Structure-Aware Hierarchical Multiple-Object Tracking for Three-Dimensional Nanoparticle Trajectory Reconstruction

SAH-PSF is the research code associated with the manuscript "Structure-Aware Hierarchical Multiparticle Tracking for Three-Dimensional Nanoparticle Trajectory Reconstruction". It detects and tracks optimized ring-spliced symmetric exponential point-spread-function (RSSE-PSF) patterns in fluorescence microscopy sequences. The method treats each complete double-lobe pattern as a particle-level parent target and its two lobes as child targets, allowing optical structure and temporal identity to be modeled together. The repository is based on a modified Ultralytics 8.3.89 codebase. The customized detection-stage structural processing and tracking callbacks are implemented in ; is the tracking entry point and exports results in MOTChallenge format.ultralytics/trackers/track.pytrack_MOT.py