Pedestrian Attribute Recognition as Label-balanced Multi-label Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929337665585152 |
|---|---|
| author | Zhou, Yibo Hu, Hai-Miao Xiang, Yirong Zhang, Xiaokang Wu, Haotian |
| author_facet | Zhou, Yibo Hu, Hai-Miao Xiang, Yirong Zhang, Xiaokang Wu, Haotian |
| contents | Rooting in the scarcity of most attributes, realistic pedestrian attribute datasets exhibit unduly skewed data distribution, from which two types of model failures are delivered: (1) label imbalance: model predictions lean greatly towards the side of majority labels; (2) semantics imbalance: model is easily overfitted on the under-represented attributes due to their insufficient semantic diversity. To render perfect label balancing, we propose a novel framework that successfully decouples label-balanced data re-sampling from the curse of attributes co-occurrence, i.e., we equalize the sampling prior of an attribute while not biasing that of the co-occurred others. To diversify the attributes semantics and mitigate the feature noise, we propose a Bayesian feature augmentation method to introduce true in-distribution novelty. Handling both imbalances jointly, our work achieves best accuracy on various popular benchmarks, and importantly, with minimal computational budget. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_04858 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Pedestrian Attribute Recognition as Label-balanced Multi-label Learning Zhou, Yibo Hu, Hai-Miao Xiang, Yirong Zhang, Xiaokang Wu, Haotian Computer Vision and Pattern Recognition Rooting in the scarcity of most attributes, realistic pedestrian attribute datasets exhibit unduly skewed data distribution, from which two types of model failures are delivered: (1) label imbalance: model predictions lean greatly towards the side of majority labels; (2) semantics imbalance: model is easily overfitted on the under-represented attributes due to their insufficient semantic diversity. To render perfect label balancing, we propose a novel framework that successfully decouples label-balanced data re-sampling from the curse of attributes co-occurrence, i.e., we equalize the sampling prior of an attribute while not biasing that of the co-occurred others. To diversify the attributes semantics and mitigate the feature noise, we propose a Bayesian feature augmentation method to introduce true in-distribution novelty. Handling both imbalances jointly, our work achieves best accuracy on various popular benchmarks, and importantly, with minimal computational budget. |
| title | Pedestrian Attribute Recognition as Label-balanced Multi-label Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2405.04858 |