Super-resolution on a single image by using an adaptive MaxViT transformer with MaxSR
Super-resolution on a single image by using an adaptive MaxViT transformer with MaxSR
MaxSR: Image Super-Resolution Using Improved MaxViT
arXiv paper abstract https://arxiv.org/abs/2307.07240
arXiv PDF paper https://arxiv.org/pdf/2307.07240.pdf
... few attempts have been made to use powerful transformer models for single image super-resolution.
Because transformer ... representation capacity and ... self-attention ... leverage self-similarity prior in ... low-resolution image to improve ... super-resolution ... present ... super-resolution model based on ... hybrid vision transformer of MaxViT, named as MaxSR.
MaxSR consists of four parts, a shallow feature extraction block, multiple cascaded adaptive MaxViT blocks to extract deep hierarchical features and model global self-similarity from low-level features efficiently, a hierarchical feature fusion block, and finally a reconstruction block.
The key component ... adaptive MaxViT block, is based on MaxViT block which mixes MBConv with squeeze-and-excitation, block attention and grid attention.
... to achieve better global modelling of self-similarity in ... image ... improve block attention and grid attention in MaxViT block to adaptive block attention and adaptive grid attention which do self-attention inside each window across all grids and each grid across all windows
... MaxSR and MaxSR-light establish new state-of-the-art performance efficiently.
Please like and share this post if you enjoyed it using the buttons at the bottom!
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
Comments