metadata

license: mit
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': benign
            '1': malignant
  splits:
    - name: train
      num_bytes: 9084040
      num_examples: 400
    - name: validation
      num_bytes: 2032363
      num_examples: 100
  download_size: 11133883
  dataset_size: 11116403

Lymphnode Cancer Biopsy Dataset (100k)

Overview

This dataset contains biopsy images of lymphnode cancer tissues, divided into two classes: benign and malignant. Each sample is stored in a separate image file, organized into respective class folders. The dataset is structured to be compatible with Lumina AI's Random Contrast Learning (RCL) algorithm via the PrismRCL application or API.

Dataset Structure

The dataset is organized into the following structure:

{dataset_folder_name}/
    train_data/
        benign/
            sample_0.png
            sample_1.png
            ...
        malignant/
            sample_0.png
            sample_1.png
            ...
    test_data/
        benign/
            sample_0.png
            sample_1.png
            ...
        malignant/
            sample_0.png
            sample_1.png
            ...

Note: All image file names must be unique across all class folders.

Features

Image Data: Each file contains a biopsy image of lymphnode cancer tissue.
Classes: There are two classes, each represented by a separate folder based on the type of tissue (benign or malignant).

Usage (not pre-split)

Here is an example of how to load the dataset using PrismRCL:

C:\PrismRCL\PrismRCL.exe chisquared rclticks=10 boxdown=0 data=C:\path\to\Lymphnode_Cancer_Biopsy_100k testsize=0.1 savemodel=C:\path\to\models\mymodel.classify log=C:\path\to\log_files stopwhendone

Explanation of Command:

C:\PrismRCL\PrismRCL.exe: Path to the PrismRCL executable for classification
chisquared: Specifies Chi-squared as the training evaluation method
rclticks=10: Sets the number of RCL iterations during training to 10
boxdown=0: Configuration parameter for training behavior
data=C:\path\to\Lymphnode_Cancer_Biopsy_100k: Path to the complete dataset for Lymphnode Cancer Biopsy classification
testsize=0.1: Specifies that 10% of the data should be used for testing
savemodel=C:\path\to\models\mymodel.classify: Path to save the resulting trained model
log=C:\path\to\log_files: Directory path for storing log files of the training process
stopwhendone: Instructs PrismRCL to end the session once training is complete

License

This dataset is licensed under the Creative Commons Attribution 4.0 International License. See the LICENSE file for more details.

Original Source

This dataset was originally sourced from the GitHub Repository. Please cite the original source if you use this dataset in your research or applications.

Additional Information

The data values have been prepared to ensure compatibility with PrismRCL. No normalization is required as of version 2.4.0.

Citations

If you use this dataset in your research, please cite the following papers:

Veeling, B. S., Linmans, J., Winkens, J., Cohen, T., & Welling, M. (2018). Rotation Equivariant CNNs for Digital Pathology. arXiv preprint arXiv:1806.03962.
Ehteshami Bejnordi, B., Veta, M., Johannes van Diest, P., van Ginneken, B., Karssemeijer, N., Litjens, G., ... & the CAMELYON16 Consortium. (2017). Diagnostic Assessment of Deep Learning Algorithms for Detection of Lymph Node Metastases in Women With Breast Cancer. JAMA, 318(22), 2199–2210. https://doi.org/10.1001/jama.2017.14585