Napari-GUI for manual tracklets refinement

This project aims to develop a napari-based GUI for interactive correction of pose-estimation tracks within the movement framework. It will enable users to resolve identity swaps in multi-animal datasets and correct keypoint errors efficiently. In multi-animal tracking, automated pipelines often fail due to occlusions, noise, and identity ambiguity. The proposed tool integrates visualization and editing into a unified workflow, allowing users to inspect and refine trajectories directly while maintaining standardized data structures. This will improve the quality and reliability of pose-estimation datasets, providing the open-source community with a robust and extensible solution for behavioral data curation. This project builds on an initial napari-based prototype that I have already developed for manual trajectory correction.