Hugging Face's logo

qualcomm
/
MediaPipe-Hand-Detection

Object Detection
PyTorch
LiteRT
ONNX
real_time
android
Model card Files Community

MediaPipe-Hand-Detection: Optimized for Mobile Deployment

Real-time hand detection optimized for mobile and edge

The MediaPipe Hand Landmark Detector is a machine learning pipeline that predicts bounding boxes and pose skeletons of hands in an image.

This model is an implementation of MediaPipe-Hand-Detection found here.

This repository provides scripts to run MediaPipe-Hand-Detection on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

  • Model Type: Object detection
  • Model Stats:
    • Input resolution: 256x256
    • Number of parameters (MediaPipeHandDetector): 1.76M
    • Model size (MediaPipeHandDetector): 6.76 MB
    • Number of parameters (MediaPipeHandLandmarkDetector): 2.01M
    • Model size (MediaPipeHandLandmarkDetector): 7.71 MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
MediaPipeHandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 0.737 ms 0 - 31 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 0.712 ms 1 - 3 MB FP16 NPU MediaPipe-Hand-Detection.so
MediaPipeHandDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 1.007 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.542 ms 0 - 32 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.527 ms 0 - 18 MB FP16 NPU MediaPipe-Hand-Detection.so
MediaPipeHandDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.694 ms 0 - 38 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.609 ms 0 - 26 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.535 ms 0 - 20 MB FP16 NPU Use Export Script
MediaPipeHandDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.72 ms 1 - 22 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandDetector SA7255P ADP SA7255P TFLITE 24.617 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector SA7255P ADP SA7255P QNN 24.625 ms 1 - 10 MB FP16 NPU Use Export Script
MediaPipeHandDetector SA8255 (Proxy) SA8255P Proxy TFLITE 0.743 ms 0 - 31 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector SA8255 (Proxy) SA8255P Proxy QNN 0.712 ms 1 - 3 MB FP16 NPU Use Export Script
MediaPipeHandDetector SA8295P ADP SA8295P TFLITE 1.742 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector SA8295P ADP SA8295P QNN 1.678 ms 0 - 18 MB FP16 NPU Use Export Script
MediaPipeHandDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.738 ms 0 - 32 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector SA8650 (Proxy) SA8650P Proxy QNN 0.709 ms 1 - 3 MB FP16 NPU Use Export Script
MediaPipeHandDetector SA8775P ADP SA8775P TFLITE 1.554 ms 0 - 21 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector SA8775P ADP SA8775P QNN 1.488 ms 1 - 11 MB FP16 NPU Use Export Script
MediaPipeHandDetector QCS8275 (Proxy) QCS8275 Proxy TFLITE 24.617 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector QCS8275 (Proxy) QCS8275 Proxy QNN 24.625 ms 1 - 10 MB FP16 NPU Use Export Script
MediaPipeHandDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 0.734 ms 0 - 30 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector QCS8550 (Proxy) QCS8550 Proxy QNN 0.706 ms 1 - 4 MB FP16 NPU Use Export Script
MediaPipeHandDetector QCS9075 (Proxy) QCS9075 Proxy TFLITE 1.554 ms 0 - 21 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector QCS9075 (Proxy) QCS9075 Proxy QNN 1.488 ms 1 - 11 MB FP16 NPU Use Export Script
MediaPipeHandDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.426 ms 0 - 29 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.406 ms 1 - 27 MB FP16 NPU Use Export Script
MediaPipeHandDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 0.857 ms 1 - 1 MB FP16 NPU Use Export Script
MediaPipeHandDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.018 ms 4 - 4 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 1.017 ms 0 - 63 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 1.011 ms 0 - 2 MB FP16 NPU MediaPipe-Hand-Detection.so
MediaPipeHandLandmarkDetector Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 1.338 ms 0 - 42 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 0.756 ms 0 - 36 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 0.748 ms 1 - 20 MB FP16 NPU MediaPipe-Hand-Detection.so
MediaPipeHandLandmarkDetector Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 0.948 ms 0 - 37 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 0.706 ms 0 - 28 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 0.7 ms 0 - 18 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 0.902 ms 1 - 23 MB FP16 NPU MediaPipe-Hand-Detection.onnx
MediaPipeHandLandmarkDetector SA7255P ADP SA7255P TFLITE 35.408 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector SA7255P ADP SA7255P QNN 35.37 ms 1 - 10 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector SA8255 (Proxy) SA8255P Proxy TFLITE 1.015 ms 0 - 64 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector SA8255 (Proxy) SA8255P Proxy QNN 1.031 ms 1 - 3 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector SA8295P ADP SA8295P TFLITE 2.325 ms 0 - 23 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector SA8295P ADP SA8295P QNN 2.248 ms 0 - 18 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector SA8650 (Proxy) SA8650P Proxy TFLITE 0.999 ms 0 - 64 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector SA8650 (Proxy) SA8650P Proxy QNN 1.018 ms 1 - 3 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector SA8775P ADP SA8775P TFLITE 2.216 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector SA8775P ADP SA8775P QNN 2.143 ms 0 - 10 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector QCS8275 (Proxy) QCS8275 Proxy TFLITE 35.408 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector QCS8275 (Proxy) QCS8275 Proxy QNN 35.37 ms 1 - 10 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy TFLITE 1.01 ms 0 - 64 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector QCS8550 (Proxy) QCS8550 Proxy QNN 1.005 ms 1 - 3 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector QCS9075 (Proxy) QCS9075 Proxy TFLITE 2.216 ms 0 - 22 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector QCS9075 (Proxy) QCS9075 Proxy QNN 2.143 ms 0 - 10 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy TFLITE 1.945 ms 0 - 34 MB FP16 NPU MediaPipe-Hand-Detection.tflite
MediaPipeHandLandmarkDetector QCS8450 (Proxy) QCS8450 Proxy QNN 1.936 ms 1 - 33 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite QNN 1.218 ms 1 - 1 MB FP16 NPU Use Export Script
MediaPipeHandLandmarkDetector Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.38 ms 6 - 6 MB FP16 NPU MediaPipe-Hand-Detection.onnx

Installation

Install the package via pip:

pip install qai-hub-models

Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub with your Qualcomm® ID. Once signed in navigate to Account -> Settings -> API Token.

With this API token, you can configure your client to run models on the cloud hosted devices.

qai-hub configure --api_token API_TOKEN

Navigate to docs for more information.

Demo off target

The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input.

python -m qai_hub_models.models.mediapipe_hand.demo

The above demo runs a reference implementation of pre-processing, model inference, and post processing.

NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).

%run -m qai_hub_models.models.mediapipe_hand.demo

Run model on a cloud-hosted device

In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following:

  • Performance check on-device on a cloud-hosted device
  • Downloads compiled assets that can be deployed on-device for Android.
  • Accuracy check between PyTorch and on-device outputs.
python -m qai_hub_models.models.mediapipe_hand.export

Profiling Results
------------------------------------------------------------
MediaPipeHandDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 0.7                    
Estimated peak memory usage (MB): [0, 31]                
Total # Ops                     : 149                    
Compute Unit(s)                 : NPU (149 ops)          

------------------------------------------------------------
MediaPipeHandLandmarkDetector
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 1.0                    
Estimated peak memory usage (MB): [0, 63]                
Total # Ops                     : 158                    
Compute Unit(s)                 : NPU (158 ops)

How does this work?

This export script leverages Qualcomm® AI Hub to optimize, validate, and deploy this model on-device. Lets go through each step below in detail:

Step 1: Compile model for on-device deployment

To compile a PyTorch model for on-device deployment, we first trace the model in memory using the jit.trace and then call the submit_compile_job API.

import torch

import qai_hub as hub
from qai_hub_models.models.mediapipe_hand import Model

# Load the model
model = Model.from_pretrained()
hand_detector_model = model.hand_detector
hand_landmark_detector_model = model.hand_landmark_detector

# Device
device = hub.Device("Samsung Galaxy S23")

# Trace model
hand_detector_input_shape = hand_detector_model.get_input_spec()
hand_detector_sample_inputs = hand_detector_model.sample_inputs()

traced_hand_detector_model = torch.jit.trace(hand_detector_model, [torch.tensor(data[0]) for _, data in hand_detector_sample_inputs.items()])

# Compile model on a specific device
hand_detector_compile_job = hub.submit_compile_job(
    model=traced_hand_detector_model ,
    device=device,
    input_specs=hand_detector_model.get_input_spec(),
)

# Get target model to run on-device
hand_detector_target_model = hand_detector_compile_job.get_target_model()
# Trace model
hand_landmark_detector_input_shape = hand_landmark_detector_model.get_input_spec()
hand_landmark_detector_sample_inputs = hand_landmark_detector_model.sample_inputs()

traced_hand_landmark_detector_model = torch.jit.trace(hand_landmark_detector_model, [torch.tensor(data[0]) for _, data in hand_landmark_detector_sample_inputs.items()])

# Compile model on a specific device
hand_landmark_detector_compile_job = hub.submit_compile_job(
    model=traced_hand_landmark_detector_model ,
    device=device,
    input_specs=hand_landmark_detector_model.get_input_spec(),
)

# Get target model to run on-device
hand_landmark_detector_target_model = hand_landmark_detector_compile_job.get_target_model()

Step 2: Performance profiling on cloud-hosted device

After compiling models from step 1. Models can be profiled model on-device using the target_model. Note that this scripts runs the model on a device automatically provisioned in the cloud. Once the job is submitted, you can navigate to a provided job URL to view a variety of on-device performance metrics.

hand_detector_profile_job = hub.submit_profile_job(
    model=hand_detector_target_model,
    device=device,
)
hand_landmark_detector_profile_job = hub.submit_profile_job(
    model=hand_landmark_detector_target_model,
    device=device,
)

Step 3: Verify on-device accuracy

To verify the accuracy of the model on-device, you can run on-device inference on sample input data on the same cloud hosted device.

hand_detector_input_data = hand_detector_model.sample_inputs()
hand_detector_inference_job = hub.submit_inference_job(
    model=hand_detector_target_model,
    device=device,
    inputs=hand_detector_input_data,
)
hand_detector_inference_job.download_output_data()
hand_landmark_detector_input_data = hand_landmark_detector_model.sample_inputs()
hand_landmark_detector_inference_job = hub.submit_inference_job(
    model=hand_landmark_detector_target_model,
    device=device,
    inputs=hand_landmark_detector_input_data,
)
hand_landmark_detector_inference_job.download_output_data()

With the output of the model, you can compute like PSNR, relative errors or spot check the output with expected output.

Note: This on-device profiling and inference requires access to Qualcomm® AI Hub. Sign up for access.

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

  • TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

  • QNN (.so export ): This sample app provides instructions on how to use the .so shared library in an Android application.

View on Qualcomm® AI Hub

Get more details on MediaPipe-Hand-Detection's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of MediaPipe-Hand-Detection can be found here.
  • The license for the compiled assets for on-device deployment can be found here

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Object Detection
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The HF Inference API does not support object-detection models for pytorch library.