Sequential End-to-end Network for Efficient Person Search

Abstract

Person search aims at jointly solving Person Detection and Person Re-identification (re-ID). Existing works have designed end-to-end networks based on Faster R-CNN. However, due to the parallel structure of Faster R-CNN, the extracted features come from the low-quality proposals generated by the Region Proposal Network, rather than the detected high-quality bounding boxes. Person search is a fine-grained task and such inferior features will significantly reduce re-ID performance. To address this issue, we propose a Sequential End-to-end Network (SeqNet) to extract superior features. In SeqNet, detection and re-ID are considered as a progressive process and tackled with two sub-networks sequentially. In addition, we design a robust Context Bipartite Graph Matching (CBGM) algorithm to effectively employ context information as an important complementary cue for person matching. Extensive experiments on two widely used person search benchmarks, CUHK-SYSU and PRW, have shown that our method achieves state-of-the-art results. Also, our model runs at 11.5 fps on a single GPU and can be integrated into the existing end-to-end framework easily.

Grouped Adaptive Loss Weighting for Person Search

“…Chen et al [6] introduce a norm-aware embedding method (NAE) which relieves the contradictory goals of pedestrian detection and person re-id tasks by decomposing the feature embedding into norm and angle respectively. Based on that, SeqNet [24] gets better performance by stacking the NAE models. Kim et al [22] present a prototype-guided attention module to obtain discriminative re-id features.…”

“…In Tab. 8, we compare our method with the state-of-the-art methods, including two-stage methods [5,13,18,23,35,41,45] and one-stage methods [2,4,6,12,17,22,24,25,28,[36][37][38][39].…”

“…7 shows that our GALW achieves excellent performance across various baselines both on anchor-based method NAE [6] and anchor-free methods AlignPS [37] and ROI-AlignPS * [38], which demonstrates the generalization ability of our method. [37] 122 65 SeqNet [24] 178 97 AGWF [17] 145 80 ROI-AlignPS * [38] 127 69 ROI-AlignPS * w/ GALW 127 69…”

Dual-Attention-Driven Multiscale Fusion Object Searching Network for Remote Sensing Imagery

“…However, the network structure of this method is complex, which reduces computational speed. Li et al [47] account for the fact that the detection and re-identification in person searching is a gradual process through two subnetworks for sequential processing, and the contextual information is used to enhance re-identification. Although this method improves the searching speed, it fails to unify detection and identification tasks, and the two-step structure is still too complex.…”

“…2) Comparative Experiments: Because the proposed method is novel, comparison with existing methods is difficult. Therefore, to verify the advancement provided by the proposed method, the proposed method is compared with several exist-ing pedestrian searching algorithms, including AlignPS [1], AlignPS+ [1], Roi-AlignPS [51] and SeqNet [47]. The quantitative results from the two datasets are shown in Table III.…”