Awesome

CellSAM: A Foundation Model for Cell Segmentation

<img src="https://github.com/rdilip/cellsam_inference/assets/9993319/1f40b5a5-60f1-4980-997d-6059e20d6133" alt="Try the demo!" style="width: 100%;">

Description

This repository provides inference code for CellSAM. CellSAM is described in more detail in the preprint, and is publicly deployed at cellsam.deepcell.org. 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, etc.). Feel free to reach out for support/questions! The full dataset used to train CellSAM is available here.

Getting started

The easiest way to get started with CellSAM is with pip pip install git+https://github.com/vanvalenlab/cellSAM.git

CellSAM requires python>=3.10, but otherwise uses pure PyTorch. A sample image is included in this repository. Segmentation can be performed as follows

import numpy as np
from cellSAM import segment_cellular_image
img = np.load("sample_imgs/yeaz.npy")
mask, _, _ = segment_cellular_image(img, device='cuda')

For more details, see cellsam_introduction.ipynb.

Napari package

CellSAM includes a basic napari package for annotation functionality. To install the additional napari dependencies, use pip.

pip install git+https://github.com/vanvalenlab/cellSAM.git#egg=cellsam[napari]

To launch the napari app, run cellsam napari.

Citation

Please cite us if you use CellSAM.

@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},
}