Abstract:Given a degraded input image, image restoration aims to recover the missing high-quality image content. Numerous applications demand effective image restoration, e.g., computational photography, surveillance, autonomous vehicles, and remote sensing. Significant advances in image restoration have been made in recent years, dominated by convolutional neural networks (CNNs). The widely-used CNN-based methods typically operate either on full-resolution or on progressively low-resolution representations. In the former case, spatial details are preserved but the contextual information cannot be precisely encoded. In the latter case, generated outputs are semantically reliable but spatially less accurate. This paper presents a new architecture with a holistic goal of maintaining spatially-precise high-resolution representations through the entire network, and receiving complementary contextual information from the low-resolution representations. The core of our approach is a multi-scale residual block containing the following key elements: (a) parallel multi-resolution convolution streams for extracting multi-scale features, (b) information exchange across the multi-resolution streams, (c) non-local attention mechanism for capturing contextual information, and (d) attention based multi-scale feature aggregation. Our approach learns an enriched set of features that combines contextual information from multiple scales, while simultaneously preserving the high-resolution spatial details. Extensive experiments on six real image benchmark datasets demonstrate that our method, named as MIRNet-v2 , achieves state-of-the-art results for a variety of image processing tasks, including defocus deblurring, image denoising, super-resolution, and image enhancement. The source code and pre-trained models are available at https://github.com/swz30/MIRNetv2

Multi-scale residual network-based image restoration

Image Denoising Via Multi-Scale Gated Fusion Network

A Double Feature Fusion Network with Progressive Learning for Sharper Inpainting

Deep multi-scale gabor wavelet network for image restoration

Research on a multi‐scale degradation fusion network in all‐in‐one image restoration

Efficient Multi-scale Network with Learnable Discrete Wavelet Transform for Blind Motion Deblurring

Multi-scale Convolutional Neural Networks for Non-blind Image Deconvolution.

Multi-scale Residual Low-Pass Filter Network for Image Deblurring

Reference-Based Multi-Level Features Fusion Deblurring Network for Optical Remote Sensing Images

Learning Enriched Features for Fast Image Restoration and Enhancement

Multi-scale Grid Network for Image Deblurring with High-frequency Guidance

Multi-scale network for single image deblurring based on ensemble learning module

Multi-scale Frequency Enhancement Network for Blind Image Deblurring

A Multi-scale Generalized Shrinkage Threshold Network for Image Blind Deblurring in Remote Sensing

A Multiscale Image Denoising Algorithm Based On Dilated Residual Convolution Network

Multi-level Wavelet-CNN for Image Restoration

A Deep Convolutional Encoder–Decoder–Restorer Architecture for Image Deblurring

All-in-one Multi-degradation Image Restoration Network via Hierarchical Degradation Representation

Residual Dense Network for Image Restoration

A Multiscale Generalized Shrinkage Threshold Network for Image Blind Deblurring in Remote Sensing

Deep Likelihood Network for Image Restoration with Multiple Degradation Levels