Nuclei segmentation dataset from the 2018 Data Science Bowl (BBBC038v1, Broad Bioimage Benchmark Collection). 670 microscopy images with 29,442 individually segmented nuclei as polygon annotations, converted from the official per-nucleus masks. Spans fluorescence and brightfield/histology imagery across multiple organisms and staining protocols. License: CC0 1.0 (public domain). Source: https://bbbc.broadinstitute.org/BBBC038. Please cite: Caicedo et al., Nucleus segmentation across imaging experiments: the 2018 Data Science Bowl. Nature Methods 16, 1247-1253 (2019).

Two-channel cell segmentation dataset of Drosophila melanogaster Kc167 cells (BBBC007v1, Broad Bioimage Benchmark Collection). 16 fields of view, each as a DNA image and an actin image (32 grayscale 8-bit images, 400 to 512 px), with 1,264 hand-outlined nuclei and 1,221 hand-outlined cells converted from the official outline rasters to polygon annotations. The benchmark task is finding the boundaries between touching cells, using the nuclei as seeds. License: CC0 1.0 (public domain). Source: https://bbbc.broadinstitute.org/BBBC007. Please cite: Jones, Carpenter & Golland, Voronoi-based segmentation of cells on image manifolds, Proc. ICCV Workshop on Computer Vision for Biomedical Image Applications (CVBIA), 2005; and Ljosa, Sokolnicki & Carpenter, Annotated high-throughput microscopy image sets for validation, Nature Methods 9, 637 (2012).

Nucleus segmentation dataset of U2OS cells from a chemical screen (BBBC039v1, Broad Bioimage Benchmark Collection). 200 fluorescence microscopy fields (520x696, 16-bit TIFF, Hoechst-stained nuclei) spanning 200 bioactive-compound treatments, with 19,379 individually segmented nucleus instances decoded from the official per-pixel class masks (interior/boundary) and converted to polygon annotations (3 fields have no annotated nuclei). Includes the official train (100) / validation (50) / test (50) split. License: CC0 1.0 (public domain). Source: https://bbbc.broadinstitute.org/BBBC039. Please cite: Caicedo et al. 2018, image set BBBC039v1, available from the Broad Bioimage Benchmark Collection [Ljosa et al., Nature Methods, 2012].

Simpel dataset for instance segmentation containing 2 classes