Beyond MOT: Semantic Multi-Object Tracking
Yunhao Li, Qin Li, Hao Wang, Xue Ma, Jiali Yao, Shaohua Dong, Heng Fan, Libo Zhang
TL;DR
This work expands multi-object tracking beyond predicting trajectories to include trajectory-associated semantics, coining Semantic Multi-Object Tracking (SMOT). It introduces BenSMOT, a large-scale, human-centric benchmark with 3,292 videos and rich annotations for trajectories, instance captions, interactions, and video captions, enabling end-to-end SMOT research. The authors also propose SMOTer, an end-to-end baseline that fuses video-level and trajectory-level features to jointly predict trajectories and semantic outputs. Experimental results show SMOTer achieves competitive tracking performance while delivering superior semantic understanding compared to two-stage baselines, illustrating the viability and value of joint "where" and "what" modeling for video understanding. By releasing BenSMOT and SMOTer, the work provides a concrete platform and baseline to drive future research in SMOT and related vision-language tracking tasks.
Abstract
Current multi-object tracking (MOT) aims to predict trajectories of targets (i.e., ''where'') in videos. Yet, knowing merely ''where'' is insufficient in many crucial applications. In comparison, semantic understanding such as fine-grained behaviors, interactions, and overall summarized captions (i.e., ''what'') from videos, associated with ''where'', is highly-desired for comprehensive video analysis. Thus motivated, we introduce Semantic Multi-Object Tracking (SMOT), that aims to estimate object trajectories and meanwhile understand semantic details of associated trajectories including instance captions, instance interactions, and overall video captions, integrating ''where'' and ''what'' for tracking. In order to foster the exploration of SMOT, we propose BenSMOT, a large-scale Benchmark for Semantic MOT. Specifically, BenSMOT comprises 3,292 videos with 151K frames, covering various scenarios for semantic tracking of humans. BenSMOT provides annotations for the trajectories of targets, along with associated instance captions in natural language, instance interactions, and overall caption for each video sequence. To our best knowledge, BenSMOT is the first publicly available benchmark for SMOT. Besides, to encourage future research, we present a novel tracker named SMOTer, which is specially designed and end-to-end trained for SMOT, showing promising performance. By releasing BenSMOT, we expect to go beyond conventional MOT by predicting ''where'' and ''what'' for SMOT, opening up a new direction in tracking for video understanding. We will release BenSMOT and SMOTer at https://github.com/Nathan-Li123/SMOTer.