PSMOT: Online Occlusion-Aware Multi-Object Tracking Exploiting Position Sensitivity

Ranyang Zhao,Xinyan Zhang,Jianwei Zhang

DOI: https://doi.org/10.3390/s24041199

IF: 3.9

2024-02-13

Sensors

Abstract:Models based on joint detection and re-identification (ReID), which significantly increase the efficiency of online multi-object tracking (MOT) systems, are an evolution from separate detection and ReID models in the tracking-by-detection (TBD) paradigm. It is observed that these joint models are typically one-stage, while the two-stage models become obsolete because of their slow speed and low efficiency. However, the two-stage models have naive advantages over the one-stage anchor-based and anchor-free models in handling feature misalignment and occlusion, which suggests that the two-stage models, via meticulous design, could be on par with the state-of-the-art one-stage models. Following this intuition, we propose a robust and efficient two-stage joint model based on R–FCN, whose backbone and neck are fully convolutional, and the RoI-wise process only involves simple calculations. In the first stage, an adaptive sparse anchoring scheme is utilized to produce adequate, high-quality proposals to improve efficiency. To boost both detection and ReID, two key elements—feature aggregation and feature disentanglement—are taken into account. To improve robustness against occlusion, the position-sensitivity is exploited, first to estimate occlusion and then to direct the post-process for anti-occlusion. Finally, we link the model to a hierarchical association algorithm to form a complete MOT system called PSMOT. Compared to other cutting-edge systems, PSMOT achieves competitive performance while maintaining time efficiency.

engineering, electrical & electronic,chemistry, analytical,instruments & instrumentation

What problem does this paper attempt to address?

The paper attempts to address the issues of feature alignment and occlusion in multi-object tracking (MOT) to improve the performance and robustness of online multi-object tracking systems.

Specifically, the paper points out:
1. **Feature Alignment Issue**: Existing single-stage models have deficiencies in handling feature alignment, especially in detection and re-identification (ReID) tasks, where the feature extraction regions do not align with the actual object positions, leading to performance degradation.
2. **Occlusion Issue**: Occlusion is a major challenge in multi-object tracking, causing target detection failures or contamination of re-identification features, which results in tracking drift.

To solve these problems, the paper proposes a position-sensitive two-stage joint model (PSMOT), which combines the advantages of two-stage models and improves the multi-object tracking system through the following methods:
1. **Two-Stage Design**: Utilizing the fully convolutional structure of R-FCN, the model's efficiency is enhanced through lightweight RoI-wise processing.
2. **Adaptive Sparse Anchor Generation**: Replacing the traditional dense anchor generation scheme, reducing computational cost, and generating high-quality proposal boxes.
3. **Feature Aggregation and Decoupling**: Improving the model's performance in multi-task learning through multi-layer feature fusion and task-specific feature decoupling.
4. **Position Sensitivity**: For the first time, applying position sensitivity to occlusion resistance, by estimating occlusion areas and guiding subsequent processing, enhancing the model's robustness.

In summary, this paper aims to improve the performance and robustness of multi-object tracking systems by enhancing the design of two-stage models, particularly in handling feature alignment and occlusion issues.