摘要

In this paper, we propose a unified framework for fully automatic face photo stylization based on a single style exemplar. Constrained by the "single-exemplar" condition, where the numbers and varieties of patch samples are limited, we introduce flexibility in sample selection while preserving identity and content of the input photo. Based on the observation that many styles are characterized by unique color selections and texture patterns, we employ a two-phase procedure. The first phase searches a dense and semantic-aware correspondence between the input and the exemplar images, so that colors in the exemplar can be transferred to the input. The second phase conducts edge-preserving texture transfer, which preserves edges and contours of the input and mimics the textures of the exemplar at multiple scales. Experimental results demonstrate compelling visual effects and notable improvements over other state-of-the-art methods which are adapted for the same task.

  • 出版日期2017-11

全文