README.md

metadata

license: mit
dataset_info:
  features:
    - name: subject_id
      dtype: string
    - name: image_number
      dtype: int64
    - name: cell_count
      dtype: int64
    - name: image
      dtype:
        image:
          decode: false
    - name: label
      dtype: string
    - name: class_label
      dtype: string
    - name: fold
      dtype: int64
    - name: original_image_name
      dtype: string
    - name: relative_file_path
      dtype: string
  splits:
    - name: train
      num_bytes: 6487895691.044
      num_examples: 10661
  download_size: 1100428227
  dataset_size: 6487895691.044
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Dataset Summary

This dataset contains microscopic images of white blood cells for the purpose of identifying and classifying Acute Lymphoblastic Leukemia (ALL). It provides a valuable resource for researchers and practitioners in the field of medical imaging and hematology.

Field Name	Data Type	Description	Example Value	Usage
subject_id	String	Unique identifier for each patient	"1", "H24"	Patient-level grouping, analysis
image_number	Integer	Sequential number for images from the same patient	1, 10, 22	Image ordering, tracking
cell_count	Integer	Number of cells in the image	1, 2, 12	Feature for analysis/modeling
image	Image	Microscopic image of blood cells	(Binary image data)	Input for image analysis
label	String	Simple label (cancer/normal)	"cancer", "healthy"	Target variable for classification
class_label	String	Alias for label	"all", "hem"	Synonym for label
fold	Integer	Cross-validation fold assignment	0, 1, 2	Model training/evaluation
original_image_name	String	Original filename of the image	"UID_1_1_1_all.bmp"	Reference to source data
relative_file_path	String	Path to image relative to dataset root	"fold_0/all/UID_1_1_1_all.bmp"	Locating image files

Supported Tasks and Leaderboards

The dataset is well-suited for various machine learning tasks, including:

• Image Classification: Distinguish between ALL and healthy (HEM) cells.
• Object Detection: Locate and count individual cells within the images.
• Segmentation: Delineate the boundaries of individual cells in the images.

The ISBI 2019 ALL Challenge provided a leaderboard to benchmark performance on the classification task. You can find more information about the challenge and its results here: https://doi.org/10.7937/tcia.2019.dc64i46r

Data Splits

The dataset is provided as a single split (train) containing all 10,661 images. Researchers are encouraged to create their own validation and test splits, or utilize the pre-defined folds for cross-validation experiments.

Data Citation

Mourya, S., Kant, S., Kumar, P., Gupta, A., & Gupta, R. (2019). ALL Challenge dataset of ISBI 2019 (C-NMC 2019) (Version 1) [dataset]. The Cancer Imaging Archive. https://doi.org/10.7937/tcia.2019.dc64i46r