Second-order Attention Network for Single Image Super-resolution
We propose a second-order attention network (SAN) for more powerful feature expression and feature correlation learning.
super-resolution attention computer-vision research tutorial code paper arxiv:1909.11937

We propose a second-order attention network (SAN) for more powerful feature expression and feature correlation learning. Specifically, a novel train- able second-order channel attention (SOCA) module is developed to adaptively rescale the channel-wise features by using second-order feature statistics for more discriminative representations. Furthermore, we present a non-locally enhanced residual group (NLRG) structure, which not only incorporates non-local operations to capture long-distance spatial contextual information, but also contains repeated local-source residual attention groups (LSRAG) to learn increasingly abstract feature representations. Experimental results demonstrate the superiority of our SAN network over state-of-the-art SISR methods in terms of both quantitative metrics and visual quality.

Don't forget to tag @daitao in your comment, otherwise they may not be notified.

Authors
Share this project
Similar projects
Deep Tutorials for PyTorch
This is a series of in-depth tutorials I'm writing for implementing cool deep learning models on your own with the amazing PyTorch library.
Super-resolution Variational Auto-Encoders
VAE with RealNVP prior and Super-Resolution VAE in PyTorch.
Latest advancements in video streaming with AI
AI developments in video streaming using Super-resolution, Per-title encoding, P2P
Image Super-Resolution
In this project we learn how to train a super-resolution model ESPCN on DIV2K dataset to upscale images using AI by 3x
Top collections