Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals
At present, person reidentification based on attention mechanism has attracted many scholars’ interests. Although attention module can improve the representation ability and reidentification accuracy of Re-ID model to a certain extent, it depends on the coupling of attention module and original netw...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2021-01-01
|
| Series: | Complexity |
| Online Access: | http://dx.doi.org/10.1155/2021/6673461 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832568488159346688 |
|---|---|
| author | Yongyi Li Shiqi Wang Shuang Dong Xueling Lv Changzhi Lv Di Fan |
| author_facet | Yongyi Li Shiqi Wang Shuang Dong Xueling Lv Changzhi Lv Di Fan |
| author_sort | Yongyi Li |
| collection | DOAJ |
| description | At present, person reidentification based on attention mechanism has attracted many scholars’ interests. Although attention module can improve the representation ability and reidentification accuracy of Re-ID model to a certain extent, it depends on the coupling of attention module and original network. In this paper, a person reidentification model that combines multiple attentions and multiscale residuals is proposed. The model introduces combined attention fusion module and multiscale residual fusion module in the backbone network ResNet 50 to enhance the feature flow between residual blocks and better fuse multiscale features. Furthermore, a global branch and a local branch are designed and applied to enhance the channel aggregation and position perception ability of the network by utilizing the dual ensemble attention module, as along as the fine-grained feature expression is obtained by using multiproportion block and reorganization. Thus, the global and local features are enhanced. The experimental results on Market-1501 dataset and DukeMTMC-reID dataset show that the indexes of the presented model, especially Rank-1 accuracy, reach 96.20% and 89.59%, respectively, which can be considered as a progress in Re-ID. |
| format | Article |
| id | doaj-art-35682c8fd30f419584da5dbf404b1d48 |
| institution | Kabale University |
| issn | 1076-2787 1099-0526 |
| language | English |
| publishDate | 2021-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Complexity |
| spelling | doaj-art-35682c8fd30f419584da5dbf404b1d482025-02-03T00:58:58ZengWileyComplexity1076-27871099-05262021-01-01202110.1155/2021/66734616673461Person Reidentification Model Based on Multiattention Modules and Multiscale ResidualsYongyi Li0Shiqi Wang1Shuang Dong2Xueling Lv3Changzhi Lv4Di Fan5Shandong University of Science and Technology, Qingdao, Shandong 266590, ChinaShandong University of Science and Technology, Qingdao, Shandong 266590, ChinaShandong University of Science and Technology, Qingdao, Shandong 266590, ChinaDepartment of Management and Economics, Tianjin University, Tianjin 300072, ChinaShandong University of Science and Technology, Qingdao, Shandong 266590, ChinaShandong University of Science and Technology, Qingdao, Shandong 266590, ChinaAt present, person reidentification based on attention mechanism has attracted many scholars’ interests. Although attention module can improve the representation ability and reidentification accuracy of Re-ID model to a certain extent, it depends on the coupling of attention module and original network. In this paper, a person reidentification model that combines multiple attentions and multiscale residuals is proposed. The model introduces combined attention fusion module and multiscale residual fusion module in the backbone network ResNet 50 to enhance the feature flow between residual blocks and better fuse multiscale features. Furthermore, a global branch and a local branch are designed and applied to enhance the channel aggregation and position perception ability of the network by utilizing the dual ensemble attention module, as along as the fine-grained feature expression is obtained by using multiproportion block and reorganization. Thus, the global and local features are enhanced. The experimental results on Market-1501 dataset and DukeMTMC-reID dataset show that the indexes of the presented model, especially Rank-1 accuracy, reach 96.20% and 89.59%, respectively, which can be considered as a progress in Re-ID.http://dx.doi.org/10.1155/2021/6673461 |
| spellingShingle | Yongyi Li Shiqi Wang Shuang Dong Xueling Lv Changzhi Lv Di Fan Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals Complexity |
| title | Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals |
| title_full | Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals |
| title_fullStr | Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals |
| title_full_unstemmed | Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals |
| title_short | Person Reidentification Model Based on Multiattention Modules and Multiscale Residuals |
| title_sort | person reidentification model based on multiattention modules and multiscale residuals |
| url | http://dx.doi.org/10.1155/2021/6673461 |
| work_keys_str_mv | AT yongyili personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals AT shiqiwang personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals AT shuangdong personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals AT xuelinglv personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals AT changzhilv personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals AT difan personreidentificationmodelbasedonmultiattentionmodulesandmultiscaleresiduals |