We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Image and Video Processing

Title: Activating More Pixels in Image Super-Resolution Transformer

Abstract: Transformer-based methods have shown impressive performance in low-level vision tasks, such as image super-resolution. However, we find that these networks can only utilize a limited spatial range of input information through attribution analysis. This implies that the potential of Transformer is still not fully exploited in existing networks. In order to activate more input pixels for reconstruction, we propose a novel Hybrid Attention Transformer (HAT). It combines channel attention and self-attention schemes, thus making use of their complementary advantages. Moreover, to better aggregate the cross-window information, we introduce an overlapping cross-attention module to enhance the interaction between neighboring window features. In the training stage, we additionally propose a same-task pre-training strategy to bring further improvement. Extensive experiments show the effectiveness of the proposed modules, and the overall method significantly outperforms the state-of-the-art methods by more than 1dB. Codes and models will be available at this https URL
Comments: this https URL
Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2205.04437 [eess.IV]
  (or arXiv:2205.04437v2 [eess.IV] for this version)

Submission history

From: Xiangyu Chen [view email]
[v1] Mon, 9 May 2022 17:36:58 GMT (10813kb,D)
[v2] Mon, 16 May 2022 09:13:36 GMT (10813kb,D)

Link back to: arXiv, form interface, contact.