cellcast 0.2.0

A recast of cell segmentation models built on the Burn deep learning framework.

cellcast_python

This crate contains the Python bindings (via PyO3) for the cellcast core Rust library. Cellcast is a recast of cell segmentation models built on the Burn tensor and deep learning framework. The goal of this project is to modernize (i.e. recast) established cell segmentation models with a WebGPU backend. Cellcast aims to make access to cell segmentation models easy and reproducible.

Installation

Requirements

The cellcast Python package currently supports the following architectures:

Operating System	Architecture
Linux	x86-64, arm64
macOS	intel, arm64
Windows	x86-64

Cellcast is compatible with Python >=3.7 and requires only NumPy.

cellcast from PyPI

You can install the cellcast Python package from PyPI with:

$ pip install cellcast

Build cellcast_python from source

To build the cellcat_python package from source, use the maturin build tool (this requires the Rust toolchain). If you're using uv to manage your Python virtual environments (venv) add numpy and maturin to your environment and run the maturin develop command in the cellcast_python directory of the cellcast repository with your venv activated:

$ source ~/path/to/myenv/.venv/bin/activate
$ (myenv) cd cellcast_python
$ maturin develop

Alernatively if you're using conda or mamba you can do the following:

$ cd cellcat_python
$ mamba activate myenv
(myenv) $ mamba install numpy maturin
...
(myenv) $ maturin develop

This will compile a non-optimized cellcast binaries. Pass the --release flag to compile optimized binaries (note that compilation time may take upwards of 10 minutes).

Usage

Using cellcast

Once cellcast has been installed, cellcast will be available to import. The example below demonstrates how to use cellcast and the StarDist 2D versatile fluo segmentation model with Python. Note that this example assumes you have access to 2D data and tifffile installed in your Python environment with cellcast:

import cellcast.models.stardist_2d as sd
from tifffile import imread

# load 2D data for inference
data_2d = imread("path/to/data_2d.tif")

# run stardist inference and produce instance segmentations
labels = sd.predict_versatile_fluo(data, gpu=True)

Run help() on the predict_versatile_fluo() function to see the full function signature and default values.

License

Cellcast itself is a dual-licensed project with your choice of:

• MIT License (see LICENSE-MIT)
• The Unlicense (see LICENSE-UNLICENSE)

These licenses only apply to the cellcast project and do not apply to the individual models supported by cellcast. You can find each model's associated license listed in the MODEL-LICENSES file.