appropriately perform fluorescence separation. An information processing apparatus according to an embodiment includes: a fluorescence signal acquisition unit (112) that acquires a plurality of fluorescence spectra corresponding to each of a plurality of excitation lights having different wavelengths and irradiated to a fluorescence stained specimen (30), the fluorescence stained specimen (30) being created by staining a specimen (20) with a fluorescence reagent (10); a link unit (131) that generates a linked fluorescence spectrum by linking at least parts of the plurality of fluorescence spectra to each other in a wavelength direction; a separation unit (132) that separates the linked fluorescence spectrum into spectra for every fluorescent substance using a reference spectrum including a linked autofluorescence reference spectrum in which spectra of autofluorescent substances in the specimen are linked to each other in the wavelength direction and a linked fluorescence reference spectrum in which spectra of fluorescent substances in the fluorescence stained specimen are linked to each other in the wavelength direction; and an extraction unit (132) that updates the linked autofluorescence reference spectrum using the spectra for every fluorescent substance separated by the separation unit. | en | INFORMATION PROCESSING APPARATUS AND MICROSCOPE SYSTEM | 79653016_JP | 5580557_JP,11461350_JP,83587672_JP,10021252_JP,7922069_JP,5580560_JP | G01J 3/027,G01J 3/44,G01J2003/2873,G01N 21/6428,G01N 21/6458,G01N 21/6486,G01N2021/6417,G01N2021/6419,G01N2021/6423,G01N2021/6439,G01N2201/1293,G02B 21/16,G02B 21/365,G02B 21/367,G06V 10/143,G06V 10/60,G06V 20/695,G06V 20/698,G06V2201/03 | [
"G02B 21/16",
"G02B 21/36"
] | 161,529 |
524,271,688 | 2019-03-18 | 64,799,766 | N | An online learning facial recognition method in the technical field of computation and calculation, and in particular the technical field of computer vision for facial recognition. The method comprises: training a face feature extractor by using an external data set, extracting a reference feature in a local data set corresponding to each member to construct a reference feature space, comparing a feature vector of a sample under test and the reference features to determine a reference feature most similar to the feature vector of the sample under test; if the reference feature most similar to the feature vector of the sample under test satisfies a threshold requirement, using, as the identity of the sample under test, a member identity to which the reference feature most similar to the feature vector of the sample under test belongs, and if not, returning a message indicating that the identity recognition of the sample under test has failed; updating the reference feature space according to the difference between a predicted feature vector of the sample under test and a corresponding actual feature vector thereof in the reference feature space. The invention adapts to changes in facial features that occur over time, and is especially applicable to situations in which the composition of members frequently changes. | en | ONLINE LEARNING FACIAL RECOGNITION METHOD | 63673205_CN | 73348446_CN,69040593_CN,63739028_CN,69636763_CN,74286238_CN,73781406_CN | G06N 3/0454,G06N 3/084,G06V 20/41,G06V 40/168 | [
"G06K 9/00"
] | 135,501 |
511,990,918 | 2018-05-17 | 61,254,210 | N | Disclosed is a three-dimensional human face recognition method based on a multi-scale covariance descriptor and a local sensitive Riemann kernel sparse classification. The method comprises the following steps: respectively carrying out automatic preprocessing on original G library set human face models and P test set human face models; according to the library set human face models and the test set human face models after same have been subjected to the automatic preprocessing in step (1), establishing a scale space and detecting multi-scale key points and extracting neighborhoods thereof; extracting d × d-dimension local covariance descriptors from a neighborhood of each key point neighborhood under each scale, and carrying out multi-scale fusion on the local covariance descriptors so as to construct a multi-scale covariance descriptor; and mapping the local covariance descriptors to a renewable Hilbert space, and proposing a local sensitive Riemann kernel sparse representation to classify and recognize a three-dimensional human face. By means of the present invention, the expression capability of a single-scale local covariance descriptor can be effectively improved, and at the same time, the locality of a multi-scale descriptor can be effectively used in a local sensitive Riemann kernel sparse classification. | en | THREE-DIMENSIONAL HUMAN FACE RECOGNITION METHOD BASED ON MULTI-SCALE COVARIANCE DESCRIPTOR AND LOCAL SENSITIVE RIEMANN KERNEL SPARSE CLASSIFICATION | 63673205_CN | 63574817_CN,71771611_CN | G06K 9/624,G06V 20/653,G06V 40/168,G06V 40/172 | [
"G06K 9/00",
"G06K 9/62"
] | 127,406 |
500,033,082 | 2018-04-10 | 63,514,564 | N | The invention provides a self-organizing neural network topology preservation reinforcing method based on deep learning. The method is used for settling a technical problem of requirement for improving self-organizing neural network topology preservation effect in prior art. The method comprises the steps of setting a network structure and a parameter of the self-organizing neural network, and normalizing input layer data; setting the number of input-layer neurons and the number of competition-layer neurons; performing rough adjustment on the weight vector of the competition-layer neurons of the self-organizing neural network, and obtaining a rough-adjusted competition-layer weight vector; by means of the rough-adjusted result, performing fine adjustment on the weight vector of the competition-layer neurons of the self-organizing neural network, and obtaining a fine-adjusted competition-layer weight vector; measuring the competition-layer weight vector, thereby obtaining a topology preservation reinforcing effect. The self-organizing neural network topology preservation reinforcing method has advantages of reducing difference between each competition-layer weight vector of the self-organizing neural network and an input sample, and improving topology preservation capability of the self-organizing neural network. | en | Self-organizing neural network topology preservation reinforcing method based on deep learning | 66998455_ | 64830519_,61055194_,63627691_,65397854_,60956839_ | G06N 3/04,G06N 3/08 | [
"G06N 3/08",
"G06N 3/04"
] | 120,825 |
529,879,231 | 2018-11-22 | 64,764,349 | N | A dual-frequency comb spectral focusing coherent anti-Stokes Raman spectroscopy detection system, comprising a dual-frequency comb light source module (1), a Stokes light generating module (5), a microfocus and sample scanning module (9), a sampling triggering module (12), a coherent anti-Stokes signal detection module (11), a collection module (13) and an analysis processing module (14). The dual-frequency comb light source module (1) comprises a first frequency comb light source (1-1) and a second frequency comb light source (1-2), having a slight difference in repeat frequency, light output by the first frequency comb light source (1-1) acting as a pump light and a detection light in the detection process; the Stokes light generating module (5) uses a nonlinear effect of optical fiber to realize a frequency shift, obtaining Stokes light; the sampling triggering module (12) uses a secondary harmonic generation technique to obtain overlap positions of two frequency comb pulses, causing pulse positioning to trigger spectrum collection. The present detection system realizes non-mechanical scanning rapid broadband coherent anti-Stokes Raman spectroscopy detection, significantly increasing the capability of broadband coherent anti-Stokes spectroscopy technology in highly dynamic, multi-component detection analysis. | en | DUAL-FREQUENCY COMB SPECTRAL FOCUSING COHERENT ANTI-STOKES RAMAN SPECTROSCOPY DETECTION SYSTEM | 68332319_CN | 76996513_CN,66827799_CN,63914786_CN,76859316_CN | G01N 21/01,G01N 21/65,G01N2021/653 | [
"G01N 21/01",
"G01N 21/65"
] | 139,351 |
510,989,380 | 2018-11-09 | 66,004,052 | N | The invention discloses a space robot mechanical arm arresting control system. The space robot mechanical arm arresting control system comprises two loops, namely, an inner loop and an outer loop, inthe outer loop, the system achieves the attitude stability of a space robot mechanical arm base platform in the arresting process through a PD controller, and in the inner loop, the system controls amechanical arm to achieve arresting maneuvering on a non-cooperative target through a reinforce learning control system based on reinforce learning. The invention further discloses a reinforce learning method for controlling the reinforce learning control system of the mechanical arm in the inner loop of the system and a space robot dynamics modeling method of the space robot mechanical arm arresting control system. According to the space robot arresting control system, the reinforce learning method and the dynamics modeling method, compared with PD control, the posture disturbance of the baseplatform under reinforce learning RL control is smaller, the movement process of the tail end of the mechanical arm is more stable, the control precision is higher, moreover, the motion flexibility of the mechanical arm under the reinforce learning RL control is good, and the autonomous intelligence is achieved to the greater extent. | en | Space robot arresting control system, reinforce learning method and dynamics modeling method | 59074494_ | 65011553_,70196434_,60678294_,61570847_,67089055_ | B25J 9/1605,B25J 9/163 | [
"B25J 9/16"
] | 126,792 |
558,328,821 | 2020-12-10 | 74,169,662 | N | The present application relates to the technical field of artificial intelligence. A brain midline recognition method and apparatus, and a computer device and a storage medium. The method comprises: performing image preprocessing on a brain image associated with a user identification code so as to obtain an image to be subjected to recognition; by means of a multi-scale deep network model, performing midline feature extraction to generate a feature image to be processed and a classification recognition result; by means of a feature pyramid network model, performing feature fusion on all feature images to be processed so as to generate a fused feature image group; by using a bilinear interpolation method and by means of a weighted fusion model, performing interpolation and weighted fusion on all fused feature image groups so as to generate a feature image to be segmented, and performing midline segmentation on the feature image to be segmented, so as to obtain a brain midline segmentation recognition result; and performing synthesis to obtain a brain midline image, and outputting a final recognition result. By means of the method, a brain midline is automatically recognized and marked. The method is applicable to the field of smart medical treatment, etc., and can further promote the construction of smart cities. | en | BRAIN MIDLINE RECOGNITION METHOD AND APPARATUS, AND COMPUTER DEVICE AND STORAGE MEDIUM | 63942312_CN | 67743655_CN,76955409_CN,80678091_CN,82746501_CN | G06K 9/6267,G06N 3/0454,G06N 3/08,G06T 7/0012,G06T 7/10,G06T2207/20016,G06T2207/30016,G06V 10/464 | [
"G06T 7/00"
] | 158,362 |
53,316,059 | 1991-09-24 | 17,228,033 | Y | An apparatus for recognizing driving environments of a vehicle including a plurality of sensors for detecting various parameters relating to driving conditions of the vehicle such as throttle valve open angle, vehicle running speed, brake pedal depression amount and gear shift range of an automatic transmission, first and second neuron interfaces for converting parameter values detected by the sensors into a plurality of input patterns having predetermined configuration, first and second neural networks having input layers to which corresponding input patterns are applied, hidden layers and output layers for producing recognition results, and a multiplexer for selecting one of the recognition results produced on the output layers of the first and second neural networks. The first neural network has a superior separating or recognizing and learning faculty, while the second neural network has a superior associating faculty. A accelerating pedal depression amount is detected by a sensor and a variation of the thus detected amount is compared with a reference value. When the variation is larger than the reference value, the recognition result produced by the first neural network is selected and when the variation is smaller than the reference value, the recognition result from the second neural network is selected. | en | Apparatus for recognizing driving environment of vehicle | 5242667_JP | 6926641_JP | B60W 40/09,B60W2520/10,B60W2530/00,B60W2540/10,B60W2540/12,B60W2540/16,B60W2540/30,F16H 59/18,F16H 59/24,F16H 59/44,F16H 59/54,F16H 59/66,F16H 61/0213,F16H2059/003,F16H2061/0084,G06K 9/00,G06K 9/6293,G06N 3/0454,Y10S 706/905 | [
"F16H 59/18",
"F16H 59/66",
"B60R 16/02",
"F02D 45/00",
"G01C 21/00",
"B60W 10/10",
"F16H 61/10",
"F16H 59/54",
"F16H 61/04",
"B60W 10/00",
"F16H 59/24",
"G06K 9/68",
"G08G 1/16",
"F16H 61/02",
"G06N 3/04",
"G06K 9/62",
"F16H 61/00",
"G06K 9/00",
"F16H 59/44"
] | 45,774 |
533,329,637 | 2018-07-24 | 65,040,062 | N | Provided is an environmental equipment control apparatus capable of immediately executing control for improving a physical and mental state of a user. An environmental equipment control apparatus (100) for controlling a plurality of types of environmental equipment (10, 20, 30) includes: a grasping unit (60) that grasps current physical and mental state information of the user, environmental situation information, and target relationship information representing a relationship between a target physical and mental state and a current physical and mental state; learning control plan output means (80) that outputs a control change plan for each of combinations of the plurality of types of environmental equipment (10, 20, 30) in accordance with the current physical and mental state information, the environmental situation information, and the target relationship information; and a selection control unit (90) that selects one control plan from among a plurality of control plans output by the learning control plan output means (80) and executes the one control change plan. The learning control plan output means (80) learns to update a method for specifying the control plan to be output, by using the physical and mental state of the user changed by executing the control plan selected by the selection control unit (90). | en | ENVIRONMENTAL EQUIPMENT CONTROL APPARATUS | 6086649_JP | 74724730_JP,74771913_JP,74669801_JP,74767908_JP,74731895_JP | F24F 11/0008,F24F 11/63,F24F 11/65,F24F 11/70,F24F 11/80,F24F2110/10,F24F2110/60,F24F2110/70,F24F2120/00,G05B 13/027,G05B 15/02,G05B2219/2642,G06F 3/015,G06F 3/017,G06F2203/011,G06N 3/08,G06N 7/005,G06N 20/10,H05B 47/105,Y02B 30/70 | [
"F24F 11/70"
] | 141,892 |
267,601,714 | 2008-12-15 | 40,901,225 | Y | <P>PROBLEM TO BE SOLVED: To provide a system capable of judging right and left confirmation operation and preparation for a brake without imposing a burden on a driver so much and quantitatively measuring dangerous drive preventive intention of the driver. <P>SOLUTION: The dangerous drive preventive intention judgment system 10 includes a computer 12, which acquires angular velocity data from an angular velocity sensor 14 set on the head of a driver and an angular velocity sensor 16 set at the tip of the right foot, for example, acquires present position data from a position detection device 18, and records the acquired data at a drive data recording device 24. When an automobile is close to a dangerous place such as an intersection without traffic lights it is judged in accordance with the angular velocity data whether a visual confirmation operation or a preparation for a brake is carried out for dangerous drive prevention. A judgment result of the visual confirmation operation or the preparation for the brake in respect of the dangerous place, for example, is recorded in the drive data recording device 24. Further, when such judgment is made that the visual confirmation operation or the preparation for the brake is not carried out, an alarm is sounded from an alarm device 22. <P>COPYRIGHT: (C)2009,JPO&INPIT | en | DANGEROUS DRIVE PREVENTIVE INTENTION JUDGMENT SYSTEM AND DANGEROUS DRIVE PREVENTIVE INTENTION JUDGMENT METHOD | 32053237_ | 20034434_,30497834_,30631385_,29742813_,41064012_,30859718_ | A61B 3/113,A61B 5/11,A61B 5/18,A61B 5/6803,A61B 5/6829,A61B 5/6893,G08B 21/06,G08G 1/165 | [
"B60R 21/00",
"G08G 1/16",
"G08G 1/00",
"B60R 1/00"
] | 58,225 |
334,027,020 | 2006-09-12 | 37,865,536 | N | The present invention describes methods for treating heart diseases by pharmacological manipulation of intrinsic cardiac adrenergic cells. The activation or inhibition of δ-opioid receptors or selective β2-adrenergic receptor agonist in these cells can help regulate catecholamine synthesis and release in these cells. Thus drugs directed towards these receptors can be used in the treatment of heart diseases including acute and chronic congestive heart failure, as well as acute and chronic coronary artery disease. The present invention also describes a method of treating acute decompensated congestive heart failure using δ-opioid receptor agonists or selective β2-adrenergic receptor agonists to stimulate endogenous production of brain natriuretic peptide. This invention also provides a novel strategy for protection against myocardial ischemia through the specific activation of δ-opioid receptors or selective β2-adrenergic receptor agonists exclusively expressed by intrinsic cardiac adrenergic cells in human heart. The activation of intrinsic cardiac adrenergic cells by δ-opioid receptor stimulation enhances endogenous catecholamine release, which subsequently stimulates myocardial alpha 1-adrenoreceptors achieving myocardial protection in patients suffering from myocardial ischemia due to coronary artery disease. | en | Regulation of Brain Natriuretic Peptide and Catecholamines for the Treatment of Cardiovascular Diseases | 11747573_US,11747572_US | 11747572_US,11747573_US | A61K 38/33,A61P 9/00,A61P 9/04,A61P 25/00 | [
"A61K 38/00",
"A61P 9/00",
"A61K 31/4422",
"A61P 9/04",
"A61P 25/00"
] | 67,637 |
341,437,150 | 2011-11-22 | 45,688,944 | N | It is disclosed a process and a computer-based program for distinguishing ex vivo between benign and malign tumors in tissues such as soft tissues and particularly breast tissue through the registration and comparison of measurement data from dynamic contrast-enhanced magnetic resonance imagining (DCE-MRI) and dynamic susceptibility contrast-enhanced magnetic resonance (DSC) imagining of the tumors. The process and the computer-based program is adjusted to initially performing two dynamic MRI pulse sequences in an interleaved mode during parenteral administration of an MR contrast substance, wherein one of said pulse sequences is optimized for high spatial resolution and the other pulse sequence is adjusted for high temporal resolution, the high-temporal dissolved sequence further comprising a double echo-collection being sensitive towards both DCE and DSC for generating a number of different biomarker data such as pharmacokinetic biomarker data, descriptive DCE biomarkers and descriptive DSC biomarkers, and subsequently normalizing and comparing said data with corresponding data from corresponding benign and malign tumors, respectively. From each tumor volume there is preferably collected biomarkers for the 95-percentile identifying most of the abnormal kinetic properties (belonging to the relevant biomarker). | en | PROCESS FOR SEPARATING MALIGN AND BENIGN TUMORS EX VIVO THROUGH THE USE OF DCE-MRI AND/OR DSC-MRI | 43449929_NO,43449928_NO,43449930_NO | 43449929_NO,43449930_NO | A61B 5/055,A61B 5/4312,A61B 5/742,G01R 33/281,G01R 33/4828,G01R 33/5601,G01R 33/5602,G01R 33/5608,G01R 33/56366 | [
"G01R 33/563"
] | 72,208 |
558,612,750 | 2020-06-23 | 70,950,048 | N | A metering-level 3D super-depth-of-field microscopic system, which relates to the technical field of photoelectric nondestructive 3D measurement. The microscopic system comprises an optical microscopic unit, a photoelectric contour sensor unit, an x-y-z-axis electric displacement platform, a control module and a PC processor, wherein the x-y-z-axis electric displacement platform comprises an x-y-axis translation platform and a z-axis lifting module; an optical microscope is configured with an illumination light source and an image collection unit; the control module comprises an x-y-z electric control unit, a signal processor connected to the photoelectric contour sensor, and a power source module; and the image collection unit is used for collecting image information of a target measurement area and transmitting the image information to the PC processor. By implementing a corresponding measurement method by means of the metering-level 3D super-depth-of-field microscopic system, a sample can be observed in a nondestructive manner, and photographing, 3D true color imaging and 3D accurate measurement can be performed on the microscopic state of a surface material over an ultra-large area, thereby achieving metering-level measurement precision, and preserving the details and the real color of a microscopic sample. | en | METERING-LEVEL 3D SUPER-DEPTH-OF-FIELD MICROSCOPIC SYSTEM AND MEASUREMENT METHOD | 63944165_CN | 63944165_CN | G01B 11/24,G01B 11/2441,G01N 21/95 | [
"G01N 21/95",
"G01B 11/24"
] | 158,630 |
533,573,774 | 2020-04-01 | 71,164,173 | N | The present invention discloses a brain wave monitoring device. The brain wave monitoring device comprises a brain wave acquisition unit, a processing unit, a MCU main control unit and a Bluetooth module; the brain wave acquisition unit is used for acquiring bioelectrical signals; the processing unit is used for converting the bioelectrical signals into digital signals, and the MCU main control unit is used for receiving the digital signals, processing the digital signals and then outputting data packets to the Bluetooth module; and the Bluetooth module is provided with a Mesh networking functional component, so that at least two brain wave monitoring devices can be connected through the Bluetooth module to form Mesh networking. The brain wave monitoring device can monitor and record brainwave data of testees in real time and form the Mesh networking with an external device through the Bluetooth module, when a plurality of the testees need to be monitored simultaneously, the pluralityof the brain wave monitoring device transmit the acquired data to the external device for integration and then transmit the data to an intelligent terminal device, so that the brain wave data of allthe testees can be acquired, the terminal intelligent device does not need to be arranged for each testee, and time and cost are saved. | en | Brain wave monitoring device, system and monitoring method | 62991988_ | 72790183_,59568523_,78770015_,78163916_ | A61B 5/0006,A61B 5/0022,A61B 5/369,A61B 5/7203,A61B 5/7225 | [
"A61B 5/0476"
] | 142,075 |
53,474,087 | 1991-11-12 | 27,066,738 | Y | An object, such as a robot, is located at an initial state in a finite state space area and moves under the control of the unsupervised neural network model of the invention. The network instructs the object to move in one of several directions from the initial state. Upon reaching another state, the model again instructs the object to move in one of several directions. These instructions continue until either: a) the object has completed a cycle by ending up back at a state it has been to previously during this cycle, or b) the object has completed a cycle by reaching the goal state. If the object ends up back at a state it has been to previously during this cycle, the neural network model ends the cycle and immediately begins a new cycle from the present location. When the object reaches the goal state, the neural network model learns that this path is productive towards reaching the goal state, and is given delayed reinforcement in the form of a 'reward'. Upon reaching a state, the neural network model calculates a level of satisfaction with its progress towards reaching the goal state. If the level of satisfaction is low, the neural network model is more likely to override what has been learned thus far and deviate from a path known to lead to the goal state to experiment with new and possibly better paths. | en | Neural network model for reaching a goal state | 7500989_US | 5535140_US | G06N 3/04,G06N 3/08 | [
"G06N 3/04",
"G06N 3/08"
] | 46,163 |
530,636,617 | 2018-12-25 | 65,190,945 | N | A system for measuring a stimulation parameter of an animal robot based on machine vision, relating to the control field of animal robots. According to the system, a machine vision technology is used to obtain the movement trajectory and the rotation direction of the animal robot to analyze the movement behavior of the animal robot more precisely and quantitatively, and further quantitatively analyze the correspondence between the stipulation parameter and the controlled behavior of the animal robot so as to determine the optimal stipulation parameter suitable for the specific animal individual and provide necessary information for the practical research on the animal robot. The manual determination is replaced with the machine vision. On the basis of the method proposed in the present invention, not only the standardization of the experimental process and the objectivity of experimental data are ensured so as to obtain the objective and consistent measurement result of the stimulation parameter, but also the low-value labor caused by the original manual stimulation parameter measurement experiment is significantly reduced, and the research efficiency of the animal robot is improved. The present invention also relates to a method for measuring a stimulation parameter of an animal robot based on machine vision. | en | SYSTEM AND METHOD FOR MEASURING STIMULATION PARAMETER OF ANIMAL ROBOT BASED ON MACHINE VISION | 67711325_CN | 76309959_CN,63955905_CN,77581130_CN,76123707_CN,64037329_CN,77384372_CN,76113135_CN,76997423_CN,63499331_CN | B25J 19/0095,G06T 3/4038,G06T 7/20,G06T2207/10016,G06T2207/30241 | [
"B25J 19/00",
"A01K 15/02"
] | 139,879 |
503,624,391 | 2018-11-29 | 60,574,479 | N | The present invention relates to a system for planning and/or providing neuromodulation (10), especially neurostimulation, comprising: - first data input module (26) for stimulation related basic data, - a stimulation related basic data storage module (28) for storing the stimulation related basic data, - second data input module (30) for stimulation related response data, - a stimulation related response data storage module (32) for storing the stimulation related response data, - transfer module (34) configured and arranged such that the stimulation related basic data received by the data input module are linked and/or translated into and/or with the response data and/or artificial response data created by the transfer module, wherein the data generated by the transfer module (34) are transfer data, the transfer data comprising link data and/or translation data and/or artificial response data, - a transfer response data storage module for storing the transfer data, - mapping module (38) configured and arranged such that based on the stimulation related basic data and stimulation related response data and the transfer data a digital characteristic map (36) is generated, which describes the interrelation between the stimulation related basic data and the stimulation related response data and the transfer data. | en | A SYSTEM FOR PLANNING AND/OR PROVIDING NEUROMODULATION | 71915434_NL,7982008_CH | 72021317_NL,42586663_CH,63563807_CH,71945412_NL,50193252_CH,53213143_NL,53253109_NL,57511843_AT,71872992_NL | A61B 5/1116,A61B 5/1118,A61B 5/372,A61B 5/397,A61N 1/025,A61N 1/0551,A61N 1/36062,A61N 1/36067,A61N 1/36135,A61N 1/36139,A61N 1/3614,A61N 1/36146,G16H 10/60,G16H 20/30,G16H 20/40,G16H 40/40,G16H 40/63,G16H 40/67 | [
"A61N 1/05",
"A61B 5/00",
"G16H 40/63",
"A61N 1/36"
] | 122,905 |
353,625,955 | 2011-04-05 | 46,020,580 | N | Working memory (WM) is part of the brain's memory system that provides temporary storage and manipulation of information necessary for cognition. Although WM has limited capacity at any given time, it has vast memory content in the sense that it acts on the brain's nearly infinite repertoire of lifetime memories. As described, large memory content and WM functionality emerge spontaneously if the spike-timing nature of neuronal processing is taken into account. The memories are represented by extensively overlapping groups of neurons that exhibit stereotypical time-locked spatiotemporal spike-timing patterns, called polychronous patterns. Using computer-implemented simulations, associative synaptic plasticity in the form of short-term STDP selects such polychronous neuronal groups (PNGs) into WM by temporarily strengthening the synapses of the selected PNGs. This strengthening increases the spontaneous reactivation frequency of the selected PNGs, resulting in irregular, yet systematically changing elevated firing activity patterns consistent with those recorded in vivo during WM tasks. The computer-implemented model implements the relationship between such slowly changing firing rates and precisely timed spikes, and also reveals a novel relationship between WM and the perception of time on the order of seconds. | en | Spike-timing computer modeling of working memory | 9184520_US,12082680_US,6165749_US | 12082680_US,9184520_US | G06N 3/049,G06N 3/088,G06N 3/10,G16B 5/00 | [
"G06N 3/04"
] | 73,512 |
445,794,584 | 2014-05-03 | 54,355,648 | Y | An accelerated learning and rehabilitation system for teaching the performance of a musical instrument, a remotely operated system, a sport, a weapon, and for brain rehabilitation and other uses includes generating sensory cues including auditory, haptic and visual sensory cues capable of being perceived by a user. The generated sensory cues are applied to the user and are dependent on a position of at least one body member of a performer relative to a performance element of a performance object with which an event is performed. The sensory cues are effective for stimulating a various processing center of a brain of the user so that user learns how to position his body member corresponding to the position of the performer of the event. The sensory cues can include visual sensory cues effective for stimulating the visual processing center of the brain of the user. The visual sensory cues are synchronized with the other applied sensory cues so that the position of the body member of the performer is virtually visually indicated in synchronization with the other sensory cues so that the visual processing center is stimulated with a visual sensory cue in synchronization with the stimulation of other processing centers corresponding to the other sensory cues for teaching the user to perform a version of the event. | en | Accelerated learning, entertainment and cognitive therapy using augmented reality comprising combined haptic, auditory, and visual stimulation | 11055781_US | 11055781_US | A61M 21/00,A61M 21/0094,A61M2021/0016,A61M2021/0022,A61M2021/0027,A61M2021/005,A61M2021/0072,A61M2021/0077,A61M2205/507,A61M2210/0612,A61M2210/0662,A61M2210/083,A61M2210/086,G09B 5/065,G09B 9/00,G09B 9/003,G09B 9/06,G09B 9/063,G09B 9/08,G09B 9/52,G09B 15/00,G09B 15/003,G09B 15/06,G09B 19/00,G09B 19/003,G09B 19/0038,G09B 19/22,G09B 21/00,G10H 1/18,G16H 20/30 | [
"G09B 9/06",
"G09B 9/08",
"G09B 9/00",
"A61M 21/00",
"G09B 9/52",
"G09B 15/00",
"G09B 19/00",
"G10H 1/18",
"G09B 19/22",
"G09B 5/06"
] | 96,889 |
534,178,107 | 2019-05-31 | 66,603,778 | N | Disclosed is a TextCNN-based knowledge extraction method, comprising: S10, constructing a word vector dictionary and a phrase vector dictionary; S20, constructing a first convolutional neural network, and training the first convolutional neural network on the basis of a first optimization algorithm, wherein the first convolutional neural network comprises a first embedded layer, a first multilayer convolution and a first softmax function successively connected to each other; S30, constructing a second convolutional neural network, and training the second convolutional neural network on the basis of a second optimization algorithm, wherein the second convolutional neural network comprises a second embedded layer, a second multilayer convolution, a pooling layer, two fully connected layers and a second softmax function successively connected to each other; and S40, extracting, according to an entity annotation prediction output by the trained first convolutional neural network and an entity relationship prediction output by the trained second convolutional neural network, a knowledge graph triple of data to be predicted. The convolutional neural network has a parallel computation characteristic, and therefore, computational resources can be fully utilized to realize improvement in the efficiency of computation. | en | TEXTCNN-BASED KNOWLEDGE EXTRACTION METHOD AND APPARATUS, AND COMPUTER DEVICE AND STORAGE MEDIUM | 63942312_CN | 63604368_CN,67315619_CN,63958634_CN | G06F 16/36,G06F 40/284,G06F 40/295,G06F 40/30,G06K 9/6201,G06K 9/6232,G06K 9/6256,G06K 9/6261,G06K 9/6288,G06K 9/6296,G06N 3/0454,G06N 3/08,G06N 3/084,G06N 5/02,G06N 5/022,G06N 5/025,G06N 20/10 | [
"G06N 3/04",
"G06F 16/36",
"G06N 3/08"
] | 142,252 |
496,784,132 | 2017-12-04 | 62,562,525 | N | A device may receive information associated with first and second bug reports to be classified as duplicate or non-duplicate bug reports. The device may identify first and second descriptions associated with the first and second bug reports, respectively. The first and second descriptions may be different descriptions having a shared description type. The device may identify a neural network for encoding the first and second descriptions, based on the shared description type. The device may encode the first description into a first vector using the neural network, and may encode the second description into a second vector using the neural network. The device may classify the first and second bug reports as duplicate or non-duplicate bug reports based on the first vector and the second vector. The device may perform an action based on classifying the first and second bug reports as duplicate or non-duplicate bug reports. 4/b _ ) L. U o a) >U _0 0) CLU)-) . 1) 0 L- . ) ( 0 C: U)f 0U 0 -0 CU±) 0) a)U) 0 - U (n ~ . a) 0 - L a cua 0 -0 - a .41 - -- U) C: Z -0 : -CD r c-0 - 0 D 4- a) -m U) 0- C:a)7 C It a))U 0U 0U 3: a) Z U) 0-0 U C:~~~4 -U) - -0( a ) C: a) U-0 *-= 4 - 0 a U)c >C ~a) Cn 4-. .0~U U)) (LU a) 0U a)0 U) o rt f L U) 0 :00C:Q U) CD C ) a) C- -(' _ o -0 m ) ) a -m cLrJ an(n 4 ) 0 4- 4 Z ) C ) a o0 M 0 a - | en | DUPLICATE AND SIMILAR BUG REPORT DETECTION AND RETRIEVAL USING NEURAL NETWORKS | 54000148_IE | 44134498_,54867627_,54974301_,54057601_,57129009_ | G06F 7/023,G06F 11/00,G06F 11/3672,G06F 11/3692,G06F2201/81,G06F2207/4824,G06N 3/04,G06N 3/0445,G06N 3/0454,G06Q 10/06 | [
"G06F 11/36"
] | 119,036 |
34,394,835 | 1988-08-22 | 16,557,590 | N | PURPOSE:To improve the stability of a negative type silver halide color photographic sensitive material against variation of processing conditions by forming at least one silver halide emulsion layer contg. at least two kinds of negative type silver halide particles having substantially different content of desensitizer to each other on a base body and incorporating a DIR compd. into the silver halide emulsion layer. CONSTITUTION:In the above described silver halide emulsion layer, a mean desensitizer content of a partial particle group corresponding to 5wt.% of the whole particle group and having a higher desensitizer content than the residual particle group corresponding to 95wt.% of the whole particle group is regulated to >=10 times of a mean desensitizer content of another partial particle group corresponding to 5wt.% of the whole particle group but having lower desensitizer content than residual particle group corresponding to 95wt.% of the whole particle group. The DIR compd. to be used is a compd. capable of eliminating a compd. which releases a development restainer by the reaction with an oxidized body of a color developing agent. By this constitution, a higher stability is provided to a negative type silver halide color photographic sensitive material against the variation of processing conditions. | en | NEGATIVE TYPE SILVER HALIDE COLOR PHOTOGRAPHIC SENSITIVE MATERIAL | 30340472_ | 30468113_,30444681_,30372016_,30408798_,30451752_ | G03C 1/035,G03C 7/30,G03C 7/3022,G03C2001/03564 | [
"G03C 1/035",
"G03C 1/36",
"G03C 7/26",
"G03C 7/30"
] | 24,899 |
556,586,817 | 2020-04-21 | 70,993,896 | N | A virtual reality-based caregiving machine control system, comprising: a visual unit, used for obtaining environmental information around a caregiving machine, and transmitting the environmental information to a virtual scene generation unit and a calculation unit; the calculation unit, used for receiving control instructions for the caregiving machine, and obtaining, by calculation according to the environmental information, an action sequence of executing the control instructions by the caregiving machine; the virtual scene generation unit, used for generating a virtual reality scene from the environmental information, and displaying the virtual reality scene on a touch display screen in combination with the action sequence; and the touch display screen, used for receiving a touch screen adjusting instruction for the action sequence and feeding back same to the calculation unit for execution, and receiving a confirmation instruction for the action sequence. By using the above-mentioned solution, human-machine interaction is increased, natural and effective interactions between a caregiving machine and a person and between the caregiving machine and the environment are implemented, the probability of an unknown error is avoided, and the probability of causing harms to the user and the environment is reduced. | en | VIRTUAL REALITY-BASED CAREGIVING MACHINE CONTROL SYSTEM | 63673205_CN | 83049733_CN,82736495_CN,82734669_CN,69070670_CN,68310929_CN,67872635_CN | B25J 9/1602,B25J 11/009,G06F 3/011,G06F 3/017,G06F2203/012 | [
"B25J 11/00",
"B25J 9/16",
"G06F 3/01"
] | 157,305 |
557,439,428 | 2019-01-05 | 67,209,163 | Y | FIELD: communication equipment.SUBSTANCE: invention relates to a head phone based on bone conduction and, in particular, to a speaker based on a bone surface. A bone conduction speaker contains: a magnetic circuit component made with the possibility of providing a magnetic field; a vibration component, with at least part of the vibration component located in the magnetic field and converting the electrical signal introduced into the vibration component into a mechanical vibration signal; and a housing containing a housing panel facing the side of the human body, and a rear part of the housing opposite to the housing panel. The housing accommodates a vibration component. The vibrating component causes the housing panel and the back of the housing to vibrate. The vibration of the housing panel has the first phase, and the vibration of the rear part of the housing has the second phase. When the vibration frequency of the housing panel and the vibration frequency of the rear part of the housing are in the range from 2000 to 3000 Hz, the absolute value of the difference between the first phase and the second phase is less than 60 degrees.EFFECT: creation of a bone conduction-based speaker to simplify the structure of a bone conduction-based speaker, reducing sound leakage, and improving sound quality.10 cl, 45 dwg | en | BONE CONDUCTION-BASED SPEAKER AND ITS TESTING METHOD | 83148585_CN | 82914282_CN,82823269_CN,82794684_CN,82986620_CN | H04R 1/1091,H04R 1/2803,H04R 9/02,H04R 9/025,H04R 9/06,H04R 9/066,H04R 29/003,H04R2460/13 | [
"H04R 9/06",
"H04R 1/10",
"H04R 9/02"
] | 157,822 |
47,339,932 | 1997-01-14 | 26,680,689 | N | The extraction and pre-cueing of axon fibers and clusters is performed on transmission electron micrograph (TEM) image mosaics to reduce neuroanatomist workload. Given a multiplicity of myeling-sheathed axon fibers in a cross-sectional image containing many such fibers, the inventive method uses these grey-level operators to effectively identify co-occurrences of the axon fibers and their myelin sheaths and output a result representative of these co-occurences. The co-occurrences may then be sorted in terms of their geometry to identify groups of fibers indicative of axon clusters. Specifically, a grey-level opening and conditional dilation is performed on the image to obtain a first residue representative of potential axons. A grey-level closing and conditional erosion of the first opening image is then performed to obtain a second residue representative of potential axon fibers, including their respective myelin sheaths. The first residue may then be used as a marker to conditionally dilate over the second residue to identify the co-occurrences. The method of the invention significantly reduces the workload of the reviewers by identifying rougly 95 percent of the axons, with only a one to two percent false alarm rate, based on a reasonable computation time of about one hour per sample on a fast workstation. | en | METHOD OF EXTRACTING AXON FIBERS AND CLUSTERS | 6667580_US | 37767320_ | G06V 20/69 | [
"G06K 9/00"
] | 34,789 |
16,816,179 | 1990-06-27 | 15,855,761 | Y | A visual information processing device has a pair of neural networks which respectively comprise an upper layer and a lower layer of the device. Each of the pair of neural networks comprises a semiconductor integrated circuit having a plurality of neuron circuit regions which are disposed in a matrix form, each of the neuron circuit regions performing a neuron function; a molecule film having a photoelectric function and provided on the semiconductor integrated circuit, the molecule film having (i) a plurality of Tij signal input sections each performing a wiring function among the plurality of neuron circuit regions, in each of which a Tij signal representing the bonding strength among the plurality of neuron circuit regions is optically written, and (ii) a plurality of video input sections each performing a sensor function of sensing a visual image in which one pixel corresponds to one neuron circuit region; and a wiring for electrically connecting the semiconductor integrated circuit and the molecule film. Each of the plurality of neuron circuit regions is bonded with the neighboring neuron circuit regions in each of the pair of neural networks comprising the upper and lower layers, and each of the plurality of neuron circuit regions is bonded with the corresponding one between the pair of neural networks. | en | A visual information processing device. | 41231_JP | 2889461_JP,2889460_JP,2889462_JP,2889459_JP,2889458_JP | G06N 3/063,H01L 27/148 | [
"H01L 27/146",
"G06E 1/00",
"H01L 27/148",
"H01L 27/14",
"G02F 3/00",
"H01L 27/00",
"G06N 3/08",
"G06N 3/04",
"G06N 3/063",
"G06N 99/00",
"G06F 15/18"
] | 17,886 |
35,266,909 | 1998-08-31 | 8,229,841 | N | PROBLEM TO BE SOLVED: To provide the artificial intelligence device which actualizes architecture which is metaphorically similar to the physical desire of a human who experiences it. SOLUTION: The artificial intelligence device is equipped with one application part 13 which has one operation request and outputs a request signal (x) indicating a request state when the operation request is not complied with, a control part 11 which receives the request signal (x) and generates a control signal (y) reaching the development (a) of the presence of an insufficient operation request that an external agent or user senses, and a modulation part 14 which detects an external input or event (c). Further, the device is equipped with a conditioning part 12 which receives a signal βindicating the presence of the insufficient operation request at the application part 13, a signal χinstructing the external agent or user to sense the development of the presence of the insufficient operation request, and a signal α indicating the detection of the sufficiency of the input of an external origin or the event (c) and generates a signal ε indicating the possibility of the sufficiency of the current insufficient operation request based on values βt-n, χt-n, and αt-n of the receive signals of previous insufficient operation requests. | en | ARTIFICIAL INTELLIGENCE DEVICE | 13677984_ | 30381023_ | G06N 3/004,G06N 5/043 | [
"G06F 15/18",
"G05B 13/02",
"G06N 3/00",
"G06N 99/00",
"G06N 5/04"
] | 24,973 |
571,001,759 | 2019-08-14 | 69,643,651 | N | The present technology relates to an information processing device, an information processing method, and a program that enable a virtual object to be displayed at an appropriate position in the real world. An information processing device includes: a position estimation unit that estimates a current position in a first coordinate system; a display position setting unit that sets a display position of a virtual object in a third coordinate system on the basis of an environment database in which multiple pieces of environment data including a reference image or reference feature point data and reference meta information are registered, virtual object information including the display position of the virtual object in a second coordinate system, and an observed image captured near the current position; a meta information generation unit that generates observation meta information; and an environment database management unit that compares observation data including the observed image or observation feature point data containing data related to a feature point of the observed image and the observation meta information with environment data of the environment database to determine whether to add the observation data to the environment database. The present technology can be applied to a device that implements AR. | en | INFORMATION PROCESSING DEVICE, INFORMATION PROCESSING METHOD, AND PROGRAM | 5210885_JP | 81806341_JP,79575652_JP,6902859_JP | G06K 9/6215,G06T 19/00,G06T 19/006,G06V 10/761,G06V 20/56,G06V 20/58,G06V 20/588 | [
"G06V 20/58",
"G06V 20/56",
"G06T 19/00",
"G06V 10/74"
] | 165,146 |
2,775,994 | 2008-02-08 | 27,663,091 | Y | METHODS OF TREATING PATIENTS SUFFERING FROM MOVEMENT DISORDERS Abstract 5 The present invention is directed to methods of treating movement disorders by administering an effective amount of one or more adenosine A2 receptor antagonist to a patient in need thereof. The present invention also relates to methods of decreasing the adverse effect of L-DOPA in patients receiving L-DOPA therapy in the treatment of Parkinson's disease. The present invention further relates to methods and compositions 1o for treating Parkinson's disease patients with sub-clinically effective doses of L-DOPA by combining L-DOPA treatment with an effective amount of one or more adenosine A2 receptor antagonises (i.e., L-DOPA sparing effect). The present invention further relates to methods of effective treatment of Parkinson's disease by co-administering at least one adenosine A2 receptor antagonist, L-DOPA and a dopamine agonist and/or a COMT 15 inhibitor and/or a MAO inhibitor. The present invention further relates to methods of prolonging effective treatment of Parkinson's disease by administering an adenosine A2 receptor antagonist singly or together with a dopamine agonist, and/or a COMT inhibitor, and/or a MAO inhibitor without prior or subsequent administration of L-DOPA, delaying or removing on-set of L-DOPA motor complication. | en | Methods of treating patients suffering from movement disorders | 74215196_JP | 15101493_,15101495_,15101494_,15101492_,15101496_,13225097_ | A61K 9/0019,A61K 9/20,A61K 9/48,A61K 31/198,A61K 31/519,A61K 31/522,A61K 45/06,A61P 25/00,A61P 25/02,A61P 25/14,A61P 25/16 | [
"A61K 31/522",
"A61P 25/16",
"C07D 473/06",
"A61K 45/00",
"A61P 25/14"
] | 3,109 |
537,588,134 | 2020-07-23 | 67,990,596 | N | Object identification method comprising detecting a candidate object in an image (12, Fig. 1) and comparing candidate data with locally stored templates (exemplars) S204. Failure to match results in the identification of object identifiers. The object identifiers are used to search a further object data store S206 to identify the candidate object. Object data may comprise: lines; edges; ridges; corners; blobs; textures; shapes; gradients; regions; boundaries; surfaces; volumes; colours; and/or shadings. Object templates may comprise: 3D models; neural network weights; image sets; and/or eigenimages. Object identifiers may be from: the object: the object’s environment; or another object within the environment. Object identifiers may be machine-readable markings such as: barcodes; QR codes; matrix codes; microchips; and/or URLs (14, 16, 18, Fig. 2). Object identifiers may be used in a web-search to identify the further data store. Templates may be downloaded from the remote database to the local datastore when they provide a match. An identification confidence rating may be generated and stored with the associated template. The further object data store may provide secondary object data (e.g. best before/expiry date, manufacturer etc) S210. A user may be presented with data pertaining to the identified object. | en | Recognition apparatus and method | 48759759_GB | 28108417_GB,50329079_GB | G06K 7/10861,G06K 7/1443,G06V 10/10,G06V 10/22,G06V 10/44,G06V 20/00,G06V 20/20,G06V 20/52,G06V 20/63,G06V 20/64,G06V 20/68 | [
"G06K 9/20",
"G06K 9/00"
] | 144,587 |
566,445,971 | 2021-09-29 | 58,631,565 | N | To enable the accuracy and speed of analysis and classification to be improved while maintaining an optical system.SOLUTION: An analysis method comprises: a step of forming a structured illumination pattern having a plurality of regions into which some region of light from a light source is divided, and in which light characteristics free of time changes are different, and irradiating an observation object present in a light irradiation region with illumination light having the illumination pattern while varying the relative positions of the illumination pattern and the observation object; a step of receiving, by a receive unit, at least one light signal selected from scattered light, Raman scattered light, transmitted light, fluorescence and electromagnetic waves from the observation object; a step of converting the light signal into an electric signal by the receive unit; and a step of analyzing the observation object on the basis of a time-series signal extracted on the basis of the time base of the electric signal. Information regarding the observation object is analyzed from the time-series signal without reconstructing the image of the observation object, on the basis of the result of machine learning conducted using the time-series signal pertaining to the observation object.SELECTED DRAWING: Figure 1 | en | ANALYSIS METHOD | 28160467_,65860770_ | 66247512_,70729981_,62645651_,85318199_,58868865_,84243178_,58600617_ | C12M 1/34,G01N 15/14,G01N 15/1404,G01N 15/1429,G01N 15/1434,G01N 15/1459,G01N 15/147,G01N 21/01,G01N 21/27,G01N 21/64,G01N 21/65,G01N2015/1006,G01N2015/1413,G01N2015/1415,G01N2015/145,G06K 9/00536,G06K 9/6255,G06K 9/6263,G06V 10/82,G06V 20/698 | [
"G01N 21/53",
"G01N 15/14",
"G01N 21/17",
"G01N 21/64"
] | 162,914 |
478,082,879 | 2016-09-20 | 58,614,945 | N | The invention provides a novel method for solving the lexical semantic similarity between ontology concepts. The method comprises the steps of calculating the similarity between the ontology concepts, with the maximum depths, of to-be-compared words input into a statistical method module; calculating the word form similarity between the two to-be-compared words, calculating the influence of the most recent common ancestor depth of the two to-be-compared words on the similarity between the two to-be-compared words and constructing an impact factor function; and finally calculating the similarity between the two to-be-compared words. The novel method is closer to an empirical value of an expert in quantitative concept; the factors of the distance between the ontology concepts, with corresponding maximum depths, of the to-be-compared words, the depths, the density and the like are more fully and comprehensively considered, so that the accuracy of the semantic similarity result is greatly improved; the ontology reasoning effect is better improved; better improvement of the accuracy of the word form similarity result and the semantic similarity result between the words is considered; data processing of various influence factors is more standard; and the novel method accords with the practical application effect. | en | Novel method for solving lexical semantic similarity between ontology concepts | 65674138_ | 59108801_ | G06F 40/30 | [
"G06F 17/27"
] | 108,571 |
481,253,852 | 2016-01-19 | 59,273,352 | N | A naked-eye 3D display pixel unit and a display device having a 2D or 3D switching function. A naked-eye 3D display pixel unit (1) consists of a pixel unit (110), an optical light mixing region (120), and an optical guiding region (130). The pixel unit (110) comprises at least red, green, and blue light emitting units (110a, 110b, 110c). The light emitting units (110a, 110b, 110c) of the colors each consist of a substrate (111), a first electrode layer (112), a first doping layer (113), a second doping layer (114), and multiple second electrode layers (115, 115a, 115b). The pixel unit (110) projects light at a direct angle to form a 2D image, or alternately projects, by controlling electric fields, light along left-eye and right-eye directions according to a time sequence to form left-eye and right-eye parallax images, and then the images are merged to form a 3D image in a human brain. The naked-eye 3D display pixel unit (1) is minimized to a pixel level, and a large-area display device (20) consists of the naked-eye 3D display pixel unit (1) is not limited by the sizes of a panel and an optical lens, can be installed to have a flat or curved screen, and has the function of displaying 2D/3D images in a split-screen mode at the same time or in a full-screen mode by switching 2D/3D modes according to content. | en | NAKED-EYE 3D DISPLAY PIXEL UNIT AND DEVICE HAVING 2D OR 3D SWITCHING FUNCTION | 66813916_CN | 66813916_CN,64312009_CN | G02B 30/24,G02B 30/27,G02B 30/30,G09G 3/20,H01L 25/048,H01L 25/0753,H01L 27/3232,H01L 51/5225,H04N 13/305,H04N 13/31,H04N 13/32,H04N 13/341,H04N 13/351,H04N 13/356,H04N 13/398 | [
"G09G 3/20",
"G02B 27/22"
] | 110,143 |
490,749,726 | 2017-08-25 | 61,246,294 | N | Systems and methods for processing neural signals are provided. A neural data analysis system may comprise a feature extraction module configured to (1) extract a plurality of features from neural signal waveforms obtained by an implanted neural interface probe with a plurality of channels or electrodes, wherein the plurality of features are extracted from the neural signal waveforms without requiring prior digitization of the neural signal waveforms, and (2) to transmit the extracted features as a plurality of discrete outputs. The neural data analysis system may also comprise a feature-event coalescence module configured to: (1) receive the plurality of discrete outputs from the feature extraction module, and (2) construct a model-based inference of bioelectric activity based on feature-event statistics, prior knowledge of bioelectric signals, and/or a behavioral model of the feature-extraction module. The neural data analysis system may further comprise an approximator module configured to (1) receive a plurality of coalesced events from the feature-event coalescence module, and (2) apply a series of transformations to the coalesced event data to generate a higher entropy neural code, wherein the neural code comprises a representation of ensemble activity of a plurality of neurons recorded by the system. | en | SYSTEM AND METHODS FOR PROCESSING NEURAL SIGNALS | 56462486_US,57695202_US | 48846921_US,57665013_US,57695202_US | A61B 5/24,A61B 5/291,A61B 5/316,A61B 5/374,A61B 5/4064,A61B 5/6868,A61B 5/7225,A61B 5/7239,A61B 5/7242,A61B 5/7246,A61B 5/7253,A61B 5/7267,A61B 5/7282,A61B2560/0209,A61N 1/0531,A61N 1/0534,G16H 50/70 | [
"A61N 1/02",
"A61B 5/374",
"A61N 1/05",
"A61N 1/04"
] | 115,793 |
54,135,527 | 1992-07-17 | 27,338,898 | Y | In a focus detection device for a camera, a plurality of distance sensors each for detecting a distance to an object in a plurality of areas of a photographing image plane are provided and distance data obtained by the distance sensors for each object in each area is supplied to a main object detection circuit and a normalizing circuit. The normalizing circuit normalizes the distance data into a real number ranging from 0 to 1 and then supplies the same to a neural network. The neural network formed of a single-layered neuron units of which the synapse connection weighting factors are previously obtained by the learning process, calculates a vector difference between the distance data and the synapse connection weighting factors of each neuron unit, detects the minimum vector difference, and outputs position data of a main object corresponding to a neuron unit which gives the minimum vector difference. The position data of the main object is input to a main object detection circuit. One of the outputs from the distance sensors corresponding to the main object is selected by the main object detection circuit and is supplied to a focus detection circuit for effecting the calculation to detect the focus. An output of the focus detection circuit is supplied to a lens driving mechanism so as to adjust the focus. | en | Automatic camera | 5238396_JP | 9422709_JP,5276981_JP | G02B 7/285,G03B 7/091,G03B 7/09979,Y10S 706/903 | [
"G03B 7/091",
"G02B 7/28",
"G03B 7/099"
] | 48,215 |
54,262,468 | 1992-09-10 | 18,003,466 | Y | A method which includes performing a structure analysis on a natural sentence inputted by making use of a word dictionary DIC-WD and a configuration dictionary DIC-KT and converting letter series KNJ of the inputted natural sentence into a language structure information series IMF-LSL. The natural sentence inputted in the form of the language structure information series IMI-LSL is subjected in such a manner to application of meaning analysis grammar IMI-GRM to cause a single or a plurality of meaning frames IMF-FRM to be read out from a meaning frame dictionary DIC-IMI in accordance with commands of the meaning analysis grammar IMI-GRM. When a plurality of meaning frames IMI-FRM are read out a meaning frame which defines an abstract meaning expressed by the inputted natural sentence is synthesized by case coupling and/or logic coupling the meaning frames IMI-FRM. Words WD, particles JO and symbols KI are inserted into the meaning frames IMI-FRM read out or the meaning frame IMI-FRM synthesized to thereby determine and produce data sentence DT-S correctly expressing the meaning of the inputted natural sentence in a computer whereby the language structure information series IMF-LSL is converted into the data sentence DT-S in the form of data structure PSMW with a multi layered case-logic language structure. | en | Natural language processing method for converting a first natural language into a second natural language using data structures | 10586790_JP | 10586791_JP | G06F 40/242,G06F 40/253,G06F 40/55 | [
"G06F 17/27",
"G06F 17/28"
] | 48,591 |
538,218,457 | 2020-03-04 | 72,335,392 | Y | A method of transferring a facial expression from a subject to a computer generated character that includes receiving a plate with an image of the subject's facial expression, a three-dimensional parameterized deformable model of the subject's face where different facial expressions of the subject can be obtained by varying values of the model parameters, a model of a camera rig used to capture the plate, and a virtual lighting model that estimates lighting conditions when the image on the plate was captured. The method can solve for the facial expression in the plate by executing a deformation solver to solve for at least some parameters of the deformable model with a differentiable renderer and shape from shading techniques, using, as inputs, the three-dimensional parameterized deformable model, the model of the camera rig and the virtual lighting model over a series of iterations to infer geometry of the facial expression and generate a final facial mesh using the set of parameter values of the deformable model which result in a facial expression that closely matches the expression of the subject in the plate. Initial Facial Mesh 502 , 510 [APlate and Additional Inputs 0 Solve Deformation 506 from Plate Differentiable Renderer Final Facial Mesh ------------- - 510 Solver 1 Solver 2 Solver 3 I = Solver M | en | ON-SET FACIAL PERFORMANCE CAPTURE AND TRANSFER TO A THREE-DIMENSIONAL COMPUTER-GENERATED MODEL | 39782263_US | 77989412_,79259765_,78767390_,75035604_,75077615_,75109950_,78470742_,78071831_,77959302_ | G06T 13/40,G06T 15/50,G06T 17/20,G06T 19/20,G06T2219/2021,G06V 20/653,G06V 40/174,G06V2201/12,H04N 5/2224 | [
"G06K 9/46"
] | 144,924 |
551,186,603 | 2021-05-27 | 76,355,233 | N | The present application discloses a method and an apparatus (500) for acquiring sample deviation data and an electric device (600), which relate to the fields of artificial intelligence technology, automatic driving technology, intelligent transportation technology and deep learning technology. The specific implementation solution is: in case of acquiring the sample deviation data, a first driving behavior parameter of a vehicle and a second driving behavior parameter of the vehicle in a simulated automatic driving mode are respectively acquired in the manual driving mode; and it is determined whether there is a deviation between the first driving behavior parameter and the second driving behavior parameter, and if there is a deviation, the vehicle is controlled to acquire the sample deviation data within a preset time period. In this way, by establishing a running mechanism of the simulated automatic driving mode, and comparing the driving behavior parameters output in the simulated automatic driving mode with the driving behavior parameters in a real manual driving mode, the vehicle can automatically monitor a driving deviation behavior and can be triggered to automatically collect the sample deviation data when there is the deviation, thereby improving efficiency for acquiring the sample deviation data. | en | METHOD AND APPARATUS FOR ACQUIRING SAMPLE DEVIATION DATA, AND ELECTRONIC DEVICE | 74608609_CN | 81655142_CN,81637817_CN,81636732_CN,81750620_CN,81606293_CN | B60W 40/09,B60W 50/085,B60W 60/0053,B60W2050/0005,B60W2050/0028,B60W2050/005,B60W2050/0071,B60W2050/0088,B60W2520/28,B60W2540/10,B60W2540/12,B60W2540/16,B60W2540/18,B60W2540/30,B60W2554/00,G06F 11/3692,G06N 3/08,G06N 5/046 | [
"G06F 11/36",
"B60W 40/00"
] | 153,700 |
414,094,494 | 2013-07-12 | 49,915,206 | N | The ACSSECR invention is a biometric access control system and methodology that measures cognitive, psychophysiological responses to stimuli to confirm the identity of an individual. As an alternative to 'Logging in' with a user ID and password, this cognitive biometric authentication system is used for 'Cogging in' to a system with user ID and user-selected 'Cogkey'. ACSSECR is designed for strict access control scenarios where significant authentication confidence is required to gain access to controlled information, facilities, systems, vehicles, or devices. The system takes advantage of a behavioral and physiological characteristic of humans that is an unconscious response to a stimulus. The Event Related Potential (ERP) response (specifically the P3 ERP) involuntarily occurs when an individual perceives and reacts to an unexpected, task-relevant event. The task is for the user to recognize their Cogkey which is presented infrequently amidst more frequent non-target stimuli. There is no requirement for extensive enrollment by users, only the recognition of their Cogkey. The basic system does not store biometric data for comparison, but rather measures the user's Cogkey recognition responses in comparison to non-Cogkey stimulus responses. An individual can have multiple personas with different Cog keys. | en | Access Control System using Stimulus Evoked Cognitive Response | 45396025_US | 45396025_US | A61B 5/117,A61B 5/377,G06F 21/32,G07C 9/37 | [
"G06F 21/32"
] | 84,294 |
476,587,214 | 2016-08-18 | 54,484,633 | N | A multi-functional event-related potential (ERP) experiment chamber comprises an experiment chamber (1) and a laboratory (2) arranged neighboring the experiment chamber (1). The experiment chamber (1) is an experiment chamber having conventional electromagnetic shielding, and a device (14) for air return in the experiment chamber, an independent ground wire and power device, a background music player (11), an infrared video monitoring device (12), a chamber voice intercommunication device, an electric curtain control device (9), a lighting device (13) controllable from both inside and outside of the experiment chamber, an oxygen negative ion supply device (15), and an experiment table and chair are mounted or placed in an interior of the experiment chamber (1). A control console (5) is provided within the laboratory (2) to separately control the background music player (11), infrared video monitoring device (12), chamber voice intercommunication device, electric curtain control device (9), and lighting device (13) controllable from both inside and outside of the experiment chamber. The present invention creates an environment meeting various specific requirements for an ERP experiment with a smaller space, thereby facilitating improvement of EPR experiments and promotion and applications of ERP technology. | en | MULTI-FUNCTIONAL EVENT-RELATED POTENTIAL EXPERIMENT CHAMBER | 67849826_CN | 63527875_CN | A61B 5/369 | [
"A61B 5/0476"
] | 107,677 |
437,177,996 | 2013-12-17 | 52,475,846 | Y | The present invention relates to a method and a system for transmitting a user emotion during a game. More specifically, the method and the system are capable of: intuitionally recognizing the selection of an emotion type by a user; effectively and promptly transmitting an emotion in the game being played fast by determining, at once, the selection of the emotion and the degree of the emotion by even a single selection gesture; and promptly responding to emotion expression of a counterpart. The method according to the present invention includes the steps of: displaying an emotion image received from a counterpart terminal in a display of a terminal to play a game; displaying at least two or more emotion menus based on the display emotion image; firstly selecting a type of an emotion according to user selection for one of the displayed two or more emotion menus; secondly selecting the degree of the emotion according to a distance between a location of the display in which a user selection signal is generated and the emotion image; varying the emotion image according to the firstly selected type of the emotion and the secondly selected degree of the emotion; and generating a response emotion image from the varied emotion image and transmitting the generated response emotion image to the counterpart terminal. | en | METHOD AND SYSTEM FOR EXPRESSING EMOTION DURING GAME PLAY | 47395576_KR | 47486888_KR,12609987_KR | G06Q 50/10 | [
"G06Q 50/10"
] | 91,946 |
423,279,150 | 2013-12-18 | 49,881,121 | N | The present invention provides mild cleansing and cleaning compositions and methods for mitigating irritation induced by harsh detersive surfactants contained therein. The invention relates to a method of reducing skin irritation associated with a cleansing composition comprising at least one surfactant, the method comprising combining an effective amount of at least one nonionic amphiphilic polymer with at least one detersive surfactant selected from anionic surfactants, amphoteric surfactants, nonionic surfactants and combinations of two or more thereof. The at least one nonionic amphiphilic irritation mitigating polymer is prepared from a free radically polymerizable monomer composition comprising at least one hydrophilic monomer and at least one hydrophobic monomer, wherein said hydrophilic monomer is selected from hydroxy(C1-C5)alkyl (meth)acrylates, N-vinyl amides, amino group containing monomers, or mixtures thereof; wherein said hydrophobic monomer is selected from esters of (meth)acrylic acid with alcohols containing 1 to 30 carbon atoms, vinyl esters of aliphatic carboxylic acids containing 1 to 22 carbon atoms, vinyl ethers of alcohols containing 1 to 22 carbon atoms, vinyl aromatic monomers, vinyl halides, vinylidene halides, associative monomers, semi-hydrophobic monomers, or mixtures thereof. | en | Irritation mitigating polymers and uses therefor | 12497920_US | 47465266_US,13176392_US,24611166_US | A61K 8/442,A61K 8/8152,A61K2800/5422,A61K2800/75,A61Q 19/10,C11D 1/143,C11D 1/90,C11D 3/3757 | [
"C11D 1/83",
"C08F 20/10",
"C11D 3/37"
] | 89,854 |
494,529,731 | 2018-04-24 | 51,297,803 | N | A computer and digital processing system involving interprogram or interprocess communication for risks in a combined booked and pari-mutuel environment, for the purposes of gaming and/or wagering. For example, it may enable users to wager on the outcome of events, eg a horse race, a dog race, auto racing, baseball, basketball, boxing, football, golf, hockey, poker, political races, weather, bingo, keno, a jai alai game, which team will win a match or tournament and so on. This computer includes a network interface to communicate with remote devices over an electronic network and a processor to receive, via the network interface, data indicative of a request for a first wager on an event from a first remote device, generate data indicative of the first wager between the remote device and a gaming operator, in response to receiving the request, identify a proportion of the first wager to place with a pari-mutuel pool of wagers, generate data indicative of a second wager between the gaming operator and the pari-mutuel pool of wagers, the second wager comprising the proportion of the first wager, receive, via the network interface, data indicative of an outcome of the event, resolve the first wager based on the outcome of the event and resolve the second wager with the pool based on the outcome of the event. | en | ELECTRICAL COMPUTERS AND DIGITAL PROCESSING SYSTEMS INVOLVING INTERPROGRAM OR INTERPROCESS COMMUNICATION FOR RISKS IN A COMBINED BOOKED AND PARI-MUTUEL ENVIRONMENT | 12482523_US | 14995070_,14996764_ | G07F 17/3258,G07F 17/3272 | [
"A63F 9/24",
"G06F 17/00",
"A63F 13/00"
] | 117,739 |
522,293,954 | 2019-05-14 | 64,343,022 | N | A three-dimensional geometry and eigencomponent reconstruction method and apparatus based on light and shadow optimization. The method comprises: obtaining photographing of a three-dimensional color point cloud time series by means of an RGBD camera (S101); acquiring matching point pairs between a three-dimensional depth point cloud and vertexes of a reconstructed model, and obtaining a point pair set (S102); establishing a joint energy function based on eigendecomposition according to matching point pairs and a current view angle color image, and solving non-rigid motion position transformation parameters of each vertex on the reconstruction model (S103); solving the energy function to obtain a deformation transformation matrix of vertexes of a surface model and eigencomponents of each item on an image (S104); and deforming the geometry of a previous frame three-dimensional model according to a solved result, and completing and updating the geometry and eigencomponents of a current frame model to realize three-dimensional geometry and eigencomponent reconstruction (S105). The method can improve the robustness of dynamic object tracking and deformation under the conditions of sparse viewpoints and a single viewpoint, can realize accurate solving, has low requirements for devices, and has a broad prospect. | en | THREE-DIMENSIONAL GEOMETRY AND EIGENCOMPONENT RECONSTRUCTION METHOD AND DEVICE BASED ON LIGHT AND SHADOW OPTIMIZATION | 6155378_CN | 57530505_CN,19453064_CN,50256390_CN | G06T 15/50,G06T 17/00,G06T2210/44 | [
"G06T 17/00"
] | 134,090 |
571,825,517 | 2021-09-15 | 74,696,379 | N | The present invention discloses a real-time binocular stereo matching method based on an adaptive candidate parallax prediction network. The method comprises: first performing multi-scale feature extraction on a stereo image pair by using a two-dimensional convolutional neural network so as to obtain high-resolution and low-resolution feature maps; then, in the first-stage parallax estimation, using the low-resolution feature map to perform parallax coarse estimation; after an adaptive candidate parallax is predicted by using a coarse estimation result and left image information, in the second-stage parallax estimation, performing fine parallax estimation by using a prediction result and the high-resolution feature map; and finally, performing hierarchical refinement on a parallax map to obtain a full-size parallax map. Compared with existing coarse-to-fine stereo matching neural networks, the present invention allows for more accurate dynamic offset prediction for a fine parallax estimation stage so as to meet different parallax correction requirements of various targets in the images. Due to the effectiveness of dynamic prediction, a two-stage processing structure is designed in the present invention to greatly improve the calculation precision and speed of a real-time binocular stereo matching network. | en | REAL-TIME BINOCULAR STEREO MATCHING METHOD BASED ON ADAPTIVE CANDIDATE PARALLAX PREDICTION NETWORK | 68578237_CN | 63529655_CN,85169658_CN,64130423_CN,63800975_CN,67717664_CN | G06N 3/0454,G06T 7/33,G06T 7/593,G06T2207/10012,G06T2207/20081,G06T2207/20084,G06T2207/20228 | [
"G06T 7/33",
"G06T 7/593",
"G06N 3/04"
] | 165,661 |
39,692,946 | 2000-01-29 | 19,577,391 | N | A real-time brain wave measuring apparatus and method for measuring the brain wave of a subject and transmitting the measured brain wave via a serial port to a brain wave analyzing computer in real time, is provided. The brain wave measuring apparatus includes a brain wave sensor for sensing brain wave signals of multiple channels at predetermined portions of the subject's scalp using a plurality of electrodes, an interface and a serial port connector connected to the interface via a cable and connected to the serial port. The interface includes an amplifier for amplifying the amplitudes of the brain wave signals of multiple channels, the brain wave signals which are sensed by the brain wave sensor; an analog-to-digital converter for sampling and digitizing the amplified brain wave signals of multiple channels; an encoder for coding a digital value calculated by the analog-to-digital converter of each channel in real time; and a computer interface for transmitting the brain wave signals which are coded by the encoder to the serial port. Accordingly, brain wave signals of multiple channels are sensed and transmitted through the serial port to the brain wave analyzing computer in real time even if the brain wave analyzing computer is not equipped with special hardware for processing brain waves therewithin. | en | REAL-TIME BRAIN WAVE MEASURING APPARATUS USING HEADBAND AND BRAIN WAVE TRANSMISSION METHOD | 32589768_KR,32589776_KR,32589777_KR,32589770_KR,32589769_KR,32589771_KR | 32589770_KR,32589776_KR,32589777_KR,32589771_KR | A61B 5/282,A61B 5/291 | [
"A61B 5/0478",
"A61B 5/0408"
] | 26,079 |
4,389,043 | 1991-02-25 | 23,931,980 | N | 2051386 9113430 PCTABS00007 A method is disclosed for use in preprocessing noisy speech to minimize likelihood of error in estimation for use in a recognizer. The computationally-feasible technique, herein called Minimum-Mean-Log-Spectral-Distance (MMLSD) estimation using mixture models and Markov models, comprises the steps of calculating for each vector of speech in the presence of noise corresponding to a single time frame, an estimate of clean speech, where the basic assumptions of the method of the estimator are that the probability distribution of clean speech can be modeled by a mixture of components each representing a different speech class assuming different frequency channels are uncorrelated within each class and that noise at different frequency channels is uncorrelated. In a further embodiment of the invention, the method comprises the steps of calculating for each sequence of vectors of speech in the presence of noise corresponding to a sequence of time frames, an estimate of clean speech, where the basic assumptions of the method of the estimator are that the probability distribution of clean speech can be modeled by a Markov process assuming different frequency channels are uncorrelated within each state of the Markov process and that noise at different frequency channels is uncorrelated. | en | METHOD FOR SPECTRAL ESTIMATION TO IMPROVE NOISE ROBUSTNESS FOR SPEECH RECOGNITION | 16130079_US | 16349400_IL,16349401_IL | G10L 15/20,G10L 21/02 | [
"G10L 15/02",
"G10L 15/20",
"H04B 15/00",
"G10L 21/02"
] | 4,286 |
546,025,568 | 2019-09-11 | 68,434,927 | Y | The present application discloses a fiber distributed acoustic sensing system, including a forward pump source, a wavelength division multiplexer, an active phase-shifted grating array, a backward pump source, and a data demodulation and processing device. The active phase-shifted grating array includes several active phase-shifted gratings engraved on a same fiber, each active phase-shifted grating having a same excitation light wavelength. The active phase-shifted grating array is configured to receive a forward pump pulse and a backward pump light pulse incident from the backward pump source, so that a forward excitation light and a backward excitation light are generated in each active phase-shifted grating, and two adjacent active phase-shifted gratings are enabled to generate excitation light self interference within a same pulse duration. The data demodulation and processing device is configured to perform interference detection on a fiber phase between two adjacent active phase-shifted gratings to detect an acoustic signal acted on the fiber. According to the system in the present application, the passive reflected light interference is replaced with active excitation light interference, thus the excitation light signal has great strength and high stability, and environmental adaptability is high. | en | Fiber distributed acoustic sensing system | 57906196_CN | 56986784_CN,56856993_CN,56816532_CN,56824796_CN,56828214_CN,79538765_CN,79548054_CN,56856809_CN | G01D 5/344,G01D 5/35316,G01D 5/35387,G01H 9/004 | [
"G01D 5/353",
"G01H 9/00",
"G01D 5/34"
] | 150,047 |
197,032 | 2007-03-23 | 38,541,291 | Y | A reference-color measurement step of obtaining a reference-color measurement value by measuring a spectroscopic-radiation luminance of a light being emitted from a reference-color portion in a measurement direction, or a tristimulus value thereof, using a light-source-color measuring instrument 5, without irradiating the reference-color portion with a light source for measurement, in a predetermined measurement environment; an objective-portion measurement step of obtaining an objective-portion measurement value by measuring a spectroscopic-radiation luminance of a light being emitted from a measurement-objective portion in the measurement direction, or a tristimulus value thereof, using the light-source-color measuring instrument 5, without irradiating the measurement-objective portion with a light source for measurement, in the measurement environment; and a color identification step of finding a color of the measurement-objective portion by means of computation from a ratio of the objective-portion measurement value with respect to the reference-color measurement value are equipped. Even when measuring a color of such a body, like a body including a fluorescent material, whose reflectivity has changed depending on the type of light source, it is possible to measure the color of such a body accurately. | en | COLOUR MEASURING METHOD FOR A BODY | 8602_JP | 69839656_JP | G01J 3/02,G01J 3/0267,G01J 3/027,G01J 3/0278,G01J 3/0289,G01J 3/0291,G01J 3/46,G01J 3/52,G01J 3/524,G01J2003/468 | [
"G01J 3/46"
] | 1,058 |
54,157,401 | 1992-07-30 | 26,506,696 | Y | This invention relates to an abnormal synchronization preventing device used in a multiple value quadrature modulation type radio device and the abnormal synchronization preventing device includes a monitoring circuit for monitoring the carrier pull-out based on demodulated base band signals to output a carrier pull-out alarm signal, an abnormal synchronization detection circuit for sampling the reception signal level at a speed twice the transmission speed of reception data according to identification data and detecting abnormal synchronization according to whether or not the detection area of sampled data lies in a specified area, an abnormal synchronization preventing circuit for generating a signal for a preset period of time when abnormal synchronization is detected in a case where the carrier pull-out alarm signal is not output, and a selection circuit for receiving a reference signal used for generating a reference carrier frequency signal having a frequency close to the frequency of the carrier wave and a phase difference signal, selecting the phase difference signal as the control signal in the normal state, and selects the reference signal as the control signal while receiving an output signal of the abnormal synchronization preventing circuit and supplying the selected signal to an oscillator. | en | Quadrature amplitude modulation type digital radio communication device and method for preventing abnormal synchronization in demodulation system | 5231517_JP | 8616186_JP,5477954_JP,10504740_JP,10504741_JP | H04L 27/3827 | [
"H04L 27/38"
] | 48,279 |
549,025,588 | 2020-07-10 | 74,202,276 | N | The invention provides a system of deep learning neural network in prostate cancer bone metastasis identification based on whole body bone scan images. A pre-processing module receives an input whole body bone scanning image and processes the whole body bone scanning image; a neural network module detects whether the input whole body bone scanning image is the osseous metastasis of the CAC or not. The neural network module comprises a thoracic cavity part network module and a pelvis part network module. The thoracic cavity part network module is used for establishing a first-stage accelerated regional convolutional neural network, and segmenting a training image of a thoracic cavity part according to a whole body bone scanning image; and training a second-stage accelerated regional convolutional neural network by using the training image, and classifying the focus of cancer cell bone metastasis. The pelvic part network module uses a convolutional neural network and comprises the following steps: establishing a first-stage accelerated region convolutional neural network, and segmenting a training image of a pelvic part according to an input whole body bone scanning image; and training a convolutional neural network by using the training image to classify whether the pelvic part is a bone metastasis image. | en | System of deep learning neural network in prostate cancer bone metastasis identification based on whole body bone scan images | 72560269_ | 79944410_,80145721_,58351712_,70246931_ | G06N 3/0454,G06N 3/08,G06T 7/0012,G06T2207/20081,G06T2207/20084,G06T2207/30008,G06T2207/30081 | [
"A61B 5/00",
"G06T 7/194",
"G06T 7/00",
"G06N 3/04",
"G06T 7/136"
] | 152,140 |
571,096,755 | 2021-04-25 | 74,926,841 | N | A sustained attention regulation system for joint electroencephalogram and finger pressure feedback. The system is used for solving the problem that a subject of electroencephalogram feedback is difficult to enter a sustained attention state without a task. The system comprises: a finger pressure measurement module (1), a finger pressure sampling and transmission module (2), and a finger pressure conversion and recording module (3) which are connected in sequence; an electroencephalogram acquisition module (4) for acquiring an electroencephalogram analog signal of a subject; an electroencephalogram storage module (5) for converting the acquired electroencephalogram analog signal into a digital signal and storing the digital signal; an electroencephalogram data analysis module (6) for completing the analysis and extraction of an electroencephalogram feature signal according to the acquired electroencephalogram signal; and a joint feedback presentation module (7) for presenting the electroencephalogram feature signal and a finger pressure signal by means of visualization graphics. By means of a joint electroencephalogram and finger pressure feedback mode, a subject can quickly enter and maintain a sustained attention state, such that the sustained attention state of the subject can be accurately estimated. | en | SUSTAINED ATTENTION REGULATION SYSTEM FOR JOINT ELECTROENCEPHALOGRAM AND FINGER PRESSURE FEEDBACK | 71073783_CN | 67718918_CN,84771915_CN,63693143_CN,69637155_CN | G06F 3/015,G06F 3/0414,G06F2203/011 | [
"G06F 3/01"
] | 165,282 |
54,028,791 | 1992-06-05 | 27,527,490 | Y | In an image processing apparatus for processing inputted binary image data including a pseudo half-tone area in which data are binarized by a pseudo half-tone binarizing method and a non-half-tone area in which data are binarized using a predetermined threshold value, an image area judgment section performs an image area judgment process for judging whether each pixel of the inputted binary image data is included in the pseudo half-tone area or the non-half-tone area, based on the binary image data located in a predetermined first area including a plurality of pixels which includes a specified pixel and a plurality of pixels located in the periphery of the specified pixel and which are selected among the inputted binary image data. A first judgment section judges what is the kind of minor pixels in the first area based on the binary image data located in the first area, and a second judgment section judges whether or not at least one of the minor pixels as judged is included in a predetermined second area which is located within the first area and which is smaller than said first area. The area judgment section invalidates result data of the image area judgment process with respect to the specified pixel when the second judgment section judges that none of the minor pixels is included in the second area. | en | Image processing apparatus comprising means for judging image area of pseudo half-tone image | 5235029_JP | 5940488_JP | H04N 1/40062 | [
"H04N 1/40"
] | 47,897 |
555,862,304 | 2020-11-30 | 73,665,904 | N | The present application relates to the technical field of artificial intelligence, and provides an idiom blank-filling question answer selection method and apparatus, a computer device, and a storage medium. The method comprises: obtaining a question text of an idiom blank-filling question, wherein the question text comprises a blank-filling text and n candidate idioms, and the blank-filling text comprises m blanks to be filled with the candidate idioms (S1); obtaining explanation texts of all the candidate idioms (S2); inputting the blank-filling text and the explanation texts into a pre-trained idiom selection blank-filling model, and obtaining a confidence degree for filling each candidate idiom into each blank (S3); selecting m idioms from the n candidate idioms to form multiple answer sets (S4); calculating the sum of the confidence degrees for filling the candidate idioms in each answer set into the blanks (S5); obtaining the answer set having the highest confidence degree sum, and outputting as an answer for the idiom blank-filling question (S6). The present solution uses artificial intelligence, and obtains idiom blank-filling answers with high efficiency and high accuracy. The present solution may also be applied to the field of smart education, so as to promote the construction of smart cities. | en | IDIOM BLANK-FILLING QUESTION ANSWER SELECTION METHOD AND APPARATUS, AND COMPUTER DEVICE | 63942312_CN | 67862676_CN,63569491_CN | G06F 16/27,G06F 21/602,G06F 21/64,G06F 40/242,G06F 40/289,G06F 40/30,G06N 3/0454,G06N 3/08,G06Q 50/20 | [
"G06F 40/289"
] | 156,769 |
561,961,649 | 2020-06-30 | 72,121,477 | N | The present invention provides a parameter monitoring device and system for proton therapy. The device comprises: a recurrent neural network module, providing a trained recurrent neural network model, wherein the trained recurrent neural network model provides a nonlinear relationship between positron nuclide activity distribution and dose distribution of a proton beam; a PET obtaining module, obtaining, from a positron annihilation tomography system, a PET image of activity distribution of positron nuclides produced by a reaction between protons and tissues; a prediction module, inputting the PET image into the trained recurrent neural network model, and predicting the dose distribution and a Bragg peak position of the proton beam by means of the nonlinear relationship between the positron nuclide activity distribution and the dose distribution; a determination module, determining a positional relationship between the predicted dose distribution and Bragg peak position of the proton beam, and a target region; and an adjustment module, determining, according to a determination result of the determination module, whether a beam outlet parameter of the proton beam needs to be adjusted. The device and system provided in the present invention achieve the parameter monitoring and adjustment of proton therapy. | en | PARAMETER MONITORING DEVICE AND SYSTEM FOR PROTON THERAPY | 82750356_CN | 82687731_CN,69636812_CN,78677331_CN | A61N 5/1071,A61N2005/1087,G06N 3/0454,G06N 3/08 | [
"G01T 1/29",
"A61N 5/10",
"G06N 3/08",
"G06N 3/04"
] | 160,709 |
36,115,520 | 1988-12-21 | 22,475,877 | N | PURPOSE: To minimize leaked electric charges from capacitors by charging each capacitor of an analog neuron circuit network up to a weighted voltage, making individual connection to gate electrode of FETs, and coupling the electrodes with the weighted voltage source periodically thereby providing a refresh electric charge. CONSTITUTION: A neuron circuit network 10 receives values X depending on the algorithm. In the learning stage, one set of input values xi are fed sequentially in parallel with the neuron circuit network 10 and a pre-processor 11. Furthermore, one set of target values dj (j=1, m) are fed to a microprocessor 13 in the order of patterns with the output of the neuron circuit network 10 converted into a digital signal by the pre-processor 22. Weight information is stored in a RAM 15 and read under the control of a recycle enable program counter 16 and fed sequentially to a capacitor array of the neuron circuit network 10 via a D/A converter 17. A counter 16 advances the processing of a decoder 19 simultaneously and energizes sequentially a switch array of FETs of the neuron circuit network 10 via a line S1 to control provision of weighted voltage to each capacitor of the capacitor array. A delay device 21 adjusts the D/A converter output before the decoder gives a strobe signal to the FETs. | en | ARTIFICIAL NERVE CIRCUIT NETWORK | 30340753_ | 32069454_ | G06N 3/063,G06N 3/0635,G06V 10/20 | [
"G06N 3/06",
"G06N 99/00",
"G06F 15/18",
"G06N 3/063"
] | 25,041 |
551,115,023 | 2020-11-13 | 69,907,853 | N | The present application is applicable to the technical field of image recognition, and provided are a motion posture recognition method, a motion posture recognition apparatus, a terminal device and a medium. In the motion posture recognition method, by means of a trained dual-stream long and short-term video posture estimation model, motion posture recognition is performed on a video image to be recognized that comprises a moving body in a target period; by means of a dual-stream 3D convolutional neural network in the trained dual-stream long and short-term video posture estimation model, comprehensive features of the moving body are extracted from the video image; and then target pre-estimated posture information of the moving body in the target period is obtained by means of a recurrent neural network on the basis of first posture pre-estimated information of the moving body in a first period and the comprehensive features of the moving body, and an obtained pre-estimated posture represented by the target pre-estimated posture information of the moving body in the target period has time sequence continuity, so that the Euclidean distance between the pre-estimated posture and a preset reference posture is more accurate, thereby improving the efficiency of motion posture recognition of the moving body. | en | MOTION POSTURE RECOGNITION METHOD, MOTION POSTURE RECOGNITION APPARATUS, TERMINAL DEVICE AND MEDIUM | 66824291_CN | 80684100_CN,71269406_CN,63583101_CN | G06N 3/0445,G06N 3/0454,G06V 40/23 | [
"G06K 9/00"
] | 153,632 |
544,368,821 | 2020-06-22 | 74,114,558 | N | Provided are a reading comprehension assistance system and a reading comprehension assistance method for enabling input of natural language as a query sentence and presenting, to a reader, a location highly relevant to the input sentence. The reading comprehension assistance system includes: a document reading unit that reads a subject document; a document dividing unit that divides the subject document into a plurality of blocks; a first distributed representation acquisition unit that acquires distributed representations of words for each of the plurality of blocks; a query sentence reading unit that reads a query sentence; a second distributed representation acquisition unit that extracts words included in the query sentence and acquires distributed representations of the words; and a similarity degree acquisition unit that compares the distributed representations of the words between the query sentence and each of the plurality of blocks, and derives a similarity degree. The similarity degree acquisition unit: searches the words included in the blocks for a word that matches a word included in the query sentence; and for a matching word, derives a similarity degree between the distributed representations of the word in the blocks and the distributed representations of the word in the query sentence. | en | READING COMPREHENSION ASSISTANCE SYSTEM AND READING COMPREHENSION ASSISTANCE METHOD | 64639326_JP | 73532457_JP,71281249_JP,76339722_JP | G06F 16/332,G06F 16/3344,G06F 40/189,G06F 40/211,G06F 40/284,G06F 40/44,G06K 9/6215 | [
"G06F 40/44",
"G06F 16/332"
] | 148,997 |
16,730,323 | 1988-07-25 | 22,139,746 | N | There is disclosed apparatus and methods for sensing electrical signals generated in a living body and mapping these signals to MIDI code. The apparatus senses electrical signals generated in the body through electrodes (12, 14, 16) on the skin or implanted in various areas of the body. The electrical signals are then amplified (24), filtered (26) and converted (28) to digital data. This digital data is analyzed to determine characteristics of the electrical signals such as amplitude, frequency, change in amplitude or change in frequency. These characteristics are then mapped (36) to MIDI code or other code to control the generation of musical (38) notes or video (40) display information. There is also disclosed a method and apparatus for determining the alpha wave or other content of brain waves by taking a fast fourier transform of the brain wave and determining the coefficient for the fourier frequency component at the frequency of desired frequencies. Further, a method of altering MIDI code generated from biopotentials is disclosed where the effect of musical notes generated by the MIDI code on the biopotentials causes changes in the MIDI code. A similar method is also disclosed wherein MIDI code received from another machine is changed based upon mapping of biopotentials from the body of a subject. | en | BIOPOTENTIAL DIGITAL CONTROLLER FOR MUSIC AND VIDEO APPLICATIONS | 2229713_US | 2730612_US,2730613_US | A61B 5/316,A61B 5/374,A61B 5/38,A61B 5/398,A61B 5/7225,A61B 5/7257,G10H 1/02,G10H2220/135,G10H2220/376 | [
"A61B 5/374",
"G10H 1/02",
"G10H 1/00",
"G09G 5/36"
] | 17,698 |
515,738,792 | 2018-01-04 | 67,063,917 | N | Disclosed is a plush doll robot with a voice recognition function. A plush doll robot with a voice recognition function, according to one embodiment of the present invention, comprises: a control unit for performing an action corresponding to a command input by a user; a voice recognition unit for storing a user's voice as an input voice file, transmitting the input voice file to the control unit when a predetermined command is included in the input voice, and transmitting the input voice file to a voice recognition server when a predetermined command is not included in the input voice; a voice providing unit for receiving a voice response file corresponding to the input voice file from the voice recognition server and outputting the voice response file as a voice; a sensor unit, provided with at least one sensor of a touch sensor and a pulse sensor, for sensing a body input; a wireless communication unit for wirelessly communicating with an external terminal by using at least one of Wi-Fi, Bluetooth, and NFC; a motor unit for driving a plurality of motors to control movement of the body of the plush doll robot; and an LED for, when an LED lighting signal is included in signals received from the sensor unit and the voice recognition server, illuminating light in response to the received lighting signal. | en | PLUSH DOLL ROBOT WITH VOICE RECOGNITION FUNCTION | 67013284_KR | 64190221_KR | A63H 3/00,A63H 3/003,A63H 3/02,A63H 3/28,A63H 30/04,A63H2200/00,G10L 13/08,G10L 15/22,G10L 15/30,H04W 4/00 | [
"G10L 13/08",
"A63H 30/04",
"A63H 3/00",
"A63H 3/28",
"A63H 3/02",
"H04W 4/00",
"G10L 15/30",
"G10L 15/22"
] | 129,435 |
341,569,965 | 2011-04-04 | 43,939,129 | N | Disclosed is an apparatus and method for generating a cognitive training application program using brain waves, and a medium for recording same. The apparatus for generating a cognitive training application program using brain waves according to an embodiment of the present invention comprises: a BCI template creator which generates a BCI template describing a method for controlling a BCI apparatus and measuring brain waves, a neurofeedback template describing the kind of brain waves to be measured and the method for processing same for each brain symptom, and a cognitive training template describing the method by which a game proceeds according to the result of measuring the brain waves for each BCI game and then transmitting them to a template database; a BCI game engine which includes a BCI library consisting of functions for the BCI template, a neurofeedback library consisting of functions for the neurofeedback template, and a cognitive training library consisting of functions for the cognitive training template; and a code generator for the BCI application which receives one or more templates from the template database, reads the BCI library, neurofeedback library, and cognitive training library of the BCI game engine, and then generates the BCI application code according to the received template. | en | APPARATUS AND METHOD FOR GENERATING A COGNITIVE TRAINING APPLICATION PROGRAM USING BRAIN WAVES, AND MEDIUM FOR RECORDING SAME | 43458588_KR,41616282_KR,40835642_KR,10909276_KR | 43458588_KR,40835642_KR,41616282_KR | G06F 3/015,G06F 8/20 | [
"G06F 9/44"
] | 72,317 |
555,667,541 | 2021-07-06 | 77,236,388 | N | According to the method, a time sequence deep convolutional neural network (TCN), a self-attention Transformer encoder and a multi-modal multi-head attention mechanism are adopted, and the invention relates to a model and a recognition method for estimating continuous-dimension emotions from multi-mode (auditory and visual) time sequence information. The method comprises the following steps: obtaining embedded feature expressions of different modalities for features input by different modalities; then, using the embedded feature expressions of different modes as input, and obtaining advanced representations of different modes by using a multi-mode Transform encoder; and finally, taking the high-level feature expressions of different modes as input, and calculating an emotional state value at each moment. According to the method, more attention is paid to the influence of some key moments on the current emotional state in the past in time sequence, interference caused by long-term emotional information is eliminated, and the robustness of the model is improved. Meanwhile, the accuracy of continuous dimension emotion estimation is obviously improved through a method of simultaneously correcting a time sequence context dependency relationship and a multi-modal interaction fusion relationship in the model. | en | Continuous dimension emotion recognition method based on Transform encoder and multi-head multi-modal attention | 59751398_ | 60806212_,58909824_ | G06K 9/629,G06N 3/0454,G06N 3/0481,G06N 3/08 | [
"G06N 3/04",
"G06K 9/62",
"G06N 3/08"
] | 156,542 |
471,742,160 | 2015-05-29 | 54,572,994 | N | A digital phase shift point diffraction interferometer. The interferometer consists of a light source (1), a small-hole mask (2), a first spatial light modulator (3), a second spatial light modulator (6), a two-dimension photoelectric detector (8) and a computer (9). The first spatial light modulator (3) is configured as a raster by means of the computer (9), and serves as a splitter. The second spatial light modulator (6) is configured as a pin hole window mask, and serves as a filter for filtering diffraction orders except 0 order and +1 (or -1) order, so as to enable 0-order light to achieve diffraction by means of a pin hole (61) to generate a quasi-ideal spherical wave serving as a reference light wave, and +1 (or -1) order light serves as an object light wave by means of a window (62). The two light waves interfere, an interferogram is acquired, and an optical system wave aberration to be measured is extracted from the interferogram. Also provided is a method for measuring a wave aberration of an optical system using a digital phase shift point diffraction interferometer. A distance between an object light wave and a reference light wave at image plane convergent points is adjustable, and therefore great interference fringe density is achieved without reduction of an interference fringe contrast. | en | DIGITAL PHASE SHIFT POINT DIFFRACTION INTERFEROMETER AND OPTICAL SYSTEM WAVE ABERRATION MEASURING METHOD | 63935712_CN | 63675570_CN,63730337_CN,64073549_CN | G01J 9/02,G01M 11/02 | [
"G01J 9/02",
"G01M 11/02"
] | 104,987 |
564,304,965 | 2021-10-15 | 79,330,664 | N | The invention belongs to the technical field of modern industrial intelligent control, and particularly discloses a self-adaptive torque control method for bilateral pose positioning of a multi-station mechanical arm. The problem of multi-station mechanical arm pose operation control excessively depending on visual sensor positioning in industrial flow line production is solved. The method comprises the following steps: firstly, system modeling is carried out, and a sliding mode integral term is embedded in a self-adaptive feedback control protocol of a controlled system to carry out closed-loop torque controller design; and a loop-free topological structure integrally formed by an assembly line system is utilized, initial state information is adjusted in advance, a multi-station mechanical arm is loaded into a designed controller algorithm, then the controller algorithm is evolved to expected multiple sets of bilateral positions, and the method for mechanical arm pose positioning through real-time monitoring of a visual sensor is improved. By means of the method, precise operation of mechanical arm pose control can be conveniently achieved, the automation, intellectualization and integration levels of modern industrial production are improved, and good social benefits and economic values are achieved. | en | Self-adaptive torque control method for bilateral pose positioning of multi-station mechanical arm | 62887250_,65532067_ | 80207559_,65890174_,59377163_,67130013_,66007755_ | B25J 9/16,B25J 9/1664 | [
"B25J 9/16"
] | 162,252 |
560,316,996 | 2020-08-25 | 72,452,275 | N | Disclosed are an emotion information recognition method and apparatus, and a storage medium and a computer device, which relate to the technical field of artificial intelligence, and aim to respectively process a human body posture and facial expression, which have been obtained, into a posture matrix and emotion vector information, and to simultaneously process, by means of a pre-trained emotion intensity recognition model, the posture matrix and the vector information to obtain emotion intensity data and correspondingly recognize an emotion type, thereby improving the accuracy and efficiency of emotion information recognition. The method comprises: receiving an emotion information recognition request, wherein the emotion information recognition request carries human body posture information; transforming the human body posture information into a posture matrix including posture feature points by using a preset posture transformation algorithm; processing the posture matrix according to a preset emotion intensity algorithm, so as to obtain emotion intensity data; and retrieving and feeding back a corresponding emotion type according to the emotion intensity data. In addition, the present application further relates to blockchain technology, and the emotion intensity data may be stored in a blockchain. | en | EMOTION INFORMATION RECOGNITION METHOD AND APPARATUS, AND STORAGE MEDIUM AND COMPUTER DEVICE | 63942312_CN | 68135266_CN,77009095_CN,69072352_CN,82987545_CN | G06K 9/6267,G06N 3/0454,G06N 3/0481,G06N 3/049,G06N 3/08,G06V 20/41,G06V 40/174,G06V 40/20 | [
"G06K 9/00"
] | 159,614 |
517,945,761 | 2019-05-12 | 67,573,440 | N | The invention provides an active rehabilitation training system combined with transcranical current stimulation. The system comprises an EEG acquisition module, a transcranial current stimulating module, an electrode module, an affected limb feedback module and a middle control analysis module. According to brain-machine interface technology, transcranial current stimulation technology and a movement imagination therapy principle, the system analyzes and evaluates the movement imagination intention of a hemiplegia patient based on EEG data, and performs corresponding feedback to the patient. Furthermore, by means of transcranical current stimulation, simulation of a target brain region is performed before movement imagination, thereby finishing intervention of an unaffected brain functionregion or activation of an affected brain functional region. Furthermore corresponding anode stimulation is applied to the affected brain region in the affected limb feedback process. The system and the method can effectively improve movement imagination accuracy of the patient and effectiveness of a feedback effect, thereby realizing regulation and reconstruction of the brain function, promotingrestoring or reconstruction of a damaged movement conduction path, and performing a movement rehabilitation treatment effect. | en | Active rehabilitation training system combined with transcranial current stimulation | 61622624_ | 61213508_,58935659_,64916219_ | G16H 20/30 | [
"G16H 20/30"
] | 130,779 |
478,336,319 | 2016-07-05 | 55,507,701 | N | The invention discloses a shifted sub-carrier-based frequency domain chaotic cognitive radio system comprising a transmitter and a receiver. At the transmitter, a frequency domain chaotic signal generator is configured to generate a frequency domain-based chaotic signal c(t), a sub-carrier correspondence module is configured to dynamically change a corresponding relationship between a chaotic sequence C(k) and a sub-carrier, data information undergoes a binary phase shift keying modulation to provide b(t), b(t) and c(t) are modulated together to provide s(t), and s(t) is further modulated with a carrier and transmitted over a radio channel in an available frequency band. At the receiver, after coherent carrier demodulation, a frequency domain chaotic signal generator is configured to generate a frequency domain-based chaotic signal c(t), the chaotic signal c(t) is then synchronized and used to demodulate a receiving signal r(t) to obtain d(t), a zero-crossing determination is performed on d(t) to obtain an estimate b̑(t) of an information bit, thereby successfully receiving the transmitted information. The shifted sub-carrier-based frequency domain chaotic cognitive radio system disclosed by the invention significantly improves security of a conventional frequency domain chaotic cognitive radio system. | en | SHIFTED SUB-CARRIER-BASED FREQUENCY DOMAIN CHAOTIC COGNITIVE RADIO SYSTEM | 63752135_CN | 63617602_CN,68099369_CN | H04L 9/00,H04L 9/001 | [
"H04L 9/00"
] | 108,714 |
527,921,307 | 2019-05-17 | 65,841,318 | N | The disclosure belongs to the technical field of lower limb exoskeleton, and specifically discloses a hip-knee passive exoskeleton device based on clutch time-sharing control, comprising a waist support subassembly, connection subassemblies, thigh subassemblies, clutch subassemblies, shank subassemblies and elastic member subassemblies, the waist support subassembly is configured to be connected to the waist, the connection subassemblies are configured to include two connection subassemblies which are arranged in bilateral symmetry on two sides of the support subassembly, the thigh subassemblies are configured to include two thigh subassemblies which are respectively connected to the two connection subassemblies, the clutch subassemblies are configured to include two clutch subassemblies which are respectively mounted on the two thigh subassemblies, the shank subassemblies are configured to include two shank subassemblies which are arranged in bilateral symmetry below the two thigh subassemblies, the elastic member subassemblies are configured to include two elastic member subassemblies which are arranged in bilateral symmetry. The disclosure can assist the movements of the knee and hip joints, thereby improving the energy utilization efficiency and reducing the metabolic energy consumption of walking. | en | HIP-KNEE PASSIVE EXOSKELETON DEVICE BASED ON CLUTCH TIME-SHARING CONTROL | 9670505_CN | 72249036_CN,74789523_CN,74711747_CN,74628856_CN,57901069_CN | A61H 1/00,A61H 1/0262,A61H 3/00,A61H2003/007,A61H2201/0192,A61H2201/1253,A61H2201/1445,A61H2201/163,A61H2201/1642,A61H2201/165,A61H2201/1652,A61H2201/5053,A61H2205/10 | [
"A61H 3/00",
"A61H 1/02"
] | 138,103 |
474,612,059 | 2016-08-19 | 57,878,550 | Y | The invention discloses a variation function model optimization method in forest site index spatial-temporal estimation. The method comprises the following steps of (1) establishing a reliable site index variation function model optimization method by utilizing a computer, thereby ensuring effective combination of fixation and quantification of variation function selection; (2) establishing a unified site index variation function multi-scale nested model expression form and a computer automatic parameter optimal fitting method by utilizing the computer, and realizing an expansion algorithm of a multi-scale nested model function in mainstream statistic software, thereby ensuring effective prediction of a spatial interpolation algorithm; and (3) establishing a site index spatial-temporal variation function model expression form by utilizing the computer, and realizing a Kriging spatial-temporal interpolation algorithm, thereby improving the precision and reliability of the spatial-temporal estimation. The artificial subjective factors can be effectively reduced; the site index anisotropy and multi-scale dependence can be effectively described; the effective prediction of the spatial interpolation algorithm can be ensured; and the precision and reliability of the spatial-temporal estimation are improved. | en | Variation function model optimization method in forest site index spatial-temporal estimation | 65043498_ | 59568351_,69335572_,60131974_,65182161_,60619254_,67490186_,69299556_,58541030_,63371086_ | G06F 30/20 | [
"G06F 111/10",
"G06F 30/20",
"G06F 17/50",
"G06F 119/02"
] | 106,375 |
51,859,228 | 1990-08-07 | 9,384,525 | Y | Compound of general formula (I): <IMAGE> (I) in which: X, Y, Z and T each represent, independently of one another, a nitrogen atom, a -CH= group or a group CW, in which W represents a halogen atom or a lower alkyl or alkoxy group optionally substituted with one or more halogen atoms, such as a trifluoromethyl group, on condition that only one of them represents a nitrogen atom, Ra, Rb, Rc, Rd and Re represent, independently of one another, a hydrogen atom, a halogen atom or a lower alkyl or alkoxy group optionally substituted with one or more halogen atoms, such as a trifluoromethyl group, B represents a hydrogen atom and D a hydroxyl group, or B and D simultaneously represent an oxygen atom, n represents 1 or 2, and R1 represents a hydrogen atom, a linear or branched lower alkyl, linear or branched lower alkenyl, aryl or lower arylalkyl group or a cycloalkyl group having 3 to 7 carbon atoms, it being possible for each of the lower alkyl, lower alkenyl, aryl, lower arylalkyl and cycloalkyl groups to be substituted with one or more halogen atoms or a lower alkyl or alkoxy group optionally substituted with one or more halogen atoms, such as a trifluoromethyl group, their isomers, epimers and diastereoisomers, and addition salts with a pharmaceutically acceptable acid are useful in the treatment of pain. | en | Analgesic oxazolopyridine compounds | 8867508_FR | 5574409_FR,6276754_FR,8867509_FR | A61P 25/04,C07D 498/04 | [
"A61K 31/53",
"C07D 498/04",
"A61K 31/50",
"A61P 25/04",
"A61K 31/505",
"A61K 31/44"
] | 42,935 |
569,286,922 | 2021-08-18 | 74,855,304 | N | Disclosed is an end-to-end speckle projection three-dimensional measurement method based on deep learning. The method comprises: first using a projector for projection and a binocular camera for synchronously collecting a speckle pattern; performing stereo correction on the speckle pattern, inputting the pattern into a stereo matching network, and a feature extraction sub-network based on a shared weight processing the speckle pattern obtain a series of low-resolution three-dimensional feature tensors; inputting the feature tensors into a salient object detection sub-network for detecting foreground information in the speckle pattern to generate a full-resolution effective mask; generating a four-dimensional matching cost body by means of the feature tensors of two viewing angles according to a candidate disparity range, achieving cost aggregation by means of a series of three-dimensional convolutional layer filtering, and obtaining an initial disparity map by means of disparity regression; and obtaining a final disparity map by combining the effective mask and the initial disparity map to achieve single-frame robust absolute three-dimensional surface contouring. The present invention can achieve single-frame robust absolute three-dimensional surface contouring simply by projecting a speckle pattern. | en | END-TO-END SPECKLE PROJECTION THREE-DIMENSIONAL MEASUREMENT METHOD BASED ON DEEP LEARNING | 63938471_CN | 67325776_CN,67175757_CN,68351136_CN,81389251_CN,64230496_CN,81192646_CN,63724780_CN | G06N 3/0454,G06T 7/55 | [
"G06N 3/04",
"G06T 7/55"
] | 164,443 |
480,010,608 | 2016-11-17 | 59,017,327 | Y | PROBLEM TO BE SOLVED: To reduce a user's discomfort due to flicker due to lighting and extinction of the light source.SOLUTION: An image reading apparatus comprises: a plurality of light sources capable of applying light from different irradiation positions predetermined with respect to an object to be read; illumination control means for controlling so that a plurality of light sources are sequentially and alternately turned on and off sequentially at intervals at which blinking cannot be recognized by human eyes; an imaging unit that photoelectrically converts the reflected light of the light irradiated by the plurality of light sources sequentially and alternately to the object to be read for each pixel and captures a plurality of read images; a storage unit that stores results of imaging by the imaging unit as one or more read images when at least one of the plurality of light sources applies light to the reading target; and an image combining unit using a plurality of light sources to sequentially and alternately apply light to the object to be read so as to combine each predetermined setting area when the change in the image level due to the specular reflection light of each of the plurality of read images taken by the imaging unit is small, to combine the read images.SELECTED DRAWING: Figure 4 | en | IMAGE READING APPARATUS, IMAGE FORMING APPARATUS, AND IMAGE READING METHOD | 63139233_ | 60539572_,58178794_,59737118_,69407441_,58634888_,66111991_,60175625_,58934657_,59207785_,62215688_,59288636_,65074203_,66697042_ | G03G 15/011,G03G 15/043,H04N 1/00766,H04N 1/04,H04N 1/19594 | [
"H04N 1/10",
"H04N 5/238",
"H04N 5/225",
"H04N 1/04",
"H04N 1/387",
"G06T 1/00",
"H04N 1/107",
"G03B 27/54"
] | 109,534 |
52,911,747 | 2000-11-07 | 25,025,744 | Y | The invention features a method for identifying a cDNA nucleic acid encoding a mammalian protein having a signal sequence, which method includes the following steps: (a) providing library of mammalian cDNA; (b) ligating the library of mammalian cDNA to DNA encoding alkaline phosphatase lacking both a signal sequence and a membrane anchor sequence to form ligated DNA; (c) transforming bacterial cells with the ligated DNA to create a bacterial cell clone library; (d) isolating DNA comprising the mammalian cDNA from at least one clone in the bacterial cell clone library; (e) separately transfecting DNA isolated from clones in step (d) into mammalian cells which do not express alkaline phosphatase to create a mammalian cell clone library wherein each clone in the mammalian cell clone library corresponds to a clone in the bacterial cell clone library; (f) identifying a clone in the mammalian cell clone library which expresses alkaline phosphatase; (g) identifying the clone in the bacterial cell clone library corresponding to the clone in the mammalian cell clone library identified in step (f); and (h) isolating and sequencing a portion of the mammalian cDNA present in the bacterial cell library clone identified in step (g) to identify a mammalian cDNA encoding a mammalian protein having a signal sequence. | en | Nucleic acid molecules encoding a secreted neural adhesion protein | 5233538_US | 5566930_US,9612152_US,9612153_US | C12N 15/1034,C12N 15/1051,C12N 15/85,C12N2800/60,C12N2830/38,C12Q 1/6811 | [
"C12N 15/85",
"C12N 15/10",
"C12Q 1/68"
] | 44,912 |
509,880,901 | 2018-11-08 | 65,802,239 | Y | To provide a weak light detection system and weak light detection method that enable a simple configuration to detect weak light of an object at a high S/N ratio.SOLUTION: A control unit 15 of a weak light detection system is configured to apply an encoding system to a prescribed transmission information series by an error correction code generation unit 151, and make a generated error correction code series modulate excitation light with which an excitation light irradiation unit 171 irradiates a test piece. A light reception code series identification unit (an A/D converter 153) is configured to identify a light reception code series on the basis of a change in light reception intensity in a light reception unit 172 receiving light to be excited in the test piece using an excitation light irradiation pattern. A signal processing unit 154 is configured to obtain a decode information series by applying a decoding system of the error correction code to the light reception code series, collate the transmission information series with the decode information series, and when the transmission information series is consistent with the decode information series, determine that light emission corresponding to the excitation light irradiation pattern is performed from the test piece.SELECTED DRAWING: Figure 3 | en | WEAK LIGHT DETECTION SYSTEM, AND WEAK LIGHT DETECTION METHOD | 75271177_ | 60721977_,59641420_ | G01N 21/6428,G01N 21/645,G01N 21/8483,G01N2021/6439,G01N2201/062,G01N2201/0625 | [
"G01N 21/78",
"G01N 21/64"
] | 126,280 |
323,567,022 | 2010-03-16 | 42,739,463 | Y | The present invention provides a system for analyzing an expression profile in which a large number of expression profile data obtained by a next-generation high-speed sequencer, a similar experimental technique, or the like is analyzed at high speed by a general-purpose computer, gene expression patterns are visualized, thereby easily analyzing to which gene a novel gene is similar in function. A system for analyzing an expression profile which analyzes gene expression profile data includes a storage unit which stores the number of counts of mRNAs expressed from a subject gene to be evaluated as expression data under each of a plurality of gene expression conditions for each subject gene name, a correspondence analysis unit which reads the expression data from the storage unit for each subject gene, and carries out correspondence analysis on the basis of the number of counts under each expression condition in expression data, a coordinate conversion unit which converts n-dimensional (n: natural number) scores obtained by the correspondence analysis to coordinate values for m-dimensionally (m: natural number, m¤n) arranging each subject gene, and an image processing unit which carries out plotting along the corresponding coordinate values for each gene to display the result on an image display unit. | en | SYSTEM AND PROGRAM FOR ANALYZING EXPRESSION PROFILE | 4751814_JP,4751815_JP | 49629492_JP,49747284_JP | G16B 25/00,G16B 45/00 | [
"G06F 19/00",
"G16B 45/00"
] | 63,877 |
421,937 | 2000-03-06 | 23,022,044 | Y | A method of separating high-level prosodic behavior from purely articulatory constraints so that timing information can be extracted from human speech is presented. The extracted timing information is used to construct duration templates that are employed for speech synthesis. The duration templates are constructed so that words exhibiting the same stress pattern will be assigned the same duration template. Initially, the words of input text segmented into phonemes and syllables, and the associated stress pattern is assigned. The stress assigned words are then assigned grouping features by a text grouping module. A phoneme cluster module groups the phonemes into phoneme pairs and single phonemes. A static duration associated with each phoneme pair and single phoneme is retrieved from a global static table. A normalization module generates a normalized syllable duration value based upon the retrieved static durations associated with the phonemes that comprise the syllable. The normalized syllable duration value is stored in a duration template based upon the grouping features associated with that syllable. To produce natural human-sounding prosody in synthesized speech, the duration information is then extracted from the selected template, de-normalized and applied to the phonemic information. <IMAGE> | en | Generation and synthesis of prosody templates | 732933_JP | 732935_US,732934_US | G10L 13/08,G10L 13/10 | [
"G10L 13/08"
] | 2,400 |
37,429,378 | 2001-08-21 | 26,618,734 | N | PROBLEM TO BE SOLVED: To reduce unpleasant feelings felt by an operator even in the case of judging that voice recognition is impossible successively over two or more times. SOLUTION: A voice recognition propriety judgment part 21 of a voice recognition device 11 judges whether or not performance of a voice recognition processing is possible. When it is judged that the performance of the voice recognition processing is impossible, a recognition impossible time response change performance part 22 outputs a response message made different corresponding to the number of times of impossibility judgment and factors of the impossibility judgment, etc., from a speaker 13. When it is judged that the performance of the voice recognition processing is possible, a recognition processing part 23 performs the voice recognition processing to voice input from a microphone 12 and outputs a recognized result to other control units. A cancellation time change performance part 24 outputs the response message made different corresponding to the number of times of cancellation from the speaker 13 at the time of abandoning the recognized result at present and newly performing a recognition processing in the case that the recognized result in the recognition processing part 23 is different from the voice input or the like. | en | ON-VEHICLE VOICE RECOGNITION DEVICE | 10704891_ | 32120443_ | G10L 15/20,G10L2015/228 | [
"G10L 15/00",
"G10L 15/28",
"G10L 15/08",
"B60R 16/02",
"G10L 15/20",
"G10L 15/22",
"G10L 13/00"
] | 25,176 |
49,610,563 | 1999-06-01 | 26,435,085 | Y | A tactile sensor signal processing device includes a tactile sensor using a piezoelectric vibrator, a resonant resistance change detect unit, a resonant frequency change detect unit, a vibrator constant storage unit, and a signal processing unit. The resonant resistance change detect unit detects a change in resonant resistance of impedance characteristics of the piezoelectric vibrator. The resonant frequency change detect unit detects a change in resonant frequency of impedance characteristics of the piezoelectric vibrator. In the vibrator constant storage unit are stored in advance physical constants that include the size of the vibrating portion of the piezoelectric vibrator, piezoelectric and coupling constants associated with the piezoelectric mode of the piezoelectric vibrator, the density of the piezoelectric vibrator, and the size of a contact which is brought into contact with a subject to apply vibrations to it. The signal processing unit computes independently each of the real part and the imaginary part of shear elastic modulus indicating the viscoelastic characteristics of the subject on the basis of the results of detection by the resonant resistance change detect unit and the resonant frequency change detect unit and the physical constants stored in the vibrator constant storage unit. | en | Tactile sensor signal processing device capable of obtaining detailed living body information in short time | 5238396_JP | 7555621_JP | A61B 5/0051 | [
"A61B 5/00",
"A61B 1/00",
"A61B 5/103"
] | 39,057 |
494,667,310 | 2017-11-16 | 57,959,245 | N | Disclosed is an enhanced intelligent bionically-assisted mechanical leg for simulating a neural electrical signal, comprising a leg strap (1), a support mechanism (2), a power mechanism (3), a force applying column (4) and an assisting plate (5). The leg strap (1) is arranged at a shank portion. An inner side of the leg strap (1) is provided with a plurality of electromyographic sensors (11), and outer side of the leg strap (1) is provided with an acceleration sensor (12) and a microcontroller (13). The support mechanism (2) comprises a support plate (23) and an annular plate (22) fixedly arranged on the support plate (23). The power mechanism (3) is arranged on the support mechanism (2) and comprises an electric motor (31) arranged on the annular plate (22), a lead screw (32), a damping spring (33) and an arc-shaped plate (34). The force applying column (4) is hinged to the power mechanism (3) and the support mechanism (2). One end of the assisting plate (5) is hinged to the support mechanism (2). The present intelligent bionically-assisted mechanical leg is reasonable in structure, and has the advantages of a simple structure, convenient use, safety and reliability, and a high degree of intelligence and the like, thereby effectively solving the problem that only a few leg assisting machines exist. | en | ENHANCED INTELLIGENT BIONICALLY-ASSISTED MECHANICAL LEG FOR SIMULATING NEURAL ELECTRICAL SIGNAL | 64290412_CN | 64290412_CN | A61B 5/389,A61B 5/6828,A61B 5/6831,A61H 1/0237,A61H 3/00,A61H2003/005,A61H2003/007,A61H2201/1207,A61H2201/14,A61H2201/1642,A61H2201/165,A61H2201/1657,A61H2201/50,A61H2205/10,A61H2230/085 | [
"A61H 3/00"
] | 117,852 |
542,298,205 | 2019-09-26 | 73,500,863 | N | Disclosed in embodiments of the present invention are a personnel state detection method and apparatus based on eyelid feature information. The method comprises: obtaining a face image containing a face of a target person; detecting two-dimensional position information of a face feature point from the face image; constructing a target three-dimensional face model based on a preset three-dimensional face model; determining projection position information of an eyelid space point at a specified position in the face image based on a projection matrix; determining the current opening and closing length between upper and lower eyelids based on the two-dimensional position information of the eyelid feature point at a position corresponding to the specified position in the face image, three-dimensional position information of the eyelid space point at the specified position, and projection position information; and determining the current state of the target person based on the current opening and closing length, so as to determine the eyelid feature information of the human eyes. The accuracy of the detection result of the opening and closing state of the human eyes is improved by using the eyelid feature information, and the accuracy of the detection result of the current state of the person is improved. | en | PERSONNEL STATE DETECTION METHOD AND APPARATUS BASED ON EYELID FEATURE INFORMATION | 74289718_CN | 76124554_CN,79124455_CN,63499320_CN | G06K 9/00,G06V 40/193,G06V 40/197 | [
"G06K 9/00"
] | 147,519 |
352,171,753 | 2011-08-25 | 45,810,543 | N | To provide a dementia diagnosis support device which is capable of performing brain image analysis by using brain images that do not clearly show specific regions such as the parahippocampal gyrus or inferior horn of lateral ventricle, etc. and CT images that are clinically widely used, and which is capable of effectively supporting diagnosis of the type of dementia, the dementia diagnosis support device (100) extracts pixels which fall in a specific concentration range from an input brain image as a concentration range of interest, and represents the distribution thereof by graphics, a graph, or numerical values. When represented by graphics, the brain region is divided into a plurality of divided regions on the basis of the left-right brain symmetry, and for each divided region, a graph is displayed indicating the size of the concentration range of interest. When represented in a graph, with respect to points on the radius vector with the origin being the centroid of the brain region, a profile which shows the relationship between the distance thereof from the origin and the concentration value is obtained for all angles and normalized to produce a graph. When represented by numerical values, feature values which indicate the features of the concentration distribution are calculated and displayed. | en | DEMENTIA DIAGNOSIS SUPPORT DEVICE AND DEMENTIA DIAGNOSIS SUPPORT METHOD | 5726660_JP,22382215_JP | 22382215_JP | A61B 5/4088,A61B 6/032,A61B 6/463,A61B 6/501,G06T 7/0012,G06T2207/10072,G06T2207/30016 | [
"A61B 5/00",
"A61B 6/03"
] | 72,682 |
533,609,844 | 2020-04-23 | 71,181,062 | Y | The invention discloses a polarized light inertia rigorous integrated navigation method based on self-learning multi-rate residual correction. When a polarized light system works normally, a course angle output by an inertial navigation system is converted into an azimuth angle of a carrier body axis relative to the solar meridian and serves as a state quantity, the azimuth angle, calculated by the polarized light navigation system, of a carrier body axis relative to the solar meridian serves as an observed quantity, and CKF-ERC is used for fusing two data; the time information and the z-axisangular rate are taken as input of a long-short-term memory neural network, data resolved by a polarized light navigation system is taken as network output, and the network output is trained to realize a self-learning function; and when the polarized light system is not available, an error compensation stage is started, the data of the polarized light is predicted through LSTM, and data fusion isperformed on the predicted value and the azimuth angle calculated by inertia, thereby improving the long-time navigation precision of a downloading body in a complex environment. The autonomy of navigation orientation can be effectively improved, and then the intelligent degree of an unmanned aerial vehicle is improved. | en | Polarized light inertia rigorous integrated navigation method based on self-learning multi-rate residual correction | 65304787_ | 59391437_,60439946_,61393425_,78453931_,63237775_,66625429_,71263683_,65951396_,58301826_ | G01C 21/165,G01C 21/20,G06N 3/0445,G06N 3/0454 | [
"G01C 21/20",
"G06N 3/04",
"G01C 21/16"
] | 142,092 |
4,583,768 | 1998-04-22 | 25,313,166 | Y | A fast position tolerant and economical fingerprint verification algorithm and software implementation is disclosed. The algorithm has very important advantages in comparison to existing fingerprint verification algorithms like minutia extraction or full image analysis based on series of two dimensional fast Fourier transforms (FFT's) and two dimensional cross correlation in speed, compactness of the template and finger positioning for verification. It is based on a small 'yardstick' pattern being moved inside the image to be verified, and when a match is found the rest of the verification process is started. The yardstick is divided into segments in order to fit different DSP and microcontroller register lengths and structures. The software code implements the algorithm in the most efficient way for utilization on a digital signal processor architecture. A main area of application is in fingerprint identification and verification, but it could be implemented with same success in other image verification systems like iris, eye, face etc., where two images should be compared. The algorithm is adaptive, so various false acceptance and false rejection coefficients could be utilized very easily. It also produces a very compact template, which leads to better memory utilization in fingerprint databases. | en | SEGMENTED SLIDING YARDSTICKS ERROR TOLERANT FINGERPRINT ENROLLMENT AND VERIFICATION SYSTEM AND METHOD | 16557350_CA | 16557350_CA | G06V 40/1347 | [
"G06K 9/00"
] | 5,459 |
488,986,054 | 2016-09-06 | 57,289,670 | N | An emotion EEG recognition method providing emotion recognition model time robustness, comprising: performing pre-processing on a collected 64-lead EEG signal comprising changing a reference to a binaural average, downsampling to 500 Hz, performing 1-100 Hz bandpass filtering, and using an independent component analysis algorithm to remove EOG interference; finding an optimal discriminative frequency component in a pre-processed EEG signal by means of adaptive tracking of discriminative frequency components, and calculating a power spectral density of the optimal discriminative frequency component on each lead, respectively, forming an emotion characteristic matrix; using principal component analysis to perform dimension reduction on the characteristic matrix; using a support vector machine classifier to perform recognition on the dimension-reduced EEG power spectrum characteristics, establishing an emotion recognition model. The described solution finds an optimal discriminative frequency component by means of adaptive tracking of discriminative frequency components, strengthens emotion correlation characteristics by means of increasing training set sample days in an emotion recognition model, weakens a time specificity characteristic, and increases time robustness of an emotion recognition model. | en | EMOTION EEG RECOGNITION METHOD PROVIDING EMOTION RECOGNITION MODEL TIME ROBUSTNESS | 68332186_CN | 63694407_CN,67860103_CN,68057442_CN,69068776_CN,63511718_CN,63919400_CN,63789053_CN,64230228_CN | A61B 5/165,A61B 5/369,A61B 5/72,A61B 5/7203,A61B 5/725,A61B 5/7264,A61B 5/7271 | [
"A61B 5/0476"
] | 114,864 |
38,578,653 | 2004-07-20 | 34,797,177 | Y | <P>PROBLEM TO BE SOLVED: To provide a visual guidance information generation system by which a device is miniaturized, cost is reduced and appropriate eye flow is surely calculated. <P>SOLUTION: A layout device 100 sections a layout image into a plurality of areas, calculates attention degree feature quantity and positional information of each partial area based on image data and weights the calculated attention degree feature quantity. Next, a starting partial area is detected among a plurality of partial areas based on the calculated attention degree feature quantity and reference points of each partial area are detected among sight line guidance featured spaces based on the weighted attention degree feature quantity and the calculated positional information. Then, the distance between the respective detected reference points is calculated and visual guidance information is generated based on the calculated distance between the reference points. The visual guidance information becomes a visual guidance path which sequentially passes each partial area using the starting partial area as a starting point and using a partial area with the minimum distance between the reference points viewed from a partial area at the previous stage as a partial area at the next stage. <P>COPYRIGHT: (C)2005,JPO&NCIPI | en | SYSTEM, PROGRAM AND METHOD FOR GENERATING VISUAL GUIDANCE INFORMATION | 14217961_ | 31093446_,31906367_ | G06V 30/40 | [
"H04N 1/387",
"G06K 9/34",
"G06K 9/20",
"G06K 9/00",
"G06T 7/00"
] | 25,482 |
551,555,735 | 2020-11-13 | 70,202,545 | N | An automatic tumor segmentation system and method, and an electronic device. The method comprises: acquiring an original image of a liver tumor, and respectively naming the original image and a corresponding ground truth image and then placing same in a folder (100); performing cutting processing on the original image, reserving a non-zero region of the image, analyzing data of the image and acquiring image data such as an interval, an intensity distribution and a shape, and placing the acquired image data in a new folder (200); constructing a MILU-Net network model by combining a U-Net network, an attention mechanism and a dual-stage cascade architecture (300); inputting data in the new folder into the MILU-Net network model for training, so as to obtain a weight file (400); and loading the weight file, configuring neurons of the MILU-Net network model according to the weight, and outputting a segmentation result (500). The present invention combines the features of network architectures having a plurality of segmentation network characteristics to overcome its own shortcomings by learning from the strong points of others, improving the existing monotonic network framework, improving the generalization performance and the segmentation precision of the model, and increasing the training efficiency. | en | AUTOMATIC TUMOR SEGMENTATION SYSTEM AND METHOD, AND ELECTRONIC DEVICE | 66824291_CN | 68075506_CN,81400319_CN | G06N 3/0454,G06N 3/08,G06T 7/11,G06T2207/10081,G06T2207/20081,G06T2207/20084,G06T2207/30084,G06T2207/30096 | [
"G06T 7/11",
"G06N 3/04"
] | 153,910 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.