Deep LG-Track: An Enhanced Localization-Confidence-Guided Multi-Object Tracker

2025-04-02

AI Summary

Multi-object tracking (MOT) is critical for autonomous driving and intelligent surveillance, yet existing methods suffer from insufficient accuracy and robustness in complex scenes. To address this, we propose a confidence-guided enhanced tracking framework. First, we introduce an adaptive Kalman filter that dynamically optimizes motion model parameters based on both localization and detection confidence scores. Second, we construct a motion-appearance fusion cost matrix weighted by detection confidence to improve association reliability. Third, we design a dynamic appearance feature update mechanism jointly driven by image clarity and localization precision, effectively mitigating interference from occlusion and object deformation. Evaluated on the MOT17 and MOT20 benchmarks, our method achieves state-of-the-art performance across key metrics—including MOTA, IDF1, and HOTA—demonstrating significant improvements in tracking accuracy, identity consistency, and generalization capability.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Security and Privacy: Tracking, profiling, and countermeasures against them
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Multi-object tracking plays a crucial role in various applications, such as autonomous driving and security surveillance. This study introduces Deep LG-Track, a novel multi-object tracker that incorporates three key enhancements to improve the tracking accuracy and robustness. First, an adaptive Kalman filter is developed to dynamically update the covariance of measurement noise based on detection confidence and trajectory disappearance. Second, a novel cost matrix is formulated to adaptively fuse motion and appearance information, leveraging localization confidence and detection confidence as weighting factors. Third, a dynamic appearance feature updating strategy is introduced, adjusting the relative weighting of historical and current appearance features based on appearance clarity and localization accuracy. Comprehensive evaluations on the MOT17 and MOT20 datasets demonstrate that the proposed Deep LG-Track consistently outperforms state-of-the-art trackers across multiple performance metrics, highlighting its effectiveness in multi-object tracking tasks.

Problem

• Enhances multi-object tracking accuracy and robustness
• Adapts Kalman filter for dynamic noise covariance updates
• Improves motion-appearance fusion with confidence-based weighting

Innovation

• Adaptive Kalman filter for dynamic noise covariance
• Novel cost matrix fusing motion and appearance
• Dynamic appearance feature updating strategy