Get our free extension to see links to code for papers anywhere online!

Chrome logo  Add to Chrome

Firefox logo Add to Firefox

Deep Frequency Filtering for Domain Generalization

Add code

Mar 23, 2022
Shiqi Lin, Zhizheng Zhang, Zhipeng Huang, Yan Lu, Cuiling Lan, Peng Chu, Quanzeng You, Jiang Wang, Zicheng Liu, Amey Parulkar, Viraj Navkal, Zhibo Chen

Share this with someone who'll enjoy it:

Improving the generalization capability of Deep Neural Networks (DNNs) is critical for their practical uses, which has been a longstanding challenge. Some theoretical studies have revealed that DNNs have preferences to different frequency components in the learning process and indicated that this may affect the robustness of learned features. In this paper, we propose Deep Frequency Filtering (DFF) for learning domain-generalizable features, which is the first endeavour to explicitly modulate frequency components of different transfer difficulties across domains during training. To achieve this, we perform Fast Fourier Transform (FFT) on feature maps at different layers, then adopt a light-weight module to learn the attention masks from frequency representations after FFT to enhance transferable frequency components while suppressing the components not conductive to generalization. Further, we empirically compare different types of attention for implementing our conceptualized DFF. Extensive experiments demonstrate the effectiveness of the proposed DFF and show that applying DFF on a plain baseline outperforms the state-of-the-art methods on different domain generalization tasks, including close-set classification and open-set retrieval.

   Access Paper Source

Share this with someone who'll enjoy it: