From ByteTrack to MCMOT: A Study of Multi-Category-Multi-Objective Tracking Methods

This paper investigates the multi-category multi-target tracking problem, focuses on online multi-target tracking methods, analyzes the advantages and disadvantages of existing methods, and proposes an improved algorithm, MCMOT, based on ByteTrack.MCMOT solves ByteTrack's deficiencies in complex environments by introducing four improvement strategies: closed-loop lifecycle management, motion state classification processing, refined priority matching, and interactive multi-model motion prediction. Experimental results show that MCMOT outperforms ByteTrack in terms of tracking accuracy, precision, persistence and frame rate.