Abstract:In this paper we study the topic of signal restoration using complexity regularization, quantifying the compression bit-cost of the signal estimate. While complexity-regularized restoration is an established concept, solid practical methods were suggested only for the Gaussian denoising task, leaving more complicated restoration problems without a generally constructive approach. Here we present practical methods for complexity-regularized restoration of signals, accommodating degradations caused by a known linear operator of an arbitrary form. Our iterative procedure, obtained using the alternating direction method of multipliers (ADMM) approach, addresses the restoration task as a sequence of simpler problems involving L2-regularized estimations and rate-distortion optimizations (considering the squared-error criterion). Further, we replace the rate-distortion optimizations with an arbitrary standardized compression technique and thereby restore the signal by leveraging underlying models designed for compression. Additionally, we propose a shift-invariant complexity regularizer, measuring the bit-cost of all the shifted forms of the estimate, extending our method to use averaging of decompressed outputs gathered from compression of shifted signals. On the theoretical side, we present an analysis of complexity-regularized restoration of a cyclo-stationary Gaussian signal from deterioration by a linear shift-invariant operator and an additive white Gaussian noise. The theory shows that optimal complexity-regularized restoration relies on an elementary restoration filter and compression spreading reconstruction quality unevenly based on the energy distribution of the degradation filter. Nicely, these ideas are realized also in the proposed practical methods. The presented experiments show good results for image deblurring and inpainting using the JPEG2000 and HEVC compression standards.

Restoration by Compression

TwIST-TV regularization based image deblurring method

Experimental Research 0n Image Restoration Algorithms for Special Application

Analysis Of Image Restoration And Evaluation For Diffraction-Degraded Remote Sensing Image

Learning-based image restoration for compressed images

Quality Adaptive Low-Rank Based JPEG Decoding with Applications

Learning-Based Image Restoration for Compressed Image through Neighboring Embedding

Low-Rank Decomposition-Based Restoration of Compressed Images via Adaptive Noise Estimation.

Relaxing Tight Frame Condition in Parallel Proximal Methods for Signal Restoration

High-Accuracy Total Variation for Compressed Video Sensing

Hybrid Regularization Image Deblurring in the Presence of Impulsive Noise

Compressed Image Restoration via External-Image Assisted Band Adaptive PCA Model Learning

Guided Regularization and its Application for Image Restoration

Neural Compression-Based Feature Learning for Video Restoration

Image Restoration Algorithm Based on Sparse Regularized Optimization

Signal Restoration Combining Modified Tikhonov Regularization and Preconditioning Technology

Optimized Pre-Compensating Compression

Near-lossless L-infinity Constrained Multi-rate Image Decompression Via Deep Neural Network

Novel Deconvolution Approach to Image Restoration

Leveraging Bitstream Metadata for Fast, Accurate, Generalized Compressed Video Quality Enhancement

Robust Image Restoration Via Random Projection and Partial Sorted ℓp Norm