CellposeSAM trained model for long term evolution experiments under resource deprivation
Year: 2026
Type: software
Abstract: Trained CellposeSAM model for phase-contrast cell segmentation in long-term evolution experiments under resource deprivation conditions, including hypoxia, hypoglycemia, and hypophosphatemia. The checkpoint is a PyTorch ZIP-format CellposeSAM model archive and is intended for use with cloneidR segmentation workflows. The associated training and testing data included: 15 SUM-159 phase-contrast 10x raw image fields with associated Cellpose training annotations (_seg.npy) 50 curated crop/mask pairs from multiple cell-line contexts: HGC-27 (9), MDA-MB-231 (12), SNU-668 (15), and SUM-159 (14) The SUM-159 training images include 2N and 4N conditions across 200, 400, and 800 resource-level labels and T4/T11 timepoints. The held-out test set included 12 SUM-159 4N phase-contrast 10x image fields with associated _seg.npy annotations. Test images covered 400 and 800 resource-level labels across T4, T5, and T11 timepoints. Uploaded file: cellposeSAM_train_2025_09_30_20250930_165546 File size: 609,382,714 bytes SHA256: 7516b5265e5c82f5ea66b0a2e720934c20e7448cbcde3aac63f4b8ac90a00876 MD5: a318ee9f7fd147709e62ec3a1cca4c1f
Authors: Kumar, Rikhil, Andor, Noemi
Institution: Moffitt Cancer Center
Language: English
Subfield: Biophysics
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7167090034