Robust Partial Matching for Person Search in the Wild
This work addresses a common but overlooked issue in person search for real-world applications like surveillance, though it is incremental as it builds on existing detection and re-identification methods.
The paper tackles the problem of misaligned bounding boxes in person search due to occlusions and backgrounds by proposing the Align-to-Part Network (APNet), which refines boxes to cover holistic body regions and uses aligned part features for robust partial matching, achieving considerable performance improvement on the new LSPS dataset and competitive results on existing benchmarks.
Various factors like occlusions, backgrounds, etc., would lead to misaligned detected bounding boxes , e.g., ones covering only portions of human body. This issue is common but overlooked by previous person search works. To alleviate this issue, this paper proposes an Align-to-Part Network (APNet) for person detection and re-Identification (reID). APNet refines detected bounding boxes to cover the estimated holistic body regions, from which discriminative part features can be extracted and aligned. Aligned part features naturally formulate reID as a partial feature matching procedure, where valid part features are selected for similarity computation, while part features on occluded or noisy regions are discarded. This design enhances the robustness of person search to real-world challenges with marginal computation overhead. This paper also contributes a Large-Scale dataset for Person Search in the wild (LSPS), which is by far the largest and the most challenging dataset for person search. Experiments show that APNet brings considerable performance improvement on LSPS. Meanwhile, it achieves competitive performance on existing person search benchmarks like CUHK-SYSU and PRW.