Product of Gaussian Mixture Diffusion Models
Preprint
Abstract
In this work we tackle the problem of estimating the density $ f_X $ of a random variable $ X $ by successive smoothing, such that the smoothed random variable $ Y $ fulfills the diffusion partial differential equation $ (\partial_t - Δ_1)f_Y(\,\cdot\,, t) = 0 $ with initial condition $ f_Y(\,\cdot\,, 0) = f_X $. We propose a product-of-experts-type model utilizing Gaussian mixture experts and study configurations that admit an analytic expression for $ f_Y (\,\cdot\,, t) $. In particular, with a focus on image processing, we derive conditions for models acting on filter-, wavelet-, and shearlet responses. Our construction naturally allows the model to be trained simultaneously over the entire diffusion horizon using empirical Bayes. We show numerical results for image denoising where our models are competitive while being tractable, interpretable, and having only a small number of learnable parameters. As a byproduct, our models can be used for reliable noise level estimation, allowing blind denoising of images corrupted by heteroscedastic noise.
Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Score-based diffusion models for accelerated MRIMedical Image Analysis, 2022
- Generalized Gaussian scale mixtures: A model for wavelet coefficients of natural imagesSignal Processing: Image Communication, 2018
- Mathematical Image Processing2018
- On the density of the sum of two independent Student -random vectorsStatistics & Probability Letters, 2010
- Multiple shrinkage and subset selection in waveletsBiometrika, 1998
- Wavelet-based statistical signal processing using hidden Markov modelsIEEE Transactions on Signal Processing, 1998
- Filters, Random Fields and Maximum Entropy (FRAME): Towards a Unified Theory for Texture ModelingInternational Journal of Computer Vision, 1998
- Generalized cross validation for wavelet thresholdingSignal Processing, 1997
- A theory for multiresolution signal decomposition: the wavelet representationIEEE Transactions on Pattern Analysis and Machine Intelligence, 1989
- Sulle funzioni ipergeometriche confluentiAnnali di Matematica Pura ed Applicata (1923 -), 1947