基于小波域ADMM深度网络的图像复原算法

近年来，基于深度学习的方法在图像复原领域展现出了优秀的性能。然而现有大多数深度网络均是通过经验进行网络结构设计，较少在网络设计中考虑结合一些传统方法以提升网络可解释性。针对这一不足，本文对结合图像退化模型的深度学习方法展开研究，提出了一种基于小波域ADMM深度网络的图像复原算法。具体而言，本文首先提出了一种基于小波域ADMM的图像复原方法，该方法在小波域下使用ADMM算法将复原问题分解为一系列子问题。接着，分别对子问题求解，并根据其解的形式帮助进行网络的设计，构建了一个可解释的深度卷积神经网络用于图像复原。实验结果表明，本文提出算法取得了较好的复原结果，不论在视觉效果还是客观评价指标上都优于对比方法。

Deep ADMM Network in Wavelet Domain for Image Restoration

In recent years, deep learning based methods have shown excellent performance in the field of image restoration. However, most of the existing deep networks are designed by experience, and traditional methods are rarely considered in network design to improve the network interpretability. To solve this problem, this paper studies the deep learning method combined with the image degradation model, and proposes an image restoration algorithm based on the deep ADMM network in the wavelet domain. To be specific, the restoration problem is decomposed into a series of subproblems by using ADMM in the wavelet domain firstly. Then, the subproblems are solved separately, and network is constructed according to the form of the solutions, leading to a interpretable deep neural network for image restoration. Experimental results show that the proposed algorithm achieves better results on both visual appearance and objective indices compared with state-of-the-art methods.