Composition-aided Sketch-realistic Portrait Generation

Sketch portrait generation is of wide applications including digital entertainment and law enforcement. Despite the great progress achieved by existing face sketch generation methods, they mostly yield blurred effects and great deformation over various facial parts. In order to tackle this challenge, we propose a novel composition-aided generative adversarial network (CA-GAN) for sketch portrait generation. First, we utilize paired inputs including a face photo and the corresponding pixel-wise face labels for generating the portrait. Second, we propose an improved pixel loss, termed compositional loss, to focus training on hard-generated components and delicate facial structures. Moreover, we use stacked CA-GANs (stack-CA-GAN) to further rectify defects and add compelling details. Experimental results show that our method is capable of generating identity-preserving, sketch-realistic, and visually comfortable sketch portraits over a wide range of challenging data, and outperforms existing methods. Besides, our methods show considerable generalization ability.