A Hybrid Approach for Visual Multi-Object Tracking

2025-10-28

AI Summary

This work addresses the challenging problem of visual multi-object tracking (MOT) under nonlinear dynamics, where the number of targets varies over time and target identities are prone to ambiguity. To tackle this, we propose a novel tracking framework that synergistically integrates stochastic and deterministic mechanisms. Methodologically, we innovatively couple particle filtering with particle swarm optimization (PSO) to jointly model motion, appearance, and social interaction cues for adaptive data association. We further introduce an identity-preserving state smoothing update, a trend-guided velocity regression strategy, detection confidence weighting, trajectory penalty, and spatial consistency constraints. Our approach achieves significant improvements over state-of-the-art methods across multiple standard benchmarks. The implementation is fully open-sourced and supports both pre-recorded videos and real-time streaming input. Extensive experiments demonstrate superior robustness and identity consistency in complex, dynamic scenes.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Tracking, profiling, and countermeasures against them
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

This paper proposes a visual multi-object tracking method that jointly employs stochastic and deterministic mechanisms to ensure identifier consistency for unknown and time-varying target numbers under nonlinear dynamics. A stochastic particle filter addresses nonlinear dynamics and non-Gaussian noise, with support from particle swarm optimization (PSO) to guide particles toward state distribution modes and mitigate divergence through proposed fitness measures incorporating motion consistency, appearance similarity, and social-interaction cues with neighboring targets. Deterministic association further enforces identifier consistency via a proposed cost matrix incorporating spatial consistency between particles and current detections, detection confidences, and track penalties. Subsequently, a novel scheme is proposed for the smooth updating of target states while preserving their identities, particularly for weak tracks during interactions with other targets and prolonged occlusions. Moreover, velocity regression over past states provides trend-seed velocities, enhancing particle sampling and state updates. The proposed tracker is designed to operate flexibly for both pre-recorded videos and camera live streams, where future frames are unavailable. Experimental results confirm superior performance compared to state-of-the-art trackers. The source-code reference implementations of both the proposed method and compared-trackers are provided on GitHub: https://github.com/SDU-VelKoTek/GenTrack2

Problem

• Tracking multiple objects with unknown and changing counts under nonlinear dynamics
• Maintaining identifier consistency during interactions and occlusions
• Operating flexibly for both recorded videos and live camera streams

Innovation

• Hybrid stochastic-deterministic tracking for dynamic targets
• Particle swarm optimization with multi-cue fitness measures
• Velocity regression and smooth state update for identity preservation