{"success":true,"data":{"author":"Hussein Naji, Reinhard Büttner et al.; mirrored by MedOtter","bucketAccessType":"public","bucketName":"MedOtter/LyNSeC","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-17T19:29:33.502Z","description":"LyNSeC (Lymphoma Nuclear Segmentation and Classification) is a histopathology dataset of diffuse large B-cell lymphoma (DLBCL) lymph node tissue with per-nucleus instance masks across H&E and immunohistochemistry. It has 699 RGB tiles of 512x512 pixels at 40x magnification with 161,247 annotated nuclei in three subsets: lynsec1_ihc (379 IHC tiles stained for CD3, Ki67 or ERG, 87,316 nuclei, marker-positive versus marker-negative class per nucleus), lynsec2_he (280 H&E tiles, 65,479 nuclei, instances only) and lynsec3_he_expert (40 expert-refined H&E tiles, 8,452 nuclei, tumor versus non-tumor class). Instance masks are 16-bit PNGs with one ID per nucleus; class masks are 8-bit. Parquet files under data/, about 360 MB (measured 2026-11-17); the source has no predefined train/test split. Use for nucleus instance segmentation, IHC positivity scoring and lymphoma cell classification. Mirrored on Hugging Face by MedOtter from the authors' release, with a card documenting provenance and any normalization. The authors' Zenodo record 8065174 is CC BY 4.0; cite Naji et al., HoLy-Net, Computers in Biology and Medicine 2024.","displayTitle":"LyNSeC: diffuse large B-cell lymphoma nucleus segmentation in H&E and IHC (CD3, Ki67, ERG), 161k nuclei","doi":"10.5281/zenodo.8065174","huggingFaceUrl":"https://huggingface.co/datasets/MedOtter/LyNSeC","id":"98a50862-fb2c-5911-ae23-00a2a28cdf9f","license":"CC-BY-4.0","licenseEvidenceUrl":"https://zenodo.org/records/8065174","objectCountEstimate":5,"probeStatus":"ok","referenceUrl":"https://www.sciencedirect.com/science/article/pii/S0010482524000623","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/LyNSeC\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MedOtter/LyNSeC\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MedOtter, license:cc-by-4.0, size_categories:n<1K, format:parquet, downloads=124"],"format":"mixed","generatedAt":"2026-11-17T19:40:31.918Z","layout":{"example":"data/lynsec1_ihc-00000-of-00001.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"50415231150415a61e15de0b4c1588031500120000930f3404000000305f6c31060000003130190a003115140032150a0033150a0034150a0035150a0036150a0037150a0038150a10395f6c3105056411130031190a156d0031156d0031156d0031156d0031156d0031156d0031156d0031156d0031156d15590032156d00321914156d0032156d0032156d0032156d0032156d0032156d0032156d0032156d154f0033156d0033156d0033191e156d0033156d0033156d0033156d0033156d0033156d0033156d15450034156d0034156d0034156d00341928156d0034156d0034156d0034156d0034156d0034156d153b0035156d0035156d0035156d0035156d00351932156d0035156d0035156d0035156d0035156d15310036156d0036156d0036156d0036156d0036156d0036193c156d0036156d0036156d0036156d15270037156d0037156d0037156d0037156d0037156d0037156d00371946156d0037156d0037156d151d0038156d0038156d0038156d0038156d0038156d0038156d0038156d00381950156d0038156d15130039156d0039156d0039156d0039156d0039156d0039156d0039156d0039156d0039195a156d0109814a11610032954a0032954a0032954a0032954a0032954a0032954a0032954a0032954a0032954a0032954a75d30032954a003275f10032758e0032752b003255c800325565","encoding":"hex","key":"data/lynsec1_ihc-00000-of-00001.parquet","truncatedAt":512}],"sampleObjects":["data/lynsec1_ihc-00000-of-00001.parquet","data/lynsec2_he-00000-of-00001.parquet","data/lynsec3_he_expert-00000-of-00001.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n  - image-segmentation\ntags:\n  - medical\n  - histopathology\n  - immunohistochemistry\n  - ihc\n  - h-and-e\n  - lymphoma\n  - dlbcl\n  - nuclei\n  - instance-segmentation\n  - lymphocytes\npretty_name: LyNSeC\nsize_categories:\n  - n<1K\n---\n\n# LyNSeC - Lymphoma Nuclear Segmentation and Classification\n\nHistopathology dataset of diffuse large B-cell lymphoma (DLBCL) tissue with\nper-nucleus instance segmentation masks across both H&E and IHC stains.\nReleased alongside the HoLy-Net paper (Naji et al., Comput Biol Med 2024).\n\nSource: <https://zenodo.org/records/8065174>\nPaper:  <https://www.sciencedirect.com/science/article/pii/S0010482524000623>\n\n## Overview\n- **Modality:** Histopathology (H&E + IHC brightfield microscopy)\n- **Tissue:** Lymph-node tissue, DLBCL\n- **Stains:** H&E, plus IHC markers CD3 / Ki67 / ERG\n- **Tile spec:** 512x512 RGB, 40x magnification\n- **Total samples:** 699 tiles, 161,247 annotated nuclei\n\n## Splits\n| Split | Stain | Tiles | Nuclei | Notes |\n|---|---|---|---|---|\n| `lynsec1_ihc` | IHC (CD3 / Ki67 / ERG) | 379 | 87,316 | Per-nucleus marker-positive vs marker-negative class label |\n| `lynsec2_he` | H&E | 280 | 65,479 | Instance-only (no per-cell class label) |\n| `lynsec3_he_expert` | H&E (expert-refined subset of LyNSeC 2) | 40 | 8,452 | Per-nucleus tumor vs non-tumor class label |\n\nThe Zenodo release does NOT ship predefined train/val/test split files;\nsplits in the original paper are defined in code, not as released folders.\n\n## Columns\n| Column | Type | Notes |\n|---|---|---|\n| `id` | string | Tile identifier from the Zenodo release (e.g. `0_l1`) |\n| `image` | Image (RGB) | 512x512 RGB tile |\n| `instance_mask` | Image (16-bit grayscale PNG, mode `I;16`) | Per-nucleus instance ID; 0 = background, 1..N = instance IDs (N up to 405 in lynsec2_he) |\n| `class_mask` | Image (mode `L`, 8-bit grayscale) - only on `lynsec1_ihc` and `lynsec3_he_expert` | Per-pixel class label: 0 = background. Non-zero class semantics are inherited from the source paper; see notes below |\n| `num_nuclei` | int32 | Number of instances in the tile (`= instance_mask.max()`) |\n\n### Class label semantics\n- **`lynsec1_ihc`**: classes {0, 1, 2}. Per the source paper, class indicates\n  marker positivity (one of marker-negative / marker-positive). The Zenodo\n  release does not document the specific mapping of `1` and `2` to those two\n  states - downstream users are expected to verify against the source code.\n- **`lynsec3_he_expert`**: classes {0, 1, 2}. Per the source paper, class\n  indicates tumor vs non-tumor cells. Same caveat about which integer maps to\n  which class.\n\nFor binary semantic segmentation (nucleus vs background) treat\n`instance_mask > 0` as foreground.\n\n## Derivation\nEach Zenodo .npy ships as `(512, 512, C)` int32 with `C in {4, 5}`:\n- channels 0..2 = RGB image (values 0..255 stored as int32),\n- channel 3 = instance map,\n- channel 4 = class map (only when `C == 5`).\n\nThis release re-encodes each tile as:\n- `image`         = `Image.fromarray(arr[:,:,:3].astype(np.uint8), 'RGB')` -> PNG\n- `instance_mask` = `Image.fromarray(arr[:,:,3].astype(np.uint16), 'I;16')` -> 16-bit PNG\n- `class_mask`    = `Image.fromarray(arr[:,:,4].astype(np.uint8),  'L')` -> 8-bit PNG (lynsec1/3 only)\n\nNo other preprocessing.\n\n## License\nCC BY 4.0 (per Zenodo record 8065174).\n\n## Citation\n```bibtex\n@article{naji2024holynet,\n  title   = {HoLy-Net: Segmentation of histological images of diffuse large\n             B-cell lymphoma},\n  author  = {Naji, Hussein and Sancere, Lucas and Simon, Adrian and Buttner,\n             Reinhard and Eich, Marie-Lisa and Lohneis, Philipp and Bozek,\n             Katarzyna},\n  journal = {Computers in Biology and Medicine},\n  volume  = {170},\n  pages   = {107978},\n  year    = {2024},\n  doi     = {10.1016/j.compbiomed.2024.107978}\n}\n```\n","encoding":"utf8","key":"README.md","name":"README.md","size":3837,"truncated":false}],"sourceObjectKey":"data/lynsec1_ihc-00000-of-00001.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/LyNSeC\")\nprint(ds)","starterSamplePath":"data/lynsec1_ihc-00000-of-00001.parquet","totalObjects":5},"summary":"LyNSeC (Lymphoma Nuclear Segmentation and Classification) is a histopathology dataset of diffuse large B-cell lymphoma (DLBCL) lymph node tissue with per-nucleus instance masks…","title":"LyNSeC: diffuse large B-cell lymphoma nucleus segmentation in H&E and IHC (CD3, Ki67, ERG), 161k nuclei","updatedAt":"2026-11-17T23:11:27.793Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/MedOtter/LyNSeC"}}