Knowledge Commons of Institute of Automation,CAS
It takes two: Dual Branch Augmentation Module for domain generalization | |
Li, Jingwei1,2![]() ![]() ![]() | |
发表期刊 | NEURAL NETWORKS
![]() |
ISSN | 0893-6080 |
2024-04-01 | |
卷号 | 172页码:12 |
摘要 | Although great success has been achieved in various computer vision tasks, deep neural networks (DNNs) suffer dramatic performance degradation when evaluated on out-of-distribution data. Domain generalization (DG) is proposed to handle this problem by learning domain-agnostic information from multiple source domains to generalize well on unseen target domains. Several methods resort to Fourier transform due to its simplicity and efficiency. They argue that amplitude spectra imply domain-specific information, which should be suppressed, while phase counterparts imply domain-agnostic information, which should be preserved. However, these methods only suppress the domain-specific information in source domains and neglect the relationship with target domains, leading to the persistence of the domain gap. Besides, these methods preserve domain-agnostic information by keeping phase components unchanged, causing them to be underutilized. In this paper, we propose Dual Branch Augmentation Module (DBAM) by leveraging Fourier transform and taking advantage of both amplitude and phase spectra. For the amplitude branch, we propose Inner-domain Amplitude Distribution Rectification (IADR) and Cross-domain Amplitude Dirichlet Mixup (CADM) to stabilize the training process and explore more feature space. In addition, we propose Test-time Amplitude Prototype Calibration (TAPC) to construct the connection between source and target domains during evaluation to further mitigate the domain gap. For the phase branch, we propose Random Symmetric Phase Perturbation (RSPP) to enhance the robustness for recognizing domain-agnostic information. With the combined contributions of the two branches, DBAM significantly surpasses other state-of-the-art (SOTA) methods. Extensive experiments on four benchmarks and further analysis demonstrate the effectiveness of DBAM. |
关键词 | Domain generalization Fourier transform Uncertainty calibration Test-time adaptation |
DOI | 10.1016/j.neunet.2023.106094 |
收录类别 | SCI |
语种 | 英语 |
WOS研究方向 | Computer Science ; Neurosciences & Neurology |
WOS类目 | Computer Science, Artificial Intelligence ; Neurosciences |
WOS记录号 | WOS:001157966200001 |
出版者 | PERGAMON-ELSEVIER SCIENCE LTD |
七大方向——子方向分类 | 机器学习 |
国重实验室规划方向分类 | 虚实融合与迁移学习 |
是否有论文关联数据集需要存交 | 否 |
引用统计 | |
文献类型 | 期刊论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/55644 |
专题 | 中科院工业视觉智能装备工程实验室 |
通讯作者 | Liu, Chengbao |
作者单位 | 1.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China |
第一作者单位 | 中国科学院自动化研究所 |
通讯作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Li, Jingwei,Li, Yuan,Tan, Jie,et al. It takes two: Dual Branch Augmentation Module for domain generalization[J]. NEURAL NETWORKS,2024,172:12. |
APA | Li, Jingwei,Li, Yuan,Tan, Jie,&Liu, Chengbao.(2024).It takes two: Dual Branch Augmentation Module for domain generalization.NEURAL NETWORKS,172,12. |
MLA | Li, Jingwei,et al."It takes two: Dual Branch Augmentation Module for domain generalization".NEURAL NETWORKS 172(2024):12. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
paper3.pdf(1630KB) | 期刊论文 | 作者接受稿 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论