Skip to content

Latest commit

 

History

History
89 lines (60 loc) · 3.54 KB

ISIC2020.md

File metadata and controls

89 lines (60 loc) · 3.54 KB

ISIC 2020

Dataset Information

The ISIC 2020 dataset, provided by the International Skin Imaging Collaboration (ISIC), is a large-scale dermoscopy image classification dataset. It consists of 33,126 images of benign and malignant skin lesions from 2,056 patients. The images have been sourced from 6 medical institutions across the globe. Unlike previous datasets, ISIC 2020 emphasizes the importance of patient-level information, which is particularly crucial in clinical practice. Each image comes with a unique patient identifier, aiding in more accurate melanoma diagnosis and reducing false positives. All malignant diagnoses have been confirmed through histopathology, while benign diagnoses have also been reliably verified.

Dataset Meta Information

Dimensions Modality Task Type Anatomical Structures Anatomical Area Number of Categories Data Volume File Format
2D dermoscopic Classification Skin Melanoma 2 33,126 DICOM, JPG

Resolution Details

Dataset Statistics size
min (640, 480)
median (5184, 3456)
max (6000, 6000)

Label Information Statistics

Category Malignant (malignant) Benign (benign)
Image Count 584 32542
Percentage 1.8% 98.2%

Visualization

Paper Visualization.

File Structure

The ISIC2020 dataset includes a train directory for storing training images, and an ISIC_2020_Training_GroundTruth.csv file, which lists patient information and diagnosis results corresponding to each image.

ISIC2020
│
├── train
│   ├── ISIC_0015719.jpg
│   └── ...
└── ISIC_2020_Training_GroundTruth.csv

Authors and Institutions

Peter Soyer (The University of Queensland, Dermatology Research Centre, Australia)

Allan Halpern (Memorial Sloan Kettering Cancer Center, USA)

Pascale Guitera (Melanoma Institute Australia, Australia)

Source Information

Official Website: https://challenge2020.isic-archive.com/

Download Link: https://challenge2020.isic-archive.com/

Article Address: https://www.nature.com/articles/s41597-021-00815-z

Publication Date: 2020

Citation

@article{rotemberg2021patient,
  title={A patient-centric dataset of images and metadata for identifying melanomas using clinical context},
  author={Rotemberg, Veronica and Kurtansky, Nicholas and Betz-Stablein, Brigid and Caffery, Liam and Chousakos, Emmanouil and Codella, Noel and Combalia, Marc and Dusza, Stephen and Guitera, Pascale and Gutman, David and others},
  journal={Scientific data},
  volume={8},
  number={1},
  pages={34},
  year={2021},
  publisher={Nature Publishing Group UK London}
}

Original introduction article is here.