einarolafsson/live-cell-segmentation-dataset

Light-microscopy cell segmentation dataset

Transmitted-light fields -- phase contrast, brightfield, DIC and quantitative phase -- with uint16 instance masks, from 14 public datasets, curated for spaCR and used to fine-tune live-cell-segmentation-cpsam.

11,007 fields: 6,778 train, 2,030 valid, 2,199 test. Split by ACQUISITION -- a well, a dish, a time-lapse or a z-stack is never divided between sets -- so the test set is fields the model never saw anything of.

Layout

images/<source>/<source>__<stem>.tif   the field
masks/<source>/<source>__<stem>.tif    uint16 instance labels, background 0, cells 1..N
splits.csv                      name, source, modality, group, split, n_objects
manifest.csv                    provenance and statistics for every field
SOURCES.md                      citation and licence for every source

Retrain

from huggingface_hub import snapshot_download
root = snapshot_download("einarolafsson/live-cell-segmentation-dataset", repo_type="dataset")
# images/ and masks/ share names, one folder per source (Hugging Face caps a folder at
# 10,000 files); splits.csv says which split each field belongs to.

Licence

Mixed -- each source keeps its own terms, listed in SOURCES.md. LIVECell and Revvity-25 are CC BY-NC 4.0, so this dataset as a whole is non-commercial. Cite every source you use.