Convolutional Dictionary Learning: A Comparative Review and New Algorithms

Abstract

Convolutional sparse representations are a form of sparse representation with a dictionary that has a structure that is equivalent to convolution with a set of linear filters. While effective algorithms have recently been developed for the convolutional sparse coding problem, the corresponding dictionary learning problem is substantially more challenging. Furthermore, although a number of different approaches have been proposed, the absence of thorough comparisons between them makes it difficult to determine which of them represents the current state of the art. The present work both addresses this deficiency and proposes some new approaches that outperform existing ones in certain contexts. A thorough set of performance comparisons indicates a very wide range of performance differences among the existing and proposed methods, and clearly identifies those that are the most effective.

Corrected typos in Eq. (18) and (19)

References in corpus (6)

• Convolutional Dictionary Learning: Acceleration and Convergence
• Convolutional Sparse Coding for High Dynamic Range Imaging
• ADMM Penalty Parameter Selection by Residual Balancing
• Scalable Online Convolutional Sparse Coding
• First and Second Order Methods for Online Convolutional Dictionary Learning
• Online Convolutional Dictionary Learning

Cited by in corpus (41)

• Image Reconstruction: From Sparsity to Data-adaptive Methods and Machine Learning
• An Online Plug-and-Play Algorithm for Regularized Image Reconstruction
• Learned Convolutional Sparse Coding
• Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration
• Dictionary Learning for Adaptive GPR Landmine Classification
• Deep Residual Autoencoders for Expectation Maximization-inspired Dictionary Learning
• Systematic Review on Learning-based Spectral CT
• Bilevel methods for image reconstruction
• Learning Filter Bank Sparsifying Transforms
• From Symmetry to Geometry: Tractable Nonconvex Problems
• Efficient ADMM-based Algorithms for Convolutional Sparse Coding
• Learning Deep Analysis Dictionaries for Image Super-Resolution
• A convex variational model for learning convolutional image atoms from incomplete data
• Heterogeneous Domain Adaptation with Adversarial Neural Representation Learning: Experiments on E-Commerce and Cybersecurity
• Fast Convolutional Dictionary Learning off the Grid
• Group Invariant Dictionary Learning
• PSF Estimation in Crowded Astronomical Imagery as a Convolutional Dictionary Learning Problem
• Unrolled Compressed Blind-Deconvolution
• Interpretable Classification of Bacterial Raman Spectra with Knockoff Wavelets
• Image Super-Resolution via RL-CSC: When Residual Learning Meets Convolutional Sparse Coding
• Quantitative MR Image Reconstruction using Parameter-Specific Dictionary Learning with Adaptive Dictionary-Size and Sparsity-Level Choice
• Adversarial Defense by Stratified Convolutional Sparse Coding
• Analysis of the Optimization Landscapes for Overcomplete Representation Learning
• Gaussian Process Convolutional Dictionary Learning
• Short-and-Sparse Deconvolution -- A Geometric Approach
• Dynamic Upsampling of Smoke through Dictionary-based Learning
• Score-based Diffusion Models for Bayesian Image Reconstruction
• Spike Sorting by Convolutional Dictionary Learning
• Convolutional Dictionary Regularizers for Tomographic Inversion
• CRNet: Image Super-Resolution Using A Convolutional Sparse Coding Inspired Network
• Adaptive Clustering of Robust Semantic Representations for Adversarial Image Purification
• On the Preservation of Spatio-temporal Information in Machine Learning Applications
• Modern GPR Target Recognition Methods
• Discriminative reconstruction via simultaneous dense and sparse coding
• Fast Rotational Sparse Coding
• Efficient Consensus Model based on Proximal Gradient Method applied to Convolutional Sparse Problems
• Convolutional Dictionary Learning in Hierarchical Networks
• Multivariate Convolutional Sparse Coding with Low Rank Tensor
• Convolutional dictionary learning based auto-encoders for natural exponential-family distributions
• Two-layer Residual Sparsifying Transform Learning for Image Reconstruction
• Tensor Convolutional Sparse Coding with Low-Rank activations, an application to EEG analysis