{"success":true,"data":{"author":"International Skin Imaging Collaboration (ISIC); repackaged by MedOtter","bucketAccessType":"public","bucketName":"MedOtter/ISIC2017","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-21T22:16:41.400Z","description":"Task 1 data from the ISBI 2017 challenge Skin Lesion Analysis Toward Melanoma Detection, hosted by the International Skin Imaging Collaboration (ISIC): dermoscopy photographs of skin lesions with expert binary lesion masks. There are 2,000 training, 150 validation and 600 test images (the final Test_v2 release), all with public ground truth, covering melanoma (521), seborrheic keratosis (386) and benign nevus (1,843). Resolutions range from about 540x722 to 4499x6748 pixels. Each row has the unmodified JPEG, the mask PNG (0 background, 255 lesion), approximate age and sex from the challenge metadata, and a flag for images that reappear in the ISIC 2018 training split; isic_2017_2018_split_ids.json lists the overlap. 63 Parquet files under data/, about 12 GB (measured 2026-11-21); read with Hugging Face datasets. Use for skin lesion segmentation and melanoma image analysis benchmarks. The ISIC challenge data page lists the 2017 data under CC-0. Cite Codella et al., ISBI 2018, doi 10.1109/ISBI.2018.8363547.","displayTitle":"ISIC 2017 Task 1: dermoscopy skin lesion segmentation, 2,750 images of melanoma, seborrheic keratosis and nevi with masks","huggingFaceUrl":"https://huggingface.co/datasets/MedOtter/ISIC2017","id":"1023ef6d-8e75-50b8-be9a-e0f5c7876edc","license":"CC0-1.0","licenseEvidenceUrl":"https://challenge.isic-archive.com/data/","objectCountEstimate":66,"probeStatus":"ok","referenceUrl":"https://doi.org/10.1109/ISBI.2018.8363547","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/ISIC2017\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MedOtter/ISIC2017\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MedOtter, license:cc0-1.0, size_categories:1K<n<10K, format:parquet, downloads=745"],"format":"mixed","generatedAt":"2026-11-21T22:32:04.359Z","layout":{"example":"data/test-00000-of-00028.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test-00000-of-00028.parquet","truncatedAt":512}],"sampleObjects":["data/test-00000-of-00028.parquet","data/test-00001-of-00028.parquet","data/test-00002-of-00028.parquet","data/test-00003-of-00028.parquet","data/test-00004-of-00028.parquet"],"sidecars":[{"bytes":"---\nlicense: cc0-1.0\ntask_categories:\n- image-segmentation\ntags:\n- medical\n- dermoscopy\n- skin\n- melanoma\n- lesion-segmentation\npretty_name: ISIC 2017 — Skin Lesion Segmentation (Task 1)\nsize_categories:\n- 1K<n<10K\ndataset_info:\n  features:\n  - name: image_id\n    dtype: string\n  - name: image\n    dtype: image\n  - name: mask\n    dtype: image\n  - name: age_approximate\n    dtype: float32\n  - name: sex\n    dtype: string\n  - name: in_isic2018_train\n    dtype: bool\n  splits:\n  - name: train\n    num_bytes: 5802730760\n    num_examples: 2000\n  - name: validation\n    num_bytes: 871819732\n    num_examples: 150\n  - name: test\n    num_bytes: 5575275700\n    num_examples: 600\n  download_size: 12250441569\n  dataset_size: 12249826192\nconfigs:\n- config_name: default\n  data_files:\n  - split: train\n    path: data/train-*\n  - split: validation\n    path: data/validation-*\n  - split: test\n    path: data/test-*\n---\n\n# ISIC 2017 — Skin Lesion Segmentation (Task 1)\n\nDermoscopic RGB images of skin lesions (melanoma / seborrheic keratosis / nevus)\nwith expert binary lesion-boundary masks, from the ISBI 2017 challenge\n*\"Skin Lesion Analysis Toward Melanoma Detection\"* hosted by the International\nSkin Imaging Collaboration (ISIC).\n\n| Split | Images | Masks | Melanoma | Seborrheic keratosis | Nevus |\n|---|---|---|---|---|---|\n| train | 2000 | 2000 | 374 | 254 | 1372 |\n| validation | 150 | 150 | 30 | 42 | 78 |\n| test | 600 | 600 | 117 | 90 | 393 |\n\nAll three splits carry public ground truth. The test split is the official\n**`Test_v2`** release — the final public revision of the challenge test set\n(v1 was withdrawn upstream). Resolutions vary widely (~540×722 up to\n4499×6748 px).\n\n## Schema\n\n| Column | Type | Notes |\n|---|---|---|\n| `image_id` | string | Stable ISIC Archive ID (`ISIC_XXXXXXX`) — cross-references all ISIC challenge editions |\n| `image` | Image | Original challenge JPEG, bytes unmodified |\n| `mask` | Image | Original expert mask PNG: **0 = background, 255 = lesion** |\n| `age_approximate` | float | From the official challenge metadata CSV; null when unknown |\n| `sex` | string | `male` / `female`; null when unknown |\n| `in_isic2018_train` | bool | True if this image also appears in ISIC 2018 Task 1's training split (see below) |\n\nMasks were created by expert clinicians via manual polyline tracing or a\nsupervised flood-fill workflow (one published mask per image; the per-image\nmethod is not disclosed). The `*_superpixels.png` scaffolding files and the\nPart-2 (dermoscopic features) / Part-3 (classification) ground truth belong to\nother challenge tasks and are not mirrored here.\n\n## ⚠️ Overlap with ISIC 2018 (leakage warning)\n\nISIC challenge editions draw from the same growing archive. Measured by exact\n`image_id` intersection against ISIC 2018 Task 1:\n\n- **2,450 / 2,750 images (89.1%) of ISIC 2017 reappear in ISIC 2018's training\n  split** — train 1800/2000, validation 121/150, test 529/600.\n- ISIC 2018's own validation/test splits share **zero** IDs with ISIC 2017.\n- Consequence: a model trained on ISIC 2018 Task 1 training data has already\n  seen 650 of the 750 ISIC 2017 validation+test images. Filter on\n  `in_isic2018_train` (or use `isic_2017_2018_split_ids.json` at the repo root)\n  before treating the two datasets as independent benchmarks.\n- Filename-level matching is a **lower bound**: pixel-level near-duplicates\n  (rescaled/re-encoded variants) exist across ISIC editions and against ISIC\n  2016 — see Cassidy et al., *Medical Image Analysis* 75:102305 (2022),\n  https://github.com/mmu-dermatology-research/isic_duplicate_removal_strategy.\n\n## Provenance\n\nOfficial author-hosted S3 bucket\n(`https://isic-archive.s3.amazonaws.com/challenges/2017/`), zips verified\nbyte-exact against Content-Length; image/mask counts match the challenge paper\n(2000/150/600). Original encoded bytes are embedded unmodified.\n\n## License\n\n**CC-0 (public domain)**, per the 2017 section of the\n[ISIC Challenge data page](https://challenge.isic-archive.com/data/).\n\n## Citation\n\n```bibtex\n@inproceedings{codella2018skin,\n  title     = {Skin lesion anal","encoding":"utf8","key":"README.md","name":"README.md","size":4767,"truncated":true}],"sourceObjectKey":"data/test-00000-of-00028.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/ISIC2017\")\nprint(ds)","starterSamplePath":"data/test-00000-of-00028.parquet","totalObjects":66},"summary":"Task 1 data from the ISBI 2017 challenge Skin Lesion Analysis Toward Melanoma Detection, hosted by the International Skin Imaging Collaboration (ISIC): dermoscopy photographs of…","title":"ISIC 2017 Task 1: dermoscopy skin lesion segmentation, 2,750 images of melanoma, seborrheic keratosis and nevi with masks","updatedAt":"2026-11-22T01:11:11.291Z","visibility":"public","websiteUrl":"https://challenge.isic-archive.com/data/"}}