Recursive Generalization Transformer for Image Super-Resolution

TL;DR

The paper introduces the Recursive Generalization Transformer (RGT) for image super-resolution, effectively capturing global context with recursive and cross-attention mechanisms, outperforming existing methods.

Contribution

It proposes a novel recursive-generalization self-attention mechanism combined with local attention and a hybrid adaptive integration for improved image super-resolution.

Findings

RGT achieves superior quantitative results on benchmark datasets.

The model effectively captures global spatial information.

Extensive experiments validate the effectiveness of the proposed approach.

Abstract

Transformer architectures have exhibited remarkable performance in image super-resolution (SR). Since the quadratic computational complexity of the self-attention (SA) in Transformer, existing methods tend to adopt SA in a local region to reduce overheads. However, the local design restricts the global context exploitation, which is crucial for accurate image reconstruction. In this work, we propose the Recursive Generalization Transformer (RGT) for image SR, which can capture global spatial information and is suitable for high-resolution images. Specifically, we propose the recursive-generalization self-attention (RG-SA). It recursively aggregates input features into representative feature maps, and then utilizes cross-attention to extract global information. Meanwhile, the channel dimensions of attention matrices (query, key, and value) are further scaled to mitigate the redundancy in the channel domain. Furthermore, we combine the RG-SA with local self-attention to enhance the exploitation of the global context, and propose the hybrid adaptive integration (HAI) for module integration. The HAI allows the direct and effective fusion between features at different levels (local or global). Extensive experiments demonstrate that our RGT outperforms recent state-of-the-art methods quantitatively and qualitatively. Code and pre-trained models are available at https://github.com/zhengchen1999/RGT.