LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion
Abstract In the task of domain generalization person re-identification (ReID), pedestrian image features exhibit significant intra-class variability and inter-class similarity. Existing methods rely on a single feature extraction architecture and struggle to capture both global context and local spa...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2025-04-01
|
| Series: | Visual Computing for Industry, Biomedicine, and Art |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s42492-025-00190-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850181854352637952 |
|---|---|
| author | Xintong Hu Peishun Liu Xuefang Wang Peiyao Wu Ruichun Tang |
| author_facet | Xintong Hu Peishun Liu Xuefang Wang Peiyao Wu Ruichun Tang |
| author_sort | Xintong Hu |
| collection | DOAJ |
| description | Abstract In the task of domain generalization person re-identification (ReID), pedestrian image features exhibit significant intra-class variability and inter-class similarity. Existing methods rely on a single feature extraction architecture and struggle to capture both global context and local spatial information, resulting in weaker generalization to unseen domains. To address this issue, an innovative domain generalization person ReID method–LViT-Net, which combines local semantics and multi-feature cross fusion, is proposed. LViT-Net adopts a dual-branch encoder with a parallel hierarchical structure to extract both local and global discriminative features. In the local branch, the local multi-scale feature fusion module is designed to fuse local feature units at different scales to ensure that the fine-grained local features at various levels are accurately captured, thereby enhancing the robustness of the features. In the global branch, the dual feature cross fusion module fuses local features and global semantic information, focusing on critical semantic information and enabling the mutual refinement and matching of local and global features. This allows the model to achieve a dynamic balance between detailed and holistic information, forming robust feature representations of pedestrians. Extensive experiments demonstrate the effectiveness of LViT-Net. In both single-source and multi-source comparison experiments, the proposed method outperforms existing state-of-the-art methods. |
| format | Article |
| id | doaj-art-082c2f3766e24bffba08808d4efec489 |
| institution | OA Journals |
| issn | 2524-4442 |
| language | English |
| publishDate | 2025-04-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | Visual Computing for Industry, Biomedicine, and Art |
| spelling | doaj-art-082c2f3766e24bffba08808d4efec4892025-08-20T02:17:49ZengSpringerOpenVisual Computing for Industry, Biomedicine, and Art2524-44422025-04-018111510.1186/s42492-025-00190-1LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusionXintong Hu0Peishun Liu1Xuefang Wang2Peiyao Wu3Ruichun Tang4Faculty of Information Science and Engineering, Ocean University of ChinaFaculty of Information Science and Engineering, Ocean University of ChinaSchool of Mathematical Sciences, Ocean University of ChinaFaculty of Information Science and Engineering, Ocean University of ChinaFaculty of Information Science and Engineering, Ocean University of ChinaAbstract In the task of domain generalization person re-identification (ReID), pedestrian image features exhibit significant intra-class variability and inter-class similarity. Existing methods rely on a single feature extraction architecture and struggle to capture both global context and local spatial information, resulting in weaker generalization to unseen domains. To address this issue, an innovative domain generalization person ReID method–LViT-Net, which combines local semantics and multi-feature cross fusion, is proposed. LViT-Net adopts a dual-branch encoder with a parallel hierarchical structure to extract both local and global discriminative features. In the local branch, the local multi-scale feature fusion module is designed to fuse local feature units at different scales to ensure that the fine-grained local features at various levels are accurately captured, thereby enhancing the robustness of the features. In the global branch, the dual feature cross fusion module fuses local features and global semantic information, focusing on critical semantic information and enabling the mutual refinement and matching of local and global features. This allows the model to achieve a dynamic balance between detailed and holistic information, forming robust feature representations of pedestrians. Extensive experiments demonstrate the effectiveness of LViT-Net. In both single-source and multi-source comparison experiments, the proposed method outperforms existing state-of-the-art methods.https://doi.org/10.1186/s42492-025-00190-1Domain generalizationPerson re-identificationFeature fusionSemantic representationDual-branch network architecture |
| spellingShingle | Xintong Hu Peishun Liu Xuefang Wang Peiyao Wu Ruichun Tang LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion Visual Computing for Industry, Biomedicine, and Art Domain generalization Person re-identification Feature fusion Semantic representation Dual-branch network architecture |
| title | LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion |
| title_full | LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion |
| title_fullStr | LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion |
| title_full_unstemmed | LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion |
| title_short | LViT-Net: a domain generalization person re-identification model combining local semantics and multi-feature cross fusion |
| title_sort | lvit net a domain generalization person re identification model combining local semantics and multi feature cross fusion |
| topic | Domain generalization Person re-identification Feature fusion Semantic representation Dual-branch network architecture |
| url | https://doi.org/10.1186/s42492-025-00190-1 |
| work_keys_str_mv | AT xintonghu lvitnetadomaingeneralizationpersonreidentificationmodelcombininglocalsemanticsandmultifeaturecrossfusion AT peishunliu lvitnetadomaingeneralizationpersonreidentificationmodelcombininglocalsemanticsandmultifeaturecrossfusion AT xuefangwang lvitnetadomaingeneralizationpersonreidentificationmodelcombininglocalsemanticsandmultifeaturecrossfusion AT peiyaowu lvitnetadomaingeneralizationpersonreidentificationmodelcombininglocalsemanticsandmultifeaturecrossfusion AT ruichuntang lvitnetadomaingeneralizationpersonreidentificationmodelcombininglocalsemanticsandmultifeaturecrossfusion |