{"success":true,"data":{"author":"Huihui Fang, Fei Li, Junde Wu, Huazhu Fu et al. (iChallenge PALM); mirrored by MedOtter","bucketAccessType":"public","bucketName":"MedOtter/iChallenge-PALM19","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-21T20:45:56.516Z","description":"PALM (PAthoLogic Myopia) is the iChallenge 2019 dataset from Zhongshan Ophthalmic Center, Sun Yat-sen University: 1,200 color fundus photographs of the left eyes of 720 myopia clinic patients, taken on a Zeiss Visucam 500 (2124x2056) or Canon CR-2 (1444x1444). Each image is labeled pathologic myopia or not (637 versus 563), and pixel masks are given for the optic disc (1,144 images), patchy retinal atrophy (870) and retinal detachment (24). Splits are patient-disjoint with 400 images each in train, validation and test. Images keep their original JPEG bytes and masks are binary PNG, stored as Parquet under data/, about 1.7 GB (measured 2026-11-21); read with Hugging Face datasets or pandas plus Pillow. Use for pathologic myopia classification, optic disc segmentation, retinal atrophy lesion segmentation and fovea localization. Mirrored by MedOtter from the authors' figshare deposit, released under CC0 1.0 with DOI 10.6084/m9.figshare.21299148.v1; cite Fang et al. and the figshare deposit.","displayTitle":"PALM (PAthoLogic Myopia, iChallenge 2019): 1,200 fundus photographs with optic disc, atrophy and detachment masks","doi":"10.6084/m9.figshare.21299148.v1","huggingFaceUrl":"https://huggingface.co/datasets/MedOtter/iChallenge-PALM19","id":"4143aec5-5464-5022-b410-627f8d56cad8","license":"CC0-1.0","licenseEvidenceUrl":"https://figshare.com/articles/dataset/21299148","objectCountEstimate":12,"probeStatus":"ok","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/iChallenge-PALM19\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MedOtter/iChallenge-PALM19\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MedOtter, license:cc0-1.0, size_categories:1K<n<10K, format:parquet, downloads=656"],"format":"mixed","generatedAt":"2026-11-21T20:48:43.320Z","layout":{"example":"data/test-00000-of-00003.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"data/test-00000-of-00003.parquet","truncatedAt":512}],"sampleObjects":["data/test-00000-of-00003.parquet","data/test-00001-of-00003.parquet","data/test-00002-of-00003.parquet","data/train-00000-of-00003.parquet","data/train-00001-of-00003.parquet"],"sidecars":[{"bytes":"---\nlicense: cc0-1.0\ntask_categories:\n  - image-segmentation\ntags:\n  - medical\n  - retinal\n  - fundus\n  - ophthalmology\n  - pathologic-myopia\n  - optic-disc\n  - lesion-segmentation\npretty_name: iChallenge-PALM19 (PALM - PAthoLogic Myopia)\nsize_categories:\n  - 1K<n<10K\n---\n\n# iChallenge-PALM19 — PALM: PAthoLogic Myopia\n\n**1200 color fundus photographs** from **720 myopia-clinic subjects** (left eyes\nonly) at Zhongshan Ophthalmic Center, Sun Yat-sen University — the PALM dataset\nfrom the **ISBI 2019 iChallenge** satellite event, with pixel-level annotation of\nthree targets.\n\n- **Modality:** color fundus photography (2D RGB)\n- **Organ:** retina / eye\n- **Resolutions:** 2124×2056 (Zeiss Visucam 500, n=1047) · 1444×1444 (Canon CR-2, n=153)\n- **Targets:** optic disc · patchy retinal atrophy · retinal detachment\n- **License:** **CC0 1.0** (public domain dedication)\n- **Source:** [figshare 21299148](https://doi.org/10.6084/m9.figshare.21299148.v1) (official, author-deposited)\n\n## Splits\n\nPatient-disjoint (images from one patient stay in one split). All counts below\nwere re-verified on the archive and match the paper's Table 1 exactly.\n\n| Split | Images | PM / non-PM | `disc_mask` | `atrophy_mask` | `detachment_mask` |\n|---|---|---|---|---|---|\n| `train` (Training) | 400 | 213 / 187 | 381 | 311 | 12 |\n| `validation` (Validation) | 400 | 211 / 189 | 379 | 271 | 6 |\n| `test` (Testing) | 400 | 213 / 187 | 384 | 288 | 6 |\n| **Total** | **1200** | **637 / 563** | **1144** | **870** | **24** |\n\n## Columns\n\n| Column | Type | Notes |\n|---|---|---|\n| `image_id` | string | Stem, e.g. `H0001`, `V0001`, `T0001` — the pairing key |\n| `image` | Image (RGB) | Fundus photograph. **Original JPEG bytes, stored verbatim** (no re-encode) |\n| `disc_mask` | Image (L), **nullable** | Optic disc. `{0,255}`, 255 = disc |\n| `atrophy_mask` | Image (L), **nullable** | Patchy retinal atrophy. `{0,255}`, 255 = lesion |\n| `detachment_mask` | Image (L), **nullable** | Retinal detachment. `{0,255}`, 255 = lesion |\n| `has_disc_mask` / `has_atrophy_mask` / `has_detachment_mask` | bool | Explicit presence flag — see *Masks are absent, not empty* |\n| `pm_label` | int32 | `1` = pathologic myopia, `0` = non-PM |\n| `fovea_x`, `fovea_y` | float32 | Fovea centre. `(0,0)` upstream means *not visible* |\n| `fovea_visible` | bool | False exactly when the coordinate is `(0,0)` |\n| `device` | string | `zeiss` \\| `canon` |\n| `shooting_center` | string | `optic_disc` \\| `fovea` \\| `midpoint` |\n| `width`, `height` | int32 | Image size (masks always match) |\n| `disc_fraction`, `atrophy_fraction`, `detachment_fraction` | float32, nullable | Foreground area fraction |\n| `id_prefix` | string | `H`/`N`/`P` (train) or `V`/`T` — **leaks the class in `train`**, see below |\n\n## ⚠️ Mask polarity was normalised here\n\n**Upstream PALM masks are inverted**: `0 = foreground`, `255 = background`.\n**Every mask in this repository has been inverted so that `255 = foreground`.**\nUse the ordinary `mask > 0`.\n\nThis was verified rather than assumed: all 2038 source masks decode to exactly\nthe value set `{0, 255}`, and polarity was confirmed by a corner test — all 8152\nframe corners across all masks are `255`, none is `0`, and the disc's `0`-region\nnever touches the frame border in any of 1144 masks.\n\n> There is **no `128` value in PALM.** The `0`/`128`/`255` (cup/rim/background)\n> convention belongs to **REFUGE**, a sibling iChallenge dataset, and must not be\n> applied here.\n\n## ⚠️ Masks are absent, not empty\n\n56 images have no disc mask, 330 no atrophy, 1176 no detachment. In this 2024\nrelease those files are **missing from the archive** — there are zero\nall-background and zero all-foreground masks among all 2038. Absence is stored as\na **null** mask plus a `has_*_mask` boolean; never infer it from an empty mask.\n\n*(The 2019 challenge release instead shipped all-255 empty disc masks for the 56\nno-OD images. That trap does not apply to this data.)*\n\n## ⚠️ The three targets overlap spatially\n\nMeasured on the 24 detachment cases: **14/24 intersect atrophy** (up to 54 % of\nt","encoding":"utf8","key":"README.md","name":"README.md","size":8180,"truncated":true}],"sourceObjectKey":"data/test-00000-of-00003.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/iChallenge-PALM19\")\nprint(ds)","starterSamplePath":"data/test-00000-of-00003.parquet","totalObjects":12},"summary":"PALM (PAthoLogic Myopia) is the iChallenge 2019 dataset from Zhongshan Ophthalmic Center, Sun Yat-sen University: 1,200 color fundus photographs of the left eyes of 720 myopia…","title":"PALM (PAthoLogic Myopia, iChallenge 2019): 1,200 fundus photographs with optic disc, atrophy and detachment masks","updatedAt":"2026-11-21T21:55:12.947Z","visibility":"public","websiteUrl":"https://doi.org/10.6084/m9.figshare.21299148.v1"}}