Person re-identification via pose-aware multi-semantic learning
Person re-identification (ReID) remains an open-ended research topic, with its variety of substantial applications such as tracking, searching, etc. Existing methods mostly explore the highest-semantic feature embedding, ignoring the insights hidden among the earlier layers. Moreover, owing to the m...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/165561 |
| _version_ | 1824456844801736704 |
|---|---|
| author | Luo, Xiangzhong Duong, Luan H. K. Liu, Weichen |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Luo, Xiangzhong Duong, Luan H. K. Liu, Weichen |
| author_sort | Luo, Xiangzhong |
| collection | NTU |
| description | Person re-identification (ReID) remains an open-ended research topic, with its variety of substantial applications such as tracking, searching, etc. Existing methods mostly explore the highest-semantic feature embedding, ignoring the insights hidden among the earlier layers. Moreover, owing to the misalignment and pose variations, pose-related information is of great significance and needs to be comprehensively utilized. In this paper, we present a novel person ReID framework called Pose-aware Multi-semantic Fusion Network (PMFN). First, taking into account multiple semantics, we propose Multi-semantic Fusion Network (MFN) as the backbone, employing several shortcuts to reserve bypass feature maps for subsequent fusion. Second, to learn a pose-sensitive embedding, pose-aware clues are considered, forming the complete PMFN and investigating the well-aligned global and local body regions. Finally, the center loss is introduced for enhancing the feature discriminability. Exhaustive experiments on two large-scale person ReID benchmarks demonstrate the strengths of our approach over recent state-of-the-art works. |
| first_indexed | 2025-02-19T04:00:34Z |
| format | Conference Paper |
| id | ntu-10356/165561 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T04:00:34Z |
| publishDate | 2023 |
| record_format | dspace |
| spelling | ntu-10356/1655612023-12-15T00:47:16Z Person re-identification via pose-aware multi-semantic learning Luo, Xiangzhong Duong, Luan H. K. Liu, Weichen School of Computer Science and Engineering 2020 IEEE International Conference on Multimedia and Expo (ICME) Parallel and Distributed Computing Centre Engineering::Computer science and engineering Person Re-Identification Multi-Level Semantics Person re-identification (ReID) remains an open-ended research topic, with its variety of substantial applications such as tracking, searching, etc. Existing methods mostly explore the highest-semantic feature embedding, ignoring the insights hidden among the earlier layers. Moreover, owing to the misalignment and pose variations, pose-related information is of great significance and needs to be comprehensively utilized. In this paper, we present a novel person ReID framework called Pose-aware Multi-semantic Fusion Network (PMFN). First, taking into account multiple semantics, we propose Multi-semantic Fusion Network (MFN) as the backbone, employing several shortcuts to reserve bypass feature maps for subsequent fusion. Second, to learn a pose-sensitive embedding, pose-aware clues are considered, forming the complete PMFN and investigating the well-aligned global and local body regions. Finally, the center loss is introduced for enhancing the feature discriminability. Exhaustive experiments on two large-scale person ReID benchmarks demonstrate the strengths of our approach over recent state-of-the-art works. Ministry of Education (MOE) Nanyang Technological University Submitted/Accepted version This work is partially supported by MoE AcRF Tier 2 MOE2019-T2-1-071 and Tier 1 MOE2019-T1-001-072, NTU NAP M4082282 and SUG M4082087, Singapore. 2023-03-31T05:23:34Z 2023-03-31T05:23:34Z 2020 Conference Paper Luo, X., Duong, L. H. K. & Liu, W. (2020). Person re-identification via pose-aware multi-semantic learning. 2020 IEEE International Conference on Multimedia and Expo (ICME). https://dx.doi.org/10.1109/ICME46284.2020.9102719 978-1-7281-1331-9 1945-788X https://hdl.handle.net/10356/165561 10.1109/ICME46284.2020.9102719 en MOE2019-T2-1-071 MOE2019-T1- 001-072 NAP (M4082282) SUG (M4082087) 10.21979/N9/DKN6CN © 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/ 10.1109/ICME46284.2020.9102719. application/pdf |
| spellingShingle | Engineering::Computer science and engineering Person Re-Identification Multi-Level Semantics Luo, Xiangzhong Duong, Luan H. K. Liu, Weichen Person re-identification via pose-aware multi-semantic learning |
| title | Person re-identification via pose-aware multi-semantic learning |
| title_full | Person re-identification via pose-aware multi-semantic learning |
| title_fullStr | Person re-identification via pose-aware multi-semantic learning |
| title_full_unstemmed | Person re-identification via pose-aware multi-semantic learning |
| title_short | Person re-identification via pose-aware multi-semantic learning |
| title_sort | person re identification via pose aware multi semantic learning |
| topic | Engineering::Computer science and engineering Person Re-Identification Multi-Level Semantics |
| url | https://hdl.handle.net/10356/165561 |
| work_keys_str_mv | AT luoxiangzhong personreidentificationviaposeawaremultisemanticlearning AT duongluanhk personreidentificationviaposeawaremultisemanticlearning AT liuweichen personreidentificationviaposeawaremultisemanticlearning |