LKFormer: Large Kernel Transformer for Infrared Image Super-Resolution

TL;DR

LKFormer introduces a large kernel transformer architecture with novel modules to enhance infrared image super-resolution, effectively capturing 2D structures and features with fewer parameters and superior results.

Contribution

The paper proposes LKFormer, a transformer model with large kernel residual attention and a gated-pixel feed-forward network, addressing infrared image super-resolution challenges.

Findings

Outperforms state-of-the-art methods in infrared super-resolution

Uses fewer parameters while achieving better performance

Effectively models non-local features with large kernels

Abstract

Given the broad application of infrared technology across diverse fields, there is an increasing emphasis on investigating super-resolution techniques for infrared images within the realm of deep learning. Despite the impressive results of current Transformer-based methods in image super-resolution tasks, their reliance on the self-attentive mechanism intrinsic to the Transformer architecture results in images being treated as one-dimensional sequences, thereby neglecting their inherent two-dimensional structure. Moreover, infrared images exhibit a uniform pixel distribution and a limited gradient range, posing challenges for the model to capture effective feature information. Consequently, we suggest a potent Transformer model, termed Large Kernel Transformer (LKFormer), to address this issue. Specifically, we have designed a Large Kernel Residual Attention (LKRA) module with linear complexity. This mainly employs depth-wise convolution with large kernels to execute non-local feature modeling, thereby substituting the standard self-attentive layer. Additionally, we have devised a novel feed-forward network structure called Gated-Pixel Feed-Forward Network (GPFN) to augment the LKFormer's capacity to manage the information flow within the network. Comprehensive experimental results reveal that our method surpasses the most advanced techniques available, using fewer parameters and yielding considerably superior performance.The source code will be available at https://github.com/sad192/large-kernel-Transformer.