A physics-informed alternative to Richardson-Lucy deconvolution across SNR regimes without iteration cutoffs
Open Access
- 30 June 2026
- journal article
- Published by Springer Nature in Nature Communications
- Vol. 17 (1)
Abstract
Richardson-Lucy deconvolution is widely used to restore imaged objects blurred by a point spread function and corrupted by noise and is known to readily overfit noise, leading to high-frequency artifacts. Practical use therefore relies on hand-tuned stopping criteria or ad hoc regularization with limited physical justification. To resolve this problem, we present DeBayes: a rigorous Bayesian deconvolution framework that builds upon a physically accurate image formation model. DeBayes performs deconvolution in the spatial domain, jointly models accurate noise sources, and infers full posterior distributions over the underlying object. It avoids assumptions of sparsity or continuity, yields strictly positive reconstructions, and converges stably without user-tuned regularization parameters or iteration cutoffs. Our method is unsupervised and designed for fast, parallelizable computation, providing a principled alternative to Richardson-Lucy for robust, physics-informed image reconstruction. We demonstrate DeBayes' stable convergence and minimal noise amplification on simulated and experimental images of mitochondria networks in HeLa cells.Keywords
Funding Information
- National Institutes of Health (R35GM148237)
- National Science Foundation (PHY- 2310610)
This publication has 37 references indexed in Scilit:
- A Proximal Approach to Image Deconvolution under Mixed Poisson-Gaussian Noise2026
- Noise amplification and ill-convergence of Richardson-Lucy deconvolutionNature Communications, 2025
- Zero-shot learning enables instant denoising and super-resolution in optical fluorescence microscopyNature Communications, 2024
- Data Modeling for the Sciences2023
- Image Reconstruction in Light-Sheet Microscopy: Spatially Varying Deconvolution and Mixed NoiseJournal of Mathematical Imaging and Vision, 2022
- Whole-cell organelle segmentation in volume electron microscopyNature, 2021
- Compressed sensing effects on quantitative analysis of undersampled human brain sodium MRIMagnetic Resonance in Medicine, 2019
- Fast, long-term, super-resolution imaging with Hessian structured illumination microscopyNature Biotechnology, 2018
- Richardson–Lucy algorithm with total variation regularization for 3D confocal microscope deconvolutionMicroscopy Research and Technique, 2006
- Nonlinear total variation based noise removal algorithmsPhysica D: Nonlinear Phenomena, 1992