Attentive Part-Based Alignment Network for Vehicle Re-Identification
Vehicle Re-identification (Re-ID) has become a research hotspot along with the rapid development of video surveillance. Attention mechanisms are utilized in vehicle Re-ID networks but often miss the attention alignment across views. In this paper, we propose a novel Attentive Part-based Alignment Ne...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/11/10/1617 |
| _version_ | 1797500215109877760 |
|---|---|
| author | Yichu Liu Haifeng Hu Dihu Chen |
| author_facet | Yichu Liu Haifeng Hu Dihu Chen |
| author_sort | Yichu Liu |
| collection | DOAJ |
| description | Vehicle Re-identification (Re-ID) has become a research hotspot along with the rapid development of video surveillance. Attention mechanisms are utilized in vehicle Re-ID networks but often miss the attention alignment across views. In this paper, we propose a novel Attentive Part-based Alignment Network (APANet) to learn robust, diverse, and discriminative features for vehicle Re-ID. To be specific, in order to enhance the discrimination of part features, two part-level alignment mechanisms are proposed in APANet, consisting of Part-level Orthogonality Loss (POL) and Part-level Attention Alignment Loss (PAAL). Furthermore, POL aims to maximize the diversity of part features via an orthogonal penalty among parts whilst PAAL learns view-invariant features by means of realizing attention alignment in a part-level fashion. Moreover, we propose a Multi-receptive-field Attention (MA) module to adopt an efficient and cost-effective pyramid structure. The pyramid structure is capable of employing more fine-grained and heterogeneous-scale spatial attention information through multi-receptive-field streams. In addition, the improved TriHard loss and Inter-group Feature Centroid Loss (IFCL) function are utilized to optimize both the inter-group and intra-group distance. Extensive experiments demonstrate the superiority of our model over multiple existing state-of-the-art approaches on two popular vehicle Re-ID benchmarks. |
| first_indexed | 2024-03-10T03:59:39Z |
| format | Article |
| id | doaj.art-19f70af20e2741109e2fa6620dedf491 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T03:59:39Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-19f70af20e2741109e2fa6620dedf4912023-11-23T10:47:50ZengMDPI AGElectronics2079-92922022-05-011110161710.3390/electronics11101617Attentive Part-Based Alignment Network for Vehicle Re-IdentificationYichu Liu0Haifeng Hu1Dihu Chen2The School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou 510006, ChinaThe School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou 510006, ChinaThe School of Electronics and Information Technology, Sun Yat-sen University, Guangzhou 510006, ChinaVehicle Re-identification (Re-ID) has become a research hotspot along with the rapid development of video surveillance. Attention mechanisms are utilized in vehicle Re-ID networks but often miss the attention alignment across views. In this paper, we propose a novel Attentive Part-based Alignment Network (APANet) to learn robust, diverse, and discriminative features for vehicle Re-ID. To be specific, in order to enhance the discrimination of part features, two part-level alignment mechanisms are proposed in APANet, consisting of Part-level Orthogonality Loss (POL) and Part-level Attention Alignment Loss (PAAL). Furthermore, POL aims to maximize the diversity of part features via an orthogonal penalty among parts whilst PAAL learns view-invariant features by means of realizing attention alignment in a part-level fashion. Moreover, we propose a Multi-receptive-field Attention (MA) module to adopt an efficient and cost-effective pyramid structure. The pyramid structure is capable of employing more fine-grained and heterogeneous-scale spatial attention information through multi-receptive-field streams. In addition, the improved TriHard loss and Inter-group Feature Centroid Loss (IFCL) function are utilized to optimize both the inter-group and intra-group distance. Extensive experiments demonstrate the superiority of our model over multiple existing state-of-the-art approaches on two popular vehicle Re-ID benchmarks.https://www.mdpi.com/2079-9292/11/10/1617Vehicle Re-identificationattention mechanismpart orthogonalitypyramid attentionfeature extractionvideo surveillance |
| spellingShingle | Yichu Liu Haifeng Hu Dihu Chen Attentive Part-Based Alignment Network for Vehicle Re-Identification Electronics Vehicle Re-identification attention mechanism part orthogonality pyramid attention feature extraction video surveillance |
| title | Attentive Part-Based Alignment Network for Vehicle Re-Identification |
| title_full | Attentive Part-Based Alignment Network for Vehicle Re-Identification |
| title_fullStr | Attentive Part-Based Alignment Network for Vehicle Re-Identification |
| title_full_unstemmed | Attentive Part-Based Alignment Network for Vehicle Re-Identification |
| title_short | Attentive Part-Based Alignment Network for Vehicle Re-Identification |
| title_sort | attentive part based alignment network for vehicle re identification |
| topic | Vehicle Re-identification attention mechanism part orthogonality pyramid attention feature extraction video surveillance |
| url | https://www.mdpi.com/2079-9292/11/10/1617 |
| work_keys_str_mv | AT yichuliu attentivepartbasedalignmentnetworkforvehiclereidentification AT haifenghu attentivepartbasedalignmentnetworkforvehiclereidentification AT dihuchen attentivepartbasedalignmentnetworkforvehiclereidentification |