MixNet: Efficient Global Modeling for Ultra-High-Definition Image Restoration

TL;DR

MixNet introduces a novel global modeling approach with efficient feature modulation layers, enabling high-quality ultra-high-definition image restoration with low computational overhead across various tasks.

Contribution

The paper proposes MixNet, a new image restoration model that effectively captures long-range dependencies using GFML, LFML, and FFL layers, optimized for UHD images.

Findings

Outperforms state-of-the-art methods on UHD restoration tasks

Achieves low inference time and computational complexity

Demonstrates effectiveness across multiple restoration tasks

Abstract

Recent advancements in image restoration methods employing global modeling have shown promising results. However, these approaches often incur substantial memory requirements, particularly when processing ultra-high-definition (UHD) images. In this paper, we propose a novel image restoration method called MixNet, which introduces an alternative approach to global modeling approaches and is more effective for UHD image restoration. To capture the longrange dependency of features without introducing excessive computational complexity, we present the Global Feature Modulation Layer (GFML). GFML associates features from different views by permuting the feature maps, enabling efficient modeling of long-range dependency. In addition, we also design the Local Feature Modulation Layer (LFML) and Feed-forward Layer (FFL) to capture local features and transform features into a compact representation. This way, our MixNetachieves effective restoration with low inference time overhead and computational complexity. We conduct extensive experiments on four UHD image restoration tasks, including low-light image enhancement, underwater image enhancement, image deblurring and image demoireing, and the comprehensive results demonstrate that our proposed method surpasses the performance of current state-of-the-art methods. The code will be available at \url{https://github.com/5chen/MixNet}.