Hungarian-matching pipeline combining IoU, centroid distance, and class cues for multi-object tracking.
2025
Tech stack
Problem
Object tracking pipeline for the VIRAT dataset using the Hungarian algorithm with a composite cost function: IoU overlap, centroid Euclidean distance, and class consistency cues.
Approach
Generates side-by-side match visualizations and CSV reports per frame. Evaluated across multiple detection confidence thresholds..
Result
See the code for full results.
Full writeup
Object tracking pipeline for the VIRAT dataset using the Hungarian algorithm with a composite cost function: IoU overlap, centroid Euclidean distance, and class consistency cues. Generates side-by-side match visualizations and CSV reports per frame. Evaluated across multiple detection confidence thresholds.