Learning Domain-Specific Features From General Features for Person Re-Identification
Person re-identification (re-id) plays a vital role in surveillance and forensics application. Since the labeled images for person re-id task is limited, the generalization ability of existed person re-id models is poor. On the other hand, images of different classes (pedestrian and non-pedestrian i...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9173660/ |
| _version_ | 1818349952158400512 |
|---|---|
| author | Chengqiu Dai Junying Feng Ruiling Zhou |
| author_facet | Chengqiu Dai Junying Feng Ruiling Zhou |
| author_sort | Chengqiu Dai |
| collection | DOAJ |
| description | Person re-identification (re-id) plays a vital role in surveillance and forensics application. Since the labeled images for person re-id task is limited, the generalization ability of existed person re-id models is poor. On the other hand, images of different classes (pedestrian and non-pedestrian images) share some general features. To this end, this paper aims to improve the performance of person re-id by designing a relearning network which can learn domain-specific features and general features simultaneously. The proposed relearning network consists of a pretrained backbone network which provides the general features, and several attention-based subnetworks that learn domain-specific features from general features of different levels. Besides, we propose a coarse-fine loss to improve the generalization of person re-id model by making full use of the massive labeled non-pedestrian images. Experimental results on the publicly available Market-1501, DukeMTMC-reID and CUHK03 pedestrian re-id datasets demonstrate the effectiveness of the proposed relearning network and coarse-fine loss. |
| first_indexed | 2024-12-13T18:14:07Z |
| format | Article |
| id | doaj.art-c619e66c9204412da14452f9716b64b1 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-13T18:14:07Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-c619e66c9204412da14452f9716b64b12022-12-21T23:35:53ZengIEEEIEEE Access2169-35362020-01-01815538915539810.1109/ACCESS.2020.30186279173660Learning Domain-Specific Features From General Features for Person Re-IdentificationChengqiu Dai0https://orcid.org/0000-0003-3827-0607Junying Feng1Ruiling Zhou2School of Computer and Information Science, Hunan Institute of Technology, Hengyang, ChinaSchool of Economics and Management, Hunan Institute of Technology, Hengyang, ChinaSchool of Computer and Information Science, Hunan Institute of Technology, Hengyang, ChinaPerson re-identification (re-id) plays a vital role in surveillance and forensics application. Since the labeled images for person re-id task is limited, the generalization ability of existed person re-id models is poor. On the other hand, images of different classes (pedestrian and non-pedestrian images) share some general features. To this end, this paper aims to improve the performance of person re-id by designing a relearning network which can learn domain-specific features and general features simultaneously. The proposed relearning network consists of a pretrained backbone network which provides the general features, and several attention-based subnetworks that learn domain-specific features from general features of different levels. Besides, we propose a coarse-fine loss to improve the generalization of person re-id model by making full use of the massive labeled non-pedestrian images. Experimental results on the publicly available Market-1501, DukeMTMC-reID and CUHK03 pedestrian re-id datasets demonstrate the effectiveness of the proposed relearning network and coarse-fine loss.https://ieeexplore.ieee.org/document/9173660/Attentioncoarse-finedomain-specific featuresgeneral features |
| spellingShingle | Chengqiu Dai Junying Feng Ruiling Zhou Learning Domain-Specific Features From General Features for Person Re-Identification IEEE Access Attention coarse-fine domain-specific features general features |
| title | Learning Domain-Specific Features From General Features for Person Re-Identification |
| title_full | Learning Domain-Specific Features From General Features for Person Re-Identification |
| title_fullStr | Learning Domain-Specific Features From General Features for Person Re-Identification |
| title_full_unstemmed | Learning Domain-Specific Features From General Features for Person Re-Identification |
| title_short | Learning Domain-Specific Features From General Features for Person Re-Identification |
| title_sort | learning domain specific features from general features for person re identification |
| topic | Attention coarse-fine domain-specific features general features |
| url | https://ieeexplore.ieee.org/document/9173660/ |
| work_keys_str_mv | AT chengqiudai learningdomainspecificfeaturesfromgeneralfeaturesforpersonreidentification AT junyingfeng learningdomainspecificfeaturesfromgeneralfeaturesforpersonreidentification AT ruilingzhou learningdomainspecificfeaturesfromgeneralfeaturesforpersonreidentification |