Discriminative Indexing for Probabilistic Image Patch Priors

Yan Wang¹, Sunghyun Cho², Jue Wang², and Shih-Fu Chang¹

¹Dept. of Electrical Engineering, Columbia University, USA
yanwang@ee.columbia.edu
sodomau@postech.ac.kr

²Adobe Research, USA
sfchang@ee.columbia.edu
juewang@adobe.com

Abstract. Newly emerged probabilistic image patch priors, such as Expected Patch Log-Likelihood (EPLL), have shown excellent performance on image restoration tasks, especially deconvolution, due to its rich expressiveness. However, its applicability is limited by the heavy computation 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 deconvolution methods by up to 40 times, with very little quality compromise.

LNCS 8692, p. 200 ff.

Full article in PDF | BibTeX

lncs@springer.com
© Springer International Publishing Switzerland 2014