CellSAM

CellSAM is a foundation model for cell segmentation that generalizes across cell types, imaging modalities, and experimental conditions. Built on the Segment Anything Model architecture and fine-tuned on a large-scale dataset of over 22 million annotated cells, CellSAM achieves state-of-the-art performance on segmentation benchmarks without per-task fine-tuning.

Published in Nature Methods (2025).

Links:

• Paper
• Code
• Web App