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 2204.13091cs.CV
收藏

单源领域泛化中视觉破坏的注意一致性

Attention Consistency on Visual Corruptions for Single-Source Domain Generalization

University of Tübingen(蒂宾根大学);MPI for Intelligent Systems(马克斯·普朗克智能系统研究所)

Ilke Cugu, Massimiliano Mancini, Yanbei Chen, Zeynep Akata

摘要:

将在单一分布上训练的视觉识别模型推广到看不见的输入分布(即域),需要使它们对训练集中的多余相关性具有鲁棒性。在这项工作中,我们通过改变训练图像来模拟新的领域,并在同一样本的不同视图中施加一致的视觉注意力来实现这一目标。我们发现,第一个目标可以简单而有效地满足通过视觉腐败。具体来说,我们使用ImageNet-C基准测试的19个损坏和基于傅立叶变换的三个额外变换来改变训练图像的内容。由于这些损坏保留了对象位置,我们提出了注意一致性损失,以确保同一训练样本的原始版本和损坏版本的类激活映射是对齐的。我们将我们的模型命名为视觉破坏的注意力一致性(ACVC)。我们表明,ACVC始终达到三个单源领域泛化基准,PACS,COCO和大规模DomainNet的最先进的。

英文摘要:

Generalizing visual recognition models trained on a single distribution to unseen input distributions (i.e. domains) requires making them robust to superfluous correlations in the training set. In this work, we achieve this goal by altering the training images to simulate new domains and imposing consistent visual attention across the different views of the same sample. We discover that the first objective can be simply and effectively met through visual corruptions. Specifically, we alter the content of the training images using the nineteen corruptions of the ImageNet-C benchmark and three additional transformations based on Fourier transform. Since these corruptions preserve object locations, we propose an attention consistency loss to ensure that class activation maps across original and corrupted versions of the same training sample are aligned. We name our model Attention Consistency on Visual Corruptions (ACVC). We show that ACVC consistently achieves the state of the art on three single-source domain generalization benchmarks, PACS, COCO, and the large-scale DomainNet.

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。

↑