MAMNet: Multi-path Adaptive Modulation Network for Image Super-Resolution

TL;DR

MAMNet introduces a multi-path adaptive modulation approach for image super-resolution, effectively capturing diverse image features with fewer parameters, leading to improved performance over existing methods.

Contribution

The paper proposes a novel multi-path adaptive modulation block (MAMB) that adaptively modulates features using three information paths, enhancing efficiency and effectiveness in super-resolution.

Findings

MAMNet outperforms many state-of-the-art SR methods.

MAMB is lightweight and parameter-efficient.

Experimental results validate the effectiveness of the proposed approach.

Abstract

In recent years, single image super-resolution (SR) methods based on deep convolutional neural networks (CNNs) have made significant progress. However, due to the non-adaptive nature of the convolution operation, they cannot adapt to various characteristics of images, which limits their representational capability and, consequently, results in unnecessarily large model sizes. To address this issue, we propose a novel multi-path adaptive modulation network (MAMNet). Specifically, we propose a multi-path adaptive modulation block (MAMB), which is a lightweight yet effective residual block that adaptively modulates residual feature responses by fully exploiting their information via three paths. The three paths model three types of information suitable for SR: 1) channel-specific information (CSI) using global variance pooling, 2) inter-channel dependencies (ICD) based on the CSI, 3) and channel-specific spatial dependencies (CSD) via depth-wise convolution. We demonstrate that the proposed MAMB is effective and parameter-efficient for image SR than other feature modulation methods. In addition, experimental results show that our MAMNet outperforms most of the state-of-the-art methods with a relatively small number of parameters.