---
annotations_creators:
- The International Skin Imaging Collaboration
language_creators:
- found
language:
- en
license: cc-by-2.0
multilinguality:
- monolingual
size_categories:
- 10K<n<100K
source_datasets:
- ISIC 2019 Challenge
task_categories:
- image-classification
task_ids: []
paperswithcode_id: ISIC-2019
pretty_name: ISIC 2019 RESIZED (224x224)
dataset_info:
  config_name: default
  features:
    - name: img
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': Melanoma
            '1': Melanocytic nevus
            '2': Basal cell carcinoma
            '3': Actinic keratosis
            '4': Benign keratosis
            '5': Dermatofibroma
            '6': Vascular lesion
            '7': Squamous cell carcinoma
            '8': None of the others
---

# Dataset Card for ISIC-2019 RESIZED

## Dataset Description

### Dataset Summary
[The original ISIC 2019](https://challenge.isic-archive.com/landing/2019/) dataset contains 25,331 images of dermoscopic images from nine different diagnostic categories.
Published by the [The International Skin Imaging Collbaroation](https://www.isic-archive.com/), it serves as a challenging image classification dataset where the distribution
across classes is heavily skewed towards the first two classes.

The ISIC-2019 RESIZED is a downsampled version of the original dataset, downsampled and archived on HuggingFace for convenience of use. The downsampling was performed by
[Carlo Metta](https://scholar.google.com/citations?user=THYM6sIAAAAJ&hl=it), and the processing was done by [Maciej Zuziak](https://scholar.google.com/citations?user=v8JvNv0AAAAJ&hl=pl).


### Citation Rules 
The aggregate 'ISIC_2019: Training' data must be cited as original work:

```
    BCN_20000 Dataset: (c) Department of Dermatology, Hospital Clínic de Barcelona

    HAM10000 Dataset: (c) by ViDIR Group, Department of Dermatology, Medical University of Vienna; https://doi.org/10.1038/sdata.2018.161   [Add to Citavi project by DOI]

    MSK Dataset: (c) Anonymous; https://arxiv.org/abs/1710.05006; https://arxiv.org/abs/1902.03368
```

When referencing this dataset in your own manuscripts and publications, please use the following full citations: 
 
 ```

    [1] Tschandl P., Rosendahl C. & Kittler H. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data 5, 180161 doi.10.1038/sdata.2018.161 (2018)

    [2] Noel C. F. Codella, David Gutman, M. Emre Celebi, Brian Helba, Michael A. Marchetti, Stephen W. Dusza, Aadi Kalloo, Konstantinos Liopyris, Nabin Mishra, Harald Kittler, Allan Halpern: "Skin Lesion Analysis Toward Melanoma Detection: A Challenge at the 2017 International Symposium on Biomedical Imaging (ISBI), Hosted by the International Skin Imaging Collaboration (ISIC)", 2017; arXiv:1710.05006.   [Add to Citavi project by ArXiv ID]

    [3] Hernández-Pérez C, Combalia M, Podlipnik S, Codella NC, Rotemberg V, Halpern AC, Reiter O, Carrera C, Barreiro A, Helba B, Puig S, Vilaplana V, Malvehy J. BCN20000: Dermoscopic lesions in the wild. Scientific Data. 2024 Jun 17;11(1):641.
```

Additionally, if the resampled version is used, please cite this repository.