Re-Attention Transformer for Weakly Supervised Object Localization
Hui Su (Zhejiang Lab), Yue Ye (Zhejiang Lab), Zhiwei Chen (Zhe jiang Lab), Mingli Song (Zhejiang University), Lechao Cheng (Zhejiang Lab)*
The 33rd British Machine Vision Conference
Abstract
Weakly supervised object localization is a challenging task which aims to localize objects with coarse annotations such as image categories. Existing deep network approaches are mainly based on class activation map, which focuses on highlighting discriminative local region while ignoring the full object. In addition, the emerging transformer-based techniques constantly put a lot of emphasis on the backdrop that impedes the ability to identify complete objects. To address these issues, we present a Re-Attention mechanism termed token refinement transformer (TRT) that captures the object-level semantics to guide the localization well. Specifically, TRT introduces a novel module named token priority scoring module (TPSM) to suppress the effects of background noise while focusing on the target object. Then, we incorporate the class activation map as the semantically aware input to restrain the attention map to the target object. Extensive experiments on two benchmarks showcase the superiority of our proposed method against existing methods with image category annotations. Source code is available in \url{https://github.com/su-hui-zz/ReAttentionTransformer}.Video
Citation
@inproceedings{Su_2022_BMVC,
author = {Hui Su and Yue Ye and Zhiwei Chen and Mingli Song and Lechao Cheng},
title = {Re-Attention Transformer for Weakly Supervised Object Localization},
booktitle = {33rd British Machine Vision Conference 2022, {BMVC} 2022, London, UK, November 21-24, 2022},
publisher = {{BMVA} Press},
year = {2022},
url = {https://bmvc2022.mpi-inf.mpg.de/0070.pdf}
}
Copyright © 2022
The British Machine Vision Conference is organised by