Restore image by preventing local and global connections from spreading noise with LIR
Restore image by preventing local and global connections from spreading noise with LIR
LIR: Efficient Degradation Removal for Lightweight Image Restoration
arXiv paper abstract https://browse.arxiv.org/abs/2402.01368
arXiv PDF paper https://browse.arxiv.org/pdf/2402.01368.pdf
… Image Restoration based on CNN and transformer … often … leading to redundant parameters and unnecessary computations and hindering the efficiency of the … restoration.
… propose a Lightweight Image Restoration network called LIR to efficiently remove degradation (blur, rain, noise, haze, etc.).
A key component in LIR is the Efficient Adaptive Attention (EAA) Block, which is mainly composed of Adaptive Filters and Attention Blocks.
It is capable of adaptively sharpening contours, removing degradation, and capturing global information in various image restoration scenes in an efficient … manner.
… through a simple structural design, LIR addresses the degradations existing in the local and global residual connections that are ignored by modern networks.
… LIR achieves comparable performance to state-of-the-art networks on most benchmarks with fewer parameters and computations …
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
