Learning Deep Gradient Descent Optimization for Image Deconvolution
- 17 February 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 31 (12) , 5468-5482
Abstract
As an integral component of blind image deblurring, non-blind deconvolution removes image blur with a given blur kernel, which is essential but difficult due to the ill-posed nature of the inverse problem. The predominant approach is based on optimization subject to regularization functions that are either manually designed or learned from examples. Existing learning-based methods have shown superior restoration quality but are not practical enough due to their restricted and static model design. They solely focus on learning a prior and require to know the noise level for deconvolution. We address the gap between the optimization- and learning-based approaches by learning a universal gradient descent optimizer. We propose a recurrent gradient descent network (RGDN) by systematically incorporating deep neural networks into a fully parameterized gradient descent scheme. A hyperparameter-free update unit shared across steps is used to generate the updates from the current estimates based on a convolutional neural network. By training on diverse examples, the RGDN learns an implicit image prior and a universal update rule through recursive supervision. The learned optimizer can be repeatedly used to improve the quality of diverse degenerated observations. The proposed method possesses strong interpretability and high generalization. Extensive experiments on synthetic benchmarks and challenging real-world images demonstrate that the proposed deep optimization method is effective and robust to produce favorable results as well as practical for real-world image deblurring applications.Keywords
All Related Versions
Funding Information
- ARC Discovery Project (DP160100703, DP200103797)
- Chang Jiang Scholars Program of China (100017GH030150, 15GH0301)
This publication has 40 references indexed in Scilit:
- Cascades of Regression Tree Fields for Image RestorationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2015
- The Pascal Visual Object Classes Challenge: A RetrospectiveInternational Journal of Computer Vision, 2014
- Proximal AlgorithmsFoundations and Trends® in Optimization, 2014
- Fast Alternating Direction Optimization MethodsSIAM Journal on Imaging Sciences, 2014
- A Machine Learning Approach for Non-blind Image Deconvolution2013
- A New Alternating Minimization Algorithm for Total Variation Image ReconstructionSIAM Journal on Imaging Sciences, 2008
- Image and depth from a conventional camera with a coded apertureACM Transactions on Graphics, 2007
- Fields of Experts: A Framework for Learning Image Priors2005
- Image quality assessment: from error visibility to structural similarityIEEE Transactions on Image Processing, 2004
- Nonlinear image recovery with half-quadratic regularizationIEEE Transactions on Image Processing, 1995