MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification
Person re-identification (Re-ID) has become a hot topic in both research and industry. We joined in a person Re-ID challenge of the First National Artificial Intelligence Challenge (China, 2019) and found some model designs and training tricks work great or not on a super big private dataset. In thi...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2020-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9052718/ |
| _version_ | 1818917606175801344 |
|---|---|
| author | Hanlin Tan Huaxin Xiao Xiaoyu Zhang Bin Dai Shiming Lai Yu Liu Maojun Zhang |
| author_facet | Hanlin Tan Huaxin Xiao Xiaoyu Zhang Bin Dai Shiming Lai Yu Liu Maojun Zhang |
| author_sort | Hanlin Tan |
| collection | DOAJ |
| description | Person re-identification (Re-ID) has become a hot topic in both research and industry. We joined in a person Re-ID challenge of the First National Artificial Intelligence Challenge (China, 2019) and found some model designs and training tricks work great or not on a super big private dataset. In this paper, we propose a model that combines the most effective designs, including multi-scale, multi-branch and attention mechanism, and report training tricks that are no less or even more important in improving person Re-ID performance. We analyze four commonly used public datasets: Market1501, DukeMTMC-ReID, CUHK03, and MSMT17, and achieve the state-of-the-art performance. Besides, we analyze and confirm the effectiveness of the designs by ablation studies. We also share strategies that play a key role in the challenge and experience of model designs that do not generalize well on large datasets. |
| first_indexed | 2024-12-20T00:36:44Z |
| format | Article |
| id | doaj.art-6d1a3cd658164098abc36ec2715b0b0e |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T00:36:44Z |
| publishDate | 2020-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-6d1a3cd658164098abc36ec2715b0b0e2022-12-21T19:59:44ZengIEEEIEEE Access2169-35362020-01-018636326364210.1109/ACCESS.2020.29849159052718MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-IdentificationHanlin Tan0https://orcid.org/0000-0001-8470-263XHuaxin Xiao1https://orcid.org/0000-0003-4524-2698Xiaoyu Zhang2https://orcid.org/0000-0003-4879-8499Bin Dai3https://orcid.org/0000-0003-1719-2208Shiming Lai4https://orcid.org/0000-0003-1345-2527Yu Liu5https://orcid.org/0000-0002-3914-1252Maojun Zhang6https://orcid.org/0000-0001-6748-0545College of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaCollege of Systems Engineering, National University of Defense Technology, Changsha, ChinaPerson re-identification (Re-ID) has become a hot topic in both research and industry. We joined in a person Re-ID challenge of the First National Artificial Intelligence Challenge (China, 2019) and found some model designs and training tricks work great or not on a super big private dataset. In this paper, we propose a model that combines the most effective designs, including multi-scale, multi-branch and attention mechanism, and report training tricks that are no less or even more important in improving person Re-ID performance. We analyze four commonly used public datasets: Market1501, DukeMTMC-ReID, CUHK03, and MSMT17, and achieve the state-of-the-art performance. Besides, we analyze and confirm the effectiveness of the designs by ablation studies. We also share strategies that play a key role in the challenge and experience of model designs that do not generalize well on large datasets.https://ieeexplore.ieee.org/document/9052718/Person re-identificationconvolutional neural network |
| spellingShingle | Hanlin Tan Huaxin Xiao Xiaoyu Zhang Bin Dai Shiming Lai Yu Liu Maojun Zhang MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification IEEE Access Person re-identification convolutional neural network |
| title | MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification |
| title_full | MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification |
| title_fullStr | MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification |
| title_full_unstemmed | MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification |
| title_short | MSBA: Multiple Scales, Branches and Attention Network With Bag of Tricks for Person Re-Identification |
| title_sort | msba multiple scales branches and attention network with bag of tricks for person re identification |
| topic | Person re-identification convolutional neural network |
| url | https://ieeexplore.ieee.org/document/9052718/ |
| work_keys_str_mv | AT hanlintan msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT huaxinxiao msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT xiaoyuzhang msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT bindai msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT shiminglai msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT yuliu msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification AT maojunzhang msbamultiplescalesbranchesandattentionnetworkwithbagoftricksforpersonreidentification |