《计算机应用研究》|Application Research of Computers

基于二阶表征的条件对抗域适应网络

Conditional adversarial domain adaptation networks based on second-order representation

免费全文下载 (已被下载 次)  
获取PDF全文
作者 徐春荞,张冰冰,李培华
机构 大连理工大学 信息与通信工程学院,辽宁 大连 116024
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2021)10-027-3040-04
DOI 10.19734/j.issn.1001-3695.2021.01.0046
摘要 域对抗学习是一种主流的域适应方法,它通过分类器和域判别器来学习具有可区分性的域不变特征;然而,现有的域对抗方法大多利用一阶特征来学习域不变特征,忽略了具有更强表达能力的二阶特征。提出了一种条件对抗域适应网络,通过联合建模图像的二阶表征以及特征和分类器预测之间的互协方差以便更有效地学习具有区分性的域不变特征;此外,引入了熵条件来平衡分类器预测的不确定性,以保证特征的可迁移性。提出的方法在两个常用的域适应数据库Office-31和ImageCLEF-DA上进行了验证,实验结果表明该方法优于同类方法并获得了领先的性能。
关键词 域适应; 二阶表征; 互协方差; 对抗网络
基金项目 国家自然科学基金资助项目(61971086)
本文URL http://www.arocmag.com/article/01-2021-10-027.html
英文标题 Conditional adversarial domain adaptation networks based on second-order representation
作者英文名 Xu Chunqiao, Zhang Bingbing, Li Peihua
机构英文名 School of Information & Communication Engineering,Dalian University of Technology,Dalian Liaoning 116024,China
英文摘要 Domain adversarial learning is a mainstream approach of domain adaptation, which learns discriminative domain-invariant feature representation through classifier and domain discriminator. However, existing adversarial domain adaptation methods often use first-order features to learn domain-invariant feature representation, ignoring the second-order features with more powerful representative ability. This paper proposed conditional adversarial domain adaptation networks based on second-order representation, which modeled the second-order moments of features and the cross-covariance between features and classifier predictions for more effectively learning discriminative domain-invariant features. Moreover, it introduced entropy conditioning to guarantee the transferability. The proposed method was evaluated on two commonly used datasets Office-31 and ImageCLEF-DA. Experiments show that the proposed method outperforms its counterpart.
英文关键词 domain adaptation; second-order representation; cross-covariance; adversarial network
参考文献 查看稿件参考文献
 
收稿日期 2021/1/14
修回日期 2021/3/6
页码 3040-3043,3048
中图分类号 TP391
文献标志码 A