Adaptive Image Denoising by Mixture Adaptation
This addresses the problem of improving image denoising accuracy for applications like photography or medical imaging, though it is incremental as it builds on existing prior-based methods with a more rigorous adaptation approach.
The paper tackles image denoising by developing an adaptive learning procedure that adapts a generic patch-based prior from an external database to a specific noisy image using an Expectation-Maximization algorithm, resulting in consistently better denoising performance than non-adaptive methods and superiority over several state-of-the-art algorithms.
We propose an adaptive learning procedure to learn patch-based image priors for image denoising. The new algorithm, called the Expectation-Maximization (EM) adaptation, takes a generic prior learned from a generic external database and adapts it to the noisy image to generate a specific prior. Different from existing methods that combine internal and external statistics in ad-hoc ways, the proposed algorithm is rigorously derived from a Bayesian hyper-prior perspective. There are two contributions of this paper: First, we provide full derivation of the EM adaptation algorithm and demonstrate methods to improve the computational complexity. Second, in the absence of the latent clean image, we show how EM adaptation can be modified based on pre-filtering. Experimental results show that the proposed adaptation algorithm yields consistently better denoising results than the one without adaptation and is superior to several state-of-the-art algorithms.