Skip to content

MedShapeNet: 106 101 anatomical and surgical-instrument shapes as STL meshes

pl.load("medshapenet") loads this dataset. The MedShapeNet collection (IKIM, University Hospital Essen): over 100 000 3-D shapes extracted from 23 medical datasets - organs, bones, vessels, tumours, teeth, instruments - as STL files, one per URL in the public download index (sciebo, HTTP Range supported). Sample ids are case/structure (426 structure names, e.g. liver, vertebrae, skull, tumoredbrain); ds.filter(structure='liver') selects one anatomy. STL carries no unit: shapes derived from medical images are in millimetres (declared here, TO CONFIRM for the instrument sets).

import pardesline as pl

ds = pl.load("medshapenet")
print(ds.describe())  # roles, frames, units, labels, licence, citation
s = ds[0]
obj = s.mesh  # fetched now, cached under ~/.cache/pardesline/datasets/medshapenet

Provenance and licence
Source URL index https://mml.ikim.nrw/ait/projects/medshapenet/MedShapeNetDataset.txt
Scheme urllist
Size of the whole source 1953.1 GB
Licence CC-BY-4.0
Redistribution unknown
Licence note MedShapeNetCore on Zenodo is CC BY 4.0 (checked 2026-10-14); the full collection inherits the licence of each source dataset, TO CONFIRM per structure before commercial use or redistribution.
Terms https://medshapenet.ikim.nrw/
Homepage https://medshapenet.ikim.nrw/
Cite Li J., Zhou Z., Yang J., et al. MedShapeNet - A Large-Scale Dataset of 3D Medical Shapes for Computer Vision. Biomedical Engineering / Biomedizinische Technik, 2024. arXiv:2308.16139.
Samples grouped by case

Roles (what each sample contains)

Role Type Frame Units Labels / layout Pattern or file
mesh mesh from the file header mm {case}_{structure}.stl
Triangle mesh of one anatomical structure (STL, millimetres)

Standardized layout

Files are fetched on first access into the dataset cache and stay in their original format; ds.export(folder) writes every sample as standard files (PLY for points and meshes, MetaImage for volumes and label maps) with a .pardesline.json sidecar carrying frame, units, labels, licence and provenance, plus a manifest for the folder. Label maps become exact surfaces with to_mesh(label=k), DICOM series become one volume, raw arrays are decoded through the layout above.

ds = pl.load("medshapenet").take(2)
ds.export("./standardized")        # PLY / MetaImage + sidecars + manifest.json

Algorithms and models that run on it

Generated from src/pardesline/data/manifests/medshapenet.json on 2026-10-14.