Cochlea SynapSeg: a napari plugin for cochlear synapse segmentation

Cochlea-SynapSeg 0.1.1

The latest updates to Cochlea-SynapSeg, a Napari plugin for segmenting cochlear ribbon synapses in 3D fluorescence image stacks.

SynapSeg provides tools for pre-processing, ground truth annotation, deep learning-based prediction, and post-analysis of synapse pairings. Built within Napari, it supports both manual and automated workflows, from image preprocessing to montage creation and export for figures.

For installation instructions and documentation, visit the project README.

Major Updates

Prediction Module Added

• Enabled deep learning-based ribbon segmentation prediction.
• Added pre-trained model path handling via Git LFS.
• Options to tune (and visualize tuning parameters) segmentation.

Ground Truth Labeling Enhancements

• Customizable channel-splitting for imported tiffs.
• Implemented auto peak detection for faster annotation.
• Z-scaling for isotropic viewing.

Post-processing tools

• Identify orphans vs synapse pairs.
• Generate and export montages.

Other

• Generate simulated sample images.
• Reorganize widget into scrollable tabs.
• Auto-refresh of layer dropdowns.

Full Changelog: https://github.com/ucsdmanorlab/cochlea-synapseg/compare/v0.1.0...v0.1.1