Face Sketch-to-Photo Synthesis from Simple Line Drawing

Face sketch-to-photo synthesis has attracted increasing attention in recent years for its useful applications on both digital entertainment and law enforcement. Although great progress has been made, previous methods only work on face sketches with rich textures which are not easily to obtain. In this paper, we propose a robust algorithm for synthesizing a face photo from a simple line drawing that contains only a few lines without any texture. In order to obtain a robust result, firstly, the input sketch is divided into several patches and edge descriptors are extracted from these local input patches. Afterwards, an MRF framework is built based on the divided local patches. Then a series of candidate photo patches are synthesized for each local sketch patch based on a coupled dictionary learned from a set of training data. Finally, the MRF is optimized to get the final estimated photo patches for each input sketch patch and a realistic face photo is synthesized. Experimental results on CUHK database have validated the effectiveness of the proposed method.