Channel Attention and Multi-level Features Fusion for Single Image Super-Resolution

TL;DR

This paper introduces a novel CNN-based super-resolution method that employs channel attention and multi-level feature fusion, enhancing detail recovery and computational efficiency.

Contribution

It proposes a recursive unit with channel attention and multi-level feature fusion, improving super-resolution performance while maintaining faster processing speeds.

Findings

Achieves competitive super-resolution results compared to state-of-the-art methods.

Uses residual branch upsampling with learnable transposed convolution for better detail learning.

Maintains faster speed despite enhanced performance.

Abstract

Convolutional neural networks (CNNs) have demonstrated superior performance in super-resolution (SR). However, most CNN-based SR methods neglect the different importance among feature channels or fail to take full advantage of the hierarchical features. To address these issues, this paper presents a novel recursive unit. Firstly, at the beginning of each unit, we adopt a compact channel attention mechanism to adaptively recalibrate the channel importance of input features. Then, the multi-level features, rather than only deep-level features, are extracted and fused. Additionally, we find that it will force our model to learn more details by using the learnable upsampling method (i.e., transposed convolution) only on residual branch (instead of using it both on residual branch and identity branch) while using the bicubic interpolation on the other branch. Analytic experiments show that our method achieves competitive results compared with the state-of-the-art methods and maintains faster speed as well.