What is it about?
Image patch priors become a popular tool for image denoising. The Gaussian Mixture Model (GMM) is remarkably effective in modelling natural image patches. However, GMM prior learning using the Expectation Maximization (EM) algorithm is sensitive to the initialization, often leading to low convergence rate of parameter estimation. In this paper, a novel sampling method called random neighbourhood resampling (RNR) is proposed to improve the accuracy and efficiency of parameter estimation. An enhanced GMM (EGMM) learning algorithm is further developed by incorporating RNR into the EM algorithm to initialize and update the GMM prior. The learned EGMM prior is applied in the expected patch log likelihood (EPLL) framework for image denoising.
Featured Image
Why is it important?
In this paper, we propose an enhanced Gaussian Mixture Model (EGMM) learning algorithm based on a novel sampling method called Random Neighborhood Resampling (RNR) to efficiently learning GMM prior in Expected Patch Log Likelihood (EPLL) framework for image denoising. The major contribution of the method is that it solves sensitive initialization and improves the convergence rate of parameter estimation, and updates the image patch adaptively in image denoising.
Perspectives
Read the Original
This page is a summary of: Image Patch Prior Learning Based on Random Neighborhood Resampling for Image Denoising, IET Image Processing, December 2019, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/iet-ipr.2018.5403.
You can read the full text:
Contributors
The following have contributed to this page