CellSAM Documentation

Welcome to the CellSAM documentation!

CellSAM is a universal model for cell segmentation. CellSAM achieves state-of-the-art performance on segmentation across a variety of cellular targets (bacteria, tissue, yeast, cell culture, etc.) and imaging modalities (brightfield, fluorescence, phase, multiplexed, etc.).

See the preprint for more information about the model and the general approach to image segmentation.

Installation

pip install git+https://github.com/vanvalenlab/cellSAM.git

Basic Usage

The primary interface is cellsam_pipeline(), which implements the inference pipeline for the CellSAM model. See the Gallery for examples of CellSAM applied to different types of images.

Citation

@article{israel2023foundation,
  title={A Foundation Model for Cell Segmentation},
  author={Israel, Uriah and Marks, Markus and Dilip, Rohit and Li, Qilin and Schwartz, Morgan and Pradhan, Elora and Pao, Edward and Li, Shenyi and Pearson-Goulart, Alexander and Perona, Pietro and others},
  journal={bioRxiv},
  publisher={Cold Spring Harbor Laboratory Preprints},
  doi = {10.1101/2023.11.17.567630},
}