A Semantic and Occlusion-Aware GM-PHD Filter

2026-05-19

AI Summary

This work addresses the challenge of delayed and inaccurate target initialization in traditional multi-object tracking methods under high-density and heavily occluded autonomous driving scenarios. To overcome this limitation, the authors propose a Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter that integrates deep semantic information with explicit occlusion awareness. The key innovation lies in the Semantic–Occlusion-Aware (S-OA) birth model, which, for the first time, jointly incorporates environmental semantic priors and explicit occlusion modeling to enable more accurate initialization of new targets. Experimental results on the KITTI dataset and Monte Carlo simulations demonstrate that the proposed approach significantly reduces initialization latency and outperforms or matches the strongest baseline in approximately 70% of test cases, achieving superior performance in terms of OSPA distance and cardinality error metrics.

Abstract

This paper proposes a new birth model including semantic information derived from deep learning to create an occlusion-aware Gaussian Mixture Probability Hypothesis Density (GM-PHD) filter. Unlike prior approaches that rely on simplistic or uniform assumptions, the proposed Semantic-Occlusion Aware (S-OA) birth model defines initialization terms by explicitly considering regions of occlusion and by leveraging semantic information about the environment. This enables the filter to accurately represent where new objects are more likely to appear, thereby improving tracking performance in complex and high-density driving scenarios. The method is evaluated through Monte Carlo simulations and experiments on the KITTI dataset. Performance is assessed by measuring the latency between first detection and track initiation, along with the mean absolute cardinality error and the Optimal Subpattern Assignment (OSPA) metric. Results demonstrate that the S-OA birth model reduces initialization delay in occlusion-heavy settings, matching or outperforming the strongest baseline in approximately 70% of cases. A sensitivity analysis of birth model weights is also provided. Overall, the findings underscore the benefits of integrating occlusion reasoning and semantic priors into Bayesian tracking frameworks for autonomous driving.

Problem

• occlusion
• birth model
• GM-PHD filter
• semantic information
• autonomous driving

Innovation

• semantic-aware tracking
• occlusion handling
• GM-PHD filter
• birth model
• autonomous driving