Towards a Unified View of Unsupervised Non-Local Methods for Image Denoising: The NL-Ridge Approach

We propose a unified view of unsupervised non-local methods for image denoising that linearly combine noisy image patches. The best methods, established in different modeling and estimation frameworks, are two-step algorithms. Leveraging Stein's unbiased risk estimate (SURE) for the first step...

Full description

Saved in:
Bibliographic Details
Main Authors: Herbreteau, Sebastien, Kervrann, Charles
Format: Conference Proceeding
Language:English
Subjects:
Adaptation models
Deep learning
Estimation
Image denoising
Minimization
Noise measurement
non-local method
Patch-based image denoising
Ridge regression
statistical aggregation
Stein's unbiased risk estimation
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We propose a unified view of unsupervised non-local methods for image denoising that linearly combine noisy image patches. The best methods, established in different modeling and estimation frameworks, are two-step algorithms. Leveraging Stein's unbiased risk estimate (SURE) for the first step and the "internal adaptation", a concept borrowed from deep learning theory, for the second one, we show that our NL-Ridge approach enables to reconcile several patch aggregation methods for image denoising. In the second step, our closed-form aggregation weights are computed through multivariate Ridge regressions. Experiments on artificially noisy images demonstrate that NL-Ridge may outperform well established state-of-the-art unsupervised denoisers such as BM3D and NL-Bayes, as well as recent unsupervised deep learning methods, while being simpler conceptually.
ISSN:2381-8549
DOI:10.1109/ICIP46576.2022.9897992