Transcending the Limit of Local Window: Advanced Super-Resolution Transformer with Adaptive Token Dictionary

TL;DR

This paper proposes an advanced super-resolution transformer with an adaptive token dictionary and category-based self-attention, significantly improving the receptive field and performance in single image super-resolution tasks.

Contribution

Introduces an Adaptive Token Dictionary and category-based self-attention mechanism to enhance Transformer-based super-resolution models.

Findings

Achieves state-of-the-art results on super-resolution benchmarks.

Effectively learns and adapts prior information from training data.

Enhances global information integration through adaptive refinement.

Abstract

Single Image Super-Resolution is a classic computer vision problem that involves estimating high-resolution (HR) images from low-resolution (LR) ones. Although deep neural networks (DNNs), especially Transformers for super-resolution, have seen significant advancements in recent years, challenges still remain, particularly in limited receptive field caused by window-based self-attention. To address these issues, we introduce a group of auxiliary Adaptive Token Dictionary to SR Transformer and establish an ATD-SR method. The introduced token dictionary could learn prior information from training data and adapt the learned prior to specific testing image through an adaptive refinement step. The refinement strategy could not only provide global information to all input tokens but also group image tokens into categories. Based on category partitions, we further propose a category-based self-attention mechanism designed to leverage distant but similar tokens for enhancing input features. The experimental results show that our method achieves the best performance on various single image super-resolution benchmarks.