RESEARCH POSTER PRESENTATION DESIGN © 2012

www.PosterPresentations.com This paper addresses the problem of learning over-complete dictionaries for the coupled feature spaces, where the learned dictionaries also reflect the relationship between the two spaces. A Bayesian method using a beta process prior is applied to learn the over-complete dictionaries. Compared to previous couple feature spaces dictionary learning algorithms [1][2][3], our algorithm not only provides dictionaries that customized to each feature space, but also adds more consistent and accurate mapping between the two feature spaces. We compare the proposed approach to several state-of-the-art dictionary learning methods by applying this method to single image super-resolution. Advantages of the algorithm 1. The proposed algorithm is able to learn sparse representations that correspond to the same dictionary atoms with the same sparsity but different values in coupled feature spaces, thus bringing consistent and accurate mapping between coupled feature spaces. 2. The number of dictionary atoms and their relative importance may be inferred non-parametrically. Potential Applications Singe image super-resolution, photo-sketch synthesis, image style transfer, intrinsic image analysis, compressed sensing. 1. Problem formulation There exists a mapping function that maps from feature space to feature space. The relation of the dictionaries and the observations and the relation of the two feature spaces can be described as