Abstract

Newly emerged probabilistic image patch priors, such as Ex- pected Patch Log-Likelihood (EPLL), have shown excellent performance on image restoration tasks, especially deconvolution, due to its rich ex- pressiveness. However, its applicability is limited by the heavy compu- tation involved in the associated optimization process. Inspired by the recent advances on using regression trees to index priors defined on a Conditional Random Field, we propose a novel discriminative indexing approach on patch-based priors to expedite the optimization process. Specifically, we propose an efficient tree indexing structure for EPLL, and overcome its training tractability challenges in high-dimensional spaces by utilizing special structures of the prior. Experimental results show that our approach accelerates state-of-the-art EPLL-based deconvolu- tion methods by up to 40 times, with very little quality compromise.