Image super-resolution by fusing strengths of CNN and transformer with CRAFT
Image super-resolution by fusing strengths of CNN and transformer with CRAFT
Feature Modulation Transformer: Cross-Refinement of Global Representation via High-Frequency Prior for Image Super-Resolution
arXiv paper abstract https://arxiv.org/abs/2308.05022
arXiv PDF paper https://arxiv.org/pdf/2308.05022.pdf
Transformer-based methods have … remarkable potential in single image super-resolution (SISR) by effectively extracting long-range dependencies.
… most … research … has prioritized the design of transformer … to capture global information, while overlooking the importance of incorporating high-frequency priors
… found that transformer structures are more adept at capturing low-frequency information, but have limited capacity in constructing high-frequency representations … compared to their convolutional counterparts.
… proposed solution, the cross-refinement adaptive feature modulation transformer (CRAFT), integrates the strengths of both convolutional and transformer structures.
It comprises three key components: the high-frequency enhancement residual block (HFERB) for extracting high-frequency information, the shift rectangle window attention block (SRWAB) for capturing global information, and the hybrid fusion block (HFB) for refining the global representation.
… experiments on multiple datasets demonstrate that CRAFT outperforms state-of-the-art methods … while using fewer parameters …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
