Dynamic Uncertainty Learning with Noisy Correspondence for Text-Based Person Search
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917300929560576 |
|---|---|
| author | Xie, Zequn Ji, Haoming Li, Chengxuan Meng, Lingwei |
| author_facet | Xie, Zequn Ji, Haoming Li, Chengxuan Meng, Lingwei |
| contents | Text-to-image person search aims to identify an individual based on a text description. To reduce data collection costs, large-scale text-image datasets are created from co-occurrence pairs found online. However, this can introduce noise, particularly mismatched pairs, which degrade retrieval performance. Existing methods often focus on negative samples, which amplify this noise. To address these issues, we propose the Dynamic Uncertainty and Relational Alignment (DURA) framework, which includes the Key Feature Selector (KFS) and a new loss function, Dynamic Softmax Hinge Loss (DSH-Loss). KFS captures and models noise uncertainty, improving retrieval reliability. The bidirectional evidence from cross-modal similarity is modeled as a Dirichlet distribution, enhancing adaptability to noisy data. DSH adjusts the difficulty of negative samples to improve robustness in noisy environments. Our experiments on three datasets show that the method offers strong noise resistance and improves retrieval performance in both low- and high-noise scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_06566 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Dynamic Uncertainty Learning with Noisy Correspondence for Text-Based Person Search Xie, Zequn Ji, Haoming Li, Chengxuan Meng, Lingwei Computer Vision and Pattern Recognition Text-to-image person search aims to identify an individual based on a text description. To reduce data collection costs, large-scale text-image datasets are created from co-occurrence pairs found online. However, this can introduce noise, particularly mismatched pairs, which degrade retrieval performance. Existing methods often focus on negative samples, which amplify this noise. To address these issues, we propose the Dynamic Uncertainty and Relational Alignment (DURA) framework, which includes the Key Feature Selector (KFS) and a new loss function, Dynamic Softmax Hinge Loss (DSH-Loss). KFS captures and models noise uncertainty, improving retrieval reliability. The bidirectional evidence from cross-modal similarity is modeled as a Dirichlet distribution, enhancing adaptability to noisy data. DSH adjusts the difficulty of negative samples to improve robustness in noisy environments. Our experiments on three datasets show that the method offers strong noise resistance and improves retrieval performance in both low- and high-noise scenarios. |
| title | Dynamic Uncertainty Learning with Noisy Correspondence for Text-Based Person Search |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2505.06566 |