A napari plugin for cellulus

• Introduction
• Installation
• Getting Started
• Citation
• Issues

Introduction

This repository hosts the code for the napari plugin built around cellulus, which was described in the preprint titled Unsupervised Learning of Object-Centric Embeddings for Cell Instance Segmentation in Microscopy Images.

cellulus is a deep learning based method which can be used to obtain instance-segmentation of objects in microscopy images in an unsupervised fashion i.e. requiring no ground truth labels during training.

The main source repository for cellulus lives here.

Installation

One could execute these lines of code below to create a new environment and install dependencies.

1. Create a new environment called napari-cellulus:

conda create -y -n napari-cellulus python==3.9

2. Activate the newly-created environment:

conda activate napari-cellulus

3a. If using a GPU, install pytorch cuda dependencies:

conda install pytorch==2.0.1 torchvision==0.15.2 pytorch-cuda=11.7 -c pytorch -c nvidia

3b. otherwise (if using a CPU or MPS), run:

pip install torch torchvision

4. Install the package from github:

pip install git+https://github.com/funkelab/napari-cellulus.git

Getting Started

Run the following commands in a terminal window:

conda activate napari-cellulus
napari

demo_cellulus.webm

Citation

If you find our work useful in your research, please consider citing:

@misc{wolf2023unsupervised,
      title={Unsupervised Learning of Object-Centric Embeddings for Cell Instance Segmentation in Microscopy Images},
      author={Steffen Wolf and Manan Lalit and Henry Westmacott and Katie McDole and Jan Funke},
      year={2023},
      eprint={2310.08501},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Issues

If you encounter any problems, please file an issue along with a description.