arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2112.00580cs.CV
收藏

弱监督目标定位中的背景激活抑制

Background Activation Suppression for Weakly Supervised Object Localization

University of Science and Technology of China(中国科学技术大学);Institute of Artificial Intelligence, Hefei Comprehensive National Science Center(合肥综合性国家科学中心人工智能研究院)

Pingyu Wu, Wei Zhai, Yang Cao

摘要:

弱监督对象定位(WSOL)旨在仅使用图像级标签来定位对象。最近出现了一种新的范例,通过生成前景预测图(FPM)来实现定位任务。现有的基于FPM的方法使用交叉熵(CE)来评估前景预测图并指导生成器的学习。我们主张使用激活值来实现更有效的学习。它是基于实验观察,对于一个训练有素的网络,CE收敛到零时,前景掩模只覆盖部分的对象区域。而激活值增加,直到遮罩扩展到对象边界,这表明使用激活值可以学习更多的对象区域。在本文中,我们提出了一种背景激活抑制(BAS)方法。具体而言,激活映射约束模块(AMC)的设计,以促进通过抑制背景激活值的发生器的学习。同时,通过前景区域引导和面积约束,BAS可以学习对象的整个区域。在推理阶段,我们将不同类别的预测图放在一起考虑,以获得最终的定位结果。大量的实验表明,BAS实现了显着的和一致的改进,在CUB-200-2011和ILSVRC数据集上的基线方法。代码和型号可在https://github.com/wpy1999/BAS上获得。

英文摘要:

Weakly supervised object localization (WSOL) aims to localize objects using only image-level labels. Recently a new paradigm has emerged by generating a foreground prediction map (FPM) to achieve localization task. Existing FPM-based methods use cross-entropy (CE) to evaluate the foreground prediction map and to guide the learning of generator. We argue for using activation value to achieve more efficient learning. It is based on the experimental observation that, for a trained network, CE converges to zero when the foreground mask covers only part of the object region. While activation value increases until the mask expands to the object boundary, which indicates that more object areas can be learned by using activation value. In this paper, we propose a Background Activation Suppression (BAS) method. Specifically, an Activation Map Constraint module (AMC) is designed to facilitate the learning of generator by suppressing the background activation value. Meanwhile, by using the foreground region guidance and the area constraint, BAS can learn the whole region of the object. In the inference phase, we consider the prediction maps of different categories together to obtain the final localization results. Extensive experiments show that BAS achieves significant and consistent improvement over the baseline methods on the CUB-200-2011 and ILSVRC datasets. Code and models are available at https://github.com/wpy1999/BAS.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑