Multi-object tracking (MOT) links detections over time so 'car 7' stays 'car 7' as it moves and is briefly hidden. It is what lets a system count unique people, measure dwell time, or trace a path, rather than re-detecting strangers each frame.
The book · Volume 6 of 7
AI for Video: A Complete Guide to Intelligence in Video Products: From Models and Inference to Computer Vision, Speech, Search, and Generative Video
Multi-object tracking is what turns per-frame detections into counts, dwell times and paths, and ID switches quietly corrupt all three. This volume covers multi-object tracking in depth: detection and tracking with DeepSORT and ByteTrack, re-identification, occlusion handling, and deployment topologies for edge and server. Written by Nikolay Sapunov, CEO at Fora Soft.
Need multi-object tracking that keeps IDs stable in crowded scenes?
Send us a clip of your scenes, and we will benchmark trackers on it and recommend one with measured ID switches, accuracy and frame rate. 250+ video products shipped since 2005.
