Generalized SURE for Exponential Families: Applications to Regularization

Abstract

Stein's unbiased risk estimate (SURE) was proposed by Stein for the independent, identically distributed (iid) Gaussian model in order to derive estimates that dominate least-squares (LS). In recent years, the SURE criterion has been employed in a variety of denoising problems for choosing regularization parameters that minimize an estimate of the mean-squared error (MSE). However, its use has been limited to the iid case which precludes many important applications. In this paper we begin by deriving a SURE counterpart for general, not necessarily iid distributions from the exponential family. This enables extending the SURE design technique to a much broader class of problems. Based on this generalization we suggest a new method for choosing regularization parameters in penalized LS estimators. We then demonstrate its superior performance over the conventional generalized cross validation approach and the discrepancy method in the context of image deblurring and deconvolution. The SURE technique can also be used to design estimates without predefining their structure. However, allowing for too many free parameters impairs the performance of the resulting estimates. To address this inherent tradeoff we propose a regularized SURE objective. Based on this design criterion, we derive a wavelet denoising strategy that is similar in sprit to the standard soft-threshold approach but can lead to improved MSE performance.

to appear in the IEEE Transactions on Signal Processing

References in corpus (1)

• Blind Minimax Estimation

Cited by in corpus (46)

• A SURE Approach for Digital Signal/Image Deconvolution Problems
• Unknown sparsity in compressed sensing: Denoising and inference
• Unsupervised Learning with Stein's Unbiased Risk Estimator
• Deep Learning for space-variant deconvolution in galaxy surveys
• Covariance estimation in decomposable Gaussian graphical models
• Estimating network degree distributions under sampling: An inverse problem, with applications to monitoring social media networks
• Hyperspectral Subspace Identification Using SURE
• Applications of Compressed Sensing in Communications Networks
• Bilevel methods for image reconstruction
• Local Behavior of Sparse Analysis Regularization: Applications to Risk Estimation
• Training Deep Learning Based Denoisers without Ground Truth Data
• Effective New Methods for Automated Parameter Selection in Regularized Inverse Problems
• ENSURE: A General Approach for Unsupervised Training of Deep Image Reconstruction Algorithms
• Tuning-free Plug-and-Play Proximal Algorithm for Inverse Imaging Problems
• Solving Inverse Problems with Score-Based Generative Priors learned from Noisy Data
• Estimating Unknown Sparsity in Compressed Sensing
• A parallel & automatically tuned algorithm for multispectral image deconvolution
• Parameter-free and fast nonlinear piecewise filtering. Application to experimental physics
• Deep Learning Techniques for Inverse Problems in Imaging
• Deep Learning Meets Adaptive Filtering: A Stein's Unbiased Risk Estimator Approach
• Robust multi-coil MRI reconstruction via self-supervised denoising
• Degrees of Freedom Analysis of Unrolled Neural Networks
• Maximum likelihood estimation of regularisation parameters in high-dimensional inverse problems: an empirical Bayesian approach. Part I: Methodology and Experiments
• Degrees of Freedom in Deep Neural Networks
• Sampling-based Estimation of In-degree Distribution with Applications to Directed Complex Networks
• TFPnP: Tuning-free Plug-and-Play Proximal Algorithm with Applications to Inverse Imaging Problems
• Equivariant Bootstrapping for Uncertainty Quantification in Imaging Inverse Problems
• Robust Equivariant Imaging: a fully unsupervised framework for learning to image from noisy and partial measurements
• Unsupervised parameter selection for denoising with the elastic net
• Scale-Equivariant Imaging: Self-Supervised Learning for Image Super-Resolution and Deblurring
• Hyperparameter selection for Discrete Mumford-Shah
• RHAPSODIE: Reconstruction of High-contrAst Polarized SOurces and Deconvolution for cIrcumstellar Environments
• Exponential Family Trend Filtering on Lattices
• MRI Recovery with Self-Calibrated Denoisers without Fully-Sampled Data
• Rethinking Deep Image Prior for Denoising
• Path Thresholding: Asymptotically Tuning-Free High-Dimensional Sparse Regression
• Model Adaptation for Image Reconstruction using Generalized Stein's Unbiased Risk Estimator
• Multi-frequency image reconstruction for radio-interferometry with self-tuned regularization parameters
• Signal Denoising Using the Minimum-Probability-of-Error Criterion
• yaglm: a Python package for fitting and tuning generalized linear models that supports structured, adaptive and non-convex penalties
• Unbiased Bregman-Risk Estimators: Application to Regularization Parameter Selection in Tomographic Image Reconstruction
• A Goodness-of-Fit Test for Sampled Subgraphs
• Statistical inference for block sparsity of complex signals
• Deep Image Prior using Stein's Unbiased Risk Estimator: SURE-DIP
• SUREMap: Predicting Uncertainty in CNN-based Image Reconstruction Using Stein's Unbiased Risk Estimate
• A Variational Bayesian Approach for Image Restoration. Application to Image Deblurring with Poisson-Gaussian Noise