Abstract:Deep neural networks (DNNs) have achieved significant success in image restoration tasks by directly learning a powerful non-linear mapping from corrupted images to their latent clean ones. However, there still exist two major limitations for these deep learning (DL)-based methods. Firstly, the noises contained in real corrupted images are very complex, usually neglected and largely under-estimated in most current methods. Secondly, existing DL methods are mostly trained on one pre-assumed degradation process for all of the training image pairs, such as the widely used bicubic downsampling assumption in the image super-resolution task, inevitably leading to poor generalization performance when the true degradation does not match with such assumed one. To address these issues, we propose a unified generative model for the image restoration, which elaborately configures the degradation process from the latent clean image to the observed corrupted one. Specifically, different from most of current methods, the pixel-wisely non-i.i.d. Gaussian distribution, being with more flexibility, is adopted in our method to fit the complex real noises. Furthermore, the method is built on the general image degradation process, making it capable of adapting diverse degradations under one single model. Besides, we design a variational inference algorithm to learn all parameters involved in the proposed model with explicit form of objective loss. Specifically, beyond traditional variational methodology, two DNNs are employed to parameterize the posteriori distributions, one to infer the distribution of the latent clean image, and another to infer the distribution of the image noise. Extensive experiments demonstrate the superiority of the proposed method on three classical image restoration tasks, including image denoising, image super-resolution and JPEG image deblocking.

Learning local regularization for variational image restoration

Variational Deep Image Restoration

Deep Graph Laplacian Regularization for Robust Denoising of Real Images

Structured and Localized Image Restoration

Deep unfolding of a proximal interior point method for image restoration

Variational Image Restoration Network

Variational Image Restoration Network.

Learning Weakly Convex Regularizers for Convergent Image-Reconstruction Algorithms

Guided Regularization and its Application for Image Restoration

Supervised Learning of Sparsity-Promoting Regularizers for Denoising

Deeply Aggregated Alternating Minimization for Image Restoration

Inverse problem regularization with hierarchical variational autoencoders

Equivariant Denoisers for Image Restoration

Learnable Nonlocal Self-Similarity of Deep Features for Image Denoising

A Neural-Network-Based Convex Regularizer for Inverse Problems

Batch-Less Stochastic Gradient Descent for Compressive Learning of Deep Regularization for Image Denoising

Variational Bayes image restoration with compressive autoencoders

Non-Local Auto-Encoder With Collaborative Stabilization for Image Restoration.

Understanding Regularization to Visualize Convolutional Neural Networks