Face Sketch Synthesis Based on Local and Nonlocal Similarity Regularization

Face sketch synthesis plays an important role in public security and digital entertainment. In this paper, wepresent a novel face sketch synthesis method via local similarity and nonlocal similarity regularization terms. The local similarity can overcome the technological bottlenecks of the patch representation scheme intraditional learning-based methods. It improves the quality of synthesized sketches by penalizing the dissimilartraining patches (thus have very small weights or are discarded). In addition, taking the redundancy of imagepatches into account, a global nonlocal similarity regularization is employed to restrain the generation of thenoise and maintain primitive facial features during the synthesized process. More robust synthesized resultscan be obtained. Extensive experiments on the public databases validate the generality, effectiveness, androbustness of the proposed algorithm. KCI Citation Count: 1