Identity-Guided Spatial Attention for Vehicle Re-Identification
In vehicle re-identification, identifying a specific vehicle from a large image dataset is challenging due to occlusion and complex backgrounds. Deep models struggle to identify vehicles accurately when critical details are occluded or the background is distracting. To mitigate the impact of these n...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-05-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/11/5152 |
| _version_ | 1797596800554631168 |
|---|---|
| author | Kai Lv Sheng Han Youfang Lin |
| author_facet | Kai Lv Sheng Han Youfang Lin |
| author_sort | Kai Lv |
| collection | DOAJ |
| description | In vehicle re-identification, identifying a specific vehicle from a large image dataset is challenging due to occlusion and complex backgrounds. Deep models struggle to identify vehicles accurately when critical details are occluded or the background is distracting. To mitigate the impact of these noisy factors, we propose Identity-guided Spatial Attention (ISA) to extract more beneficial details for vehicle re-identification. Our approach begins by visualizing the high activation regions of a strong baseline method and identifying noisy objects involved during training. ISA generates an attention map to mask most discriminative areas, without the need for manual annotation. Finally, the ISA map refines the embedding feature in an end-to-end manner to improve vehicle re-identification accuracy. Visualization experiments demonstrate ISA’s ability to capture nearly all vehicle details, while results on three vehicle re-identification datasets show that our method outperforms state-of-the-art approaches. |
| first_indexed | 2024-03-11T02:58:08Z |
| format | Article |
| id | doaj.art-e6d1d0bbaa224738a56ad315364956cf |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T02:58:08Z |
| publishDate | 2023-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-e6d1d0bbaa224738a56ad315364956cf2023-11-18T08:33:15ZengMDPI AGSensors1424-82202023-05-012311515210.3390/s23115152Identity-Guided Spatial Attention for Vehicle Re-IdentificationKai Lv0Sheng Han1Youfang Lin2Beijing Key Laboratory of Traffic Data Analysisand Mining, School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, ChinaBeijing Key Laboratory of Traffic Data Analysisand Mining, School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, ChinaBeijing Key Laboratory of Traffic Data Analysisand Mining, School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, ChinaIn vehicle re-identification, identifying a specific vehicle from a large image dataset is challenging due to occlusion and complex backgrounds. Deep models struggle to identify vehicles accurately when critical details are occluded or the background is distracting. To mitigate the impact of these noisy factors, we propose Identity-guided Spatial Attention (ISA) to extract more beneficial details for vehicle re-identification. Our approach begins by visualizing the high activation regions of a strong baseline method and identifying noisy objects involved during training. ISA generates an attention map to mask most discriminative areas, without the need for manual annotation. Finally, the ISA map refines the embedding feature in an end-to-end manner to improve vehicle re-identification accuracy. Visualization experiments demonstrate ISA’s ability to capture nearly all vehicle details, while results on three vehicle re-identification datasets show that our method outperforms state-of-the-art approaches.https://www.mdpi.com/1424-8220/23/11/5152vehicle re-identificationdeep learningmachine learningattention mechanismvehicle details |
| spellingShingle | Kai Lv Sheng Han Youfang Lin Identity-Guided Spatial Attention for Vehicle Re-Identification Sensors vehicle re-identification deep learning machine learning attention mechanism vehicle details |
| title | Identity-Guided Spatial Attention for Vehicle Re-Identification |
| title_full | Identity-Guided Spatial Attention for Vehicle Re-Identification |
| title_fullStr | Identity-Guided Spatial Attention for Vehicle Re-Identification |
| title_full_unstemmed | Identity-Guided Spatial Attention for Vehicle Re-Identification |
| title_short | Identity-Guided Spatial Attention for Vehicle Re-Identification |
| title_sort | identity guided spatial attention for vehicle re identification |
| topic | vehicle re-identification deep learning machine learning attention mechanism vehicle details |
| url | https://www.mdpi.com/1424-8220/23/11/5152 |
| work_keys_str_mv | AT kailv identityguidedspatialattentionforvehiclereidentification AT shenghan identityguidedspatialattentionforvehiclereidentification AT youfanglin identityguidedspatialattentionforvehiclereidentification |