Nuclei Detection and Segmentation
Localizing and segmenting individual cell nuclei in biomedical images is an important step in many applications, such as tracking cell lineages in developing organisms, disease diagnosis in digital pathology, or the development of new pharmaceuticals. StarDist uses star-convex shape representations to detect and segment cell nuclei, which is particularly effective in crowded and noisy imaging conditions, as is very common in biological imaging experiments. StarDist has been extremely well received by the bioimage analysis community due to its ease of use, comprehensive documentation and examples, and training opportunities.
This proposal aims to maintain and extend the functionality of the StarDist napari plugin. In particular, the project will ensure the plugin’s compatibility with the latest release of napari by setting up continuous integration to automate testing. The group will also improve plugin documentation and user support by extending the documentation pages and GitHub issue templates, as well as providing user support in the official forum, image.sc.
The project will also make several significant plugin extension updates, including supporting prediction on time-lapse images, adding a field of view prediction for large images, and improving multithreading support (progress bar, cancelation of running processes). The extension will also expose core library functionality (image normalization, CPU/GPU selection). StarDist has proven to be an increasingly useful tool for life science researchers, and these efforts will help to preserve and increase the practical value of StarDist to the bioimage community.