• English | 中文 | 日本語

Seg2Link is a napari-based software specifically designed for scientific research. The software aims to tackle a focused problem: offering an efficient toolbox for quick manual refinement of automated segmentation in large-scale 3D cellular images, particularly useful for brain images obtained through electron microscopy."

Our extensive documentation offers step-by-step tutorials, and our academic paper delves into the scientific methodology and validation behind the software.

Unlike other segmentation solutions, Seg2Link requires pre-processed predictions of cell/non-cell regions as inputs. These predictions can conveniently be generated using Seg2linkUnet2d (Documentation). This integrated approach makes the segmentation process both accurate and efficient.

• Utilizing Deep Learning Predictions -- Seg2Link takes deep learning predictions as input and refines initial inaccurate predictions into highly accurate results through semi-automatic user operations.

• User-Friendly -- Seg2Link not only auto-generates segmentation results but also allows for easy inspection and manual corrections through minimal mouse and keyboard interactions. It supports features like cell ordering, multiple-step undo and redo.

• Efficiency -- Seg2Link is engineered for the rapid processing of large 3D images with billions of voxels.

• Install Anaconda or Miniconda
• Create a new conda environment and activate it by:

conda create -n seg2link-env python=3.8 pip
conda activate seg2link-env

• Install seg2link from this repository:

pip install git+https://github.com/Mohinta2892/Seg2Link.git

Editable install:

pip install -e git+https://github.com/Mohinta2892/Seg2Link.git#egg=Seg2Link

• Activate the created environment by:

conda activate seg2link-env

• Start the software

seg2link

If you used this package in your research please cite it:

• Wen, C., Matsumoto, M., Sawada, M. et al. Seg2Link: an efficient and versatile solution for semi-automatic cell segmentation in 3D image stacks. Sci Rep 13, 7109 (2023). https://doi.org/10.1038/s41598-023-34232-6

If you use this fork in your research, we would be grateful if you acknowledge it:

• Mohinta, S. https://github.com/Mohinta2892/Seg2Link.git