Transformer-Progressive Mamba Network for Lightweight Image Super-Resolution

TL;DR

The paper introduces T-PMambaSR, a lightweight image super-resolution framework that combines window-based self-attention with Progressive Mamba and an adaptive high-frequency refinement module, achieving improved performance with lower computational cost.

Contribution

It proposes a novel integration of Progressive Mamba with window-based self-attention and an adaptive refinement module for efficient, high-quality image super-resolution.

Findings

T-PMambaSR outperforms recent Transformer- or Mamba-based methods in super-resolution tasks.

The method progressively enhances receptive fields and feature expressiveness.

It achieves better performance with lower computational cost.

Abstract

Recently, Mamba-based super-resolution (SR) methods have demonstrated the ability to capture global receptive fields with linear complexity, addressing the quadratic computational cost of Transformer-based SR approaches. However, existing Mamba-based methods lack fine-grained transitions across different modeling scales, which limits the efficiency of feature representation. In this paper, we propose T-PMambaSR, a lightweight SR framework that integrates window-based self-attention with Progressive Mamba. By enabling interactions among receptive fields of different scales, our method establishes a fine-grained modeling paradigm that progressively enhances feature representation without introducing additional computational cost. Furthermore, we introduce an Adaptive High-Frequency Refinement Module (AHFRM) to recover high-frequency details lost during Transformer and Mamba processing. Extensive experiments demonstrate that T-PMambaSR progressively enhances the model's receptive field and expressiveness, yielding better performance than recent Transformer- or Mamba-based methods while incurring lower computational cost.