Extra Proximal-Gradient Inspired Non-local Network

TL;DR

This paper introduces a novel deep neural network inspired by an accelerated extra proximal gradient algorithm, incorporating non-local operations for improved image reconstruction by exploiting image self-similarity.

Contribution

It presents a new network architecture that combines optimization-inspired design with non-local operations and learns parameters through end-to-end training.

Findings

Outperforms several state-of-the-art deep networks

Uses non-local self-similarity for better reconstruction

Achieves superior results with similar number of parameters

Abstract

Variational method and deep learning method are two mainstream powerful approaches to solve inverse problems in computer vision. To take advantages of advanced optimization algorithms and powerful representation ability of deep neural networks, we propose a novel deep network for image reconstruction. The architecture of this network is inspired by our proposed accelerated extra proximal gradient algorithm. It is able to incorporate non-local operation to exploit the non-local self-similarity of the images and to learn the nonlinear transform, under which the solution is sparse. All the parameters in our network are learned from minimizing a loss function. Our experimental results show that our network outperforms several state-of-the-art deep networks with almost the same number of learnable parameter.