Easily Segment Noisy Images
This team will develop a plugin for napari to allow users to train and/or apply previously trained DenoiSeg networks. By integrating the efforts of this translation project into the wider BioImage Model Zoo community, this plugin will standardize deep learning models in bioimaging. This project will also enable straightforward export of segmentation results to other software using labelings, a format tailored to efficiently deal with complex image label data. The implementation of an iterative labeling-finetuning loop (active learning framework) will enable users to improve the quality of dense segmentation results. A key part of this project seeks to expand this process and make it accessible for a variety of users. The team will provide extensive documentation, tutorials, online teaching materials, and guidelines for users and developers. The group will also teach DenoiSeg to users in the context of international teaching events, and to the developer community during hackathons and workshops.