{"success":true,"data":{"author":"Mohamed Amgad et al., Emory University / PathologyDataScience (re-hosted by MedOtter)","bucketAccessType":"public","bucketName":"MedOtter/BCSS","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:12.246Z","description":"BCSS (Breast Cancer Semantic Segmentation) is a histopathology dataset from Amgad et al., Bioinformatics 2019, in which pathologists, residents and medical students annotated regions of interest from TCGA breast cancer whole-slide images through a structured crowdsourcing protocol. The masks cover 22 tissue classes, among them tumor, stroma, lymphocytic infiltrate, necrosis or debris, blood vessels, fat, mucoid material, normal acinus or duct and other immune infiltrate, with an outside-ROI don't-care class. This repository, re-hosted by MedOtter, holds the 151 ROIs as colour-normalised RGB PNG images at 0.25 microns per pixel with indexed masks and columns for image ID, TCGA patient ID and ROI coordinates. There is no official split, so split by patient. As measured 2026-11-14 the data are ten Parquet shards totalling about 4.9 GB; load with datasets (load_dataset('MedOtter/BCSS')) or pyarrow and decode with PIL. Used for tissue segmentation, tumor microenvironment and tumor-infiltrating lymphocyte research, and pathology foundation model evaluation. The authors release the data under CC0 1.0 in the BCSS GitHub repository; cite Amgad et al. 2019.","displayTitle":"BCSS: Breast Cancer Semantic Segmentation of TCGA H&E regions of interest","id":"14c0519d-0fdd-5401-8eb8-4340e7218814","license":"CC0-1.0","licenseEvidenceUrl":"https://github.com/PathologyDataScience/BCSS","objectCountEstimate":12,"probeStatus":"ok","referenceUrl":"https://doi.org/10.1093/bioinformatics/btz083","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/BCSS\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MedOtter/BCSS\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MedOtter, license:cc0-1.0, size_categories:n<1K, format:parquet, downloads=516"],"format":"mixed","generatedAt":"2026-11-14T14:15:20.242Z","layout":{"example":"data/train-00000-of-00010.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/train-00000-of-00010.parquet","truncatedAt":512}],"sampleObjects":["data/train-00000-of-00010.parquet","data/train-00001-of-00010.parquet","data/train-00002-of-00010.parquet","data/train-00003-of-00010.parquet","data/train-00004-of-00010.parquet"],"sidecars":[{"bytes":"---\nlicense: cc0-1.0\ntask_categories:\n  - image-segmentation\ntags:\n  - medical\n  - pathology\n  - h-and-e\n  - breast\n  - segmentation\n  - bcss\nsize_categories:\n  - n<1K\n---\n\n# BCSS — Breast Cancer Semantic Segmentation (Amgad et al. 2019)\n\nRe-hosted mirror of the **Breast Cancer Semantic Segmentation** dataset\n(Amgad et al., *Bioinformatics* 2019), originally distributed via the\n[PathologyDataScience/BCSS](https://github.com/PathologyDataScience/BCSS)\nGitHub repo and rebuilt here from the\n[`nabil-m/bcss`](https://huggingface.co/datasets/nabil-m/bcss) HF mirror.\n\nThe data is **CC0 1.0** (public domain, no rights reserved); the upstream\ncodebase is MIT-licensed but covers software, not data. Redistribution is\nunrestricted.\n\n## Composition\n\n| Split | ROIs |\n|-------|-----:|\n| train | 151  |\n\n151 ROI patches extracted from TCGA breast cancer whole-slide images.\nPatches are **color-normalized** RGB at the upstream MPP=0.25 µm/px\n(40× equivalent), with native ROI resolution typically 2–4k px per\nside. There is **no official train/val/test split** — group-shuffle by\n`patient_id` downstream for honest evaluation.\n\n## Schema\n\n| Column      | Type     | Description                                           |\n|-------------|----------|-------------------------------------------------------|\n| `image`     | `Image`  | RGB ROI (PNG, color-normalized, variable size)        |\n| `mask`      | `Image`  | Indexed 22-class mask (`L`, values 0..21)             |\n| `image_id`  | `string` | Filename stem incl. xmin/ymin                         |\n| `patient_id`| `string` | TCGA-XX-YYYY prefix                                   |\n| `xmin`      | `int32`  | ROI bbox xmin in WSI base-magnification pixels        |\n| `ymin`      | `int32`  | ROI bbox ymin in WSI base-magnification pixels        |\n\n## Mask labels\n\n| Code | Class                  | | Code | Class                |\n|-----:|------------------------|-|-----:|----------------------|\n| 0    | outside_roi (don't care) | | 11   | other_immune_infiltrate |\n| 1    | tumor                  | | 12   | mucoid_material      |\n| 2    | stroma                 | | 13   | normal_acinus_or_duct |\n| 3    | lymphocytic_infiltrate | | 14   | lymphatics           |\n| 4    | necrosis_or_debris     | | 15   | undetermined         |\n| 5    | glandular_secretions   | | 16   | nerve                |\n| 6    | blood                  | | 17   | skin_adnexa          |\n| 7    | exclude                | | 18   | blood_vessel         |\n| 8    | metaplasia_NOS         | | 19   | angioinvasion        |\n| 9    | fat                    | | 20   | dcis                 |\n| 10   | plasma_cells           | | 21   | other                |\n\n**Code 0 (`outside_roi`) is a \"don't care\" region** — the original paper\nrecommends excluding it from any loss. For binary tumor evaluation,\nthe canonical foreground is class 1.\n\n## License\n\nCC0 1.0 Universal — public domain. No rights reserved.\n\n## Citation\n\n```bibtex\n@article{amgad2019structured,\n  title   = {Structured crowdsourcing enables convolutional segmentation of histology images},\n  author  = {Amgad, Mohamed and Elfandy, Habiba and Hussein, Hagar and others},\n  journal = {Bioinformatics},\n  volume  = {35},\n  number  = {18},\n  pages   = {3461--3467},\n  year    = {2019},\n  doi     = {10.1093/bioinformatics/btz083}\n}\n```\n","encoding":"utf8","key":"README.md","name":"README.md","size":3329,"truncated":false}],"sourceObjectKey":"data/train-00000-of-00010.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/BCSS\")\nprint(ds)","starterSamplePath":"data/train-00000-of-00010.parquet","totalObjects":12},"summary":"BCSS (Breast Cancer Semantic Segmentation) is a histopathology dataset from Amgad et al., Bioinformatics 2019, in which pathologists, residents and medical students annotated…","title":"BCSS: Breast Cancer Semantic Segmentation of TCGA H&E regions of interest","updatedAt":"2026-11-14T17:52:32.023Z","visibility":"public","websiteUrl":"https://github.com/PathologyDataScience/BCSS"}}