Learning 3D face reconstruction from a single sketch

3D face reconstruction from a single image is classic computer vision problem with many applications. However, most works achieve photos, and little attention has been paid to other portrait forms. In this paper, we propose learning-based approach reconstruct sketch. To overcome the of no paired sketch-3D data for supervised learning, introduce photo-to-sketch synthesis technique obtain training data, dual-path architecture synergistic both sketches photos. We further novel line loss function refine characteristic details depicted by lines in well preserved. Our method outperforms state-of-the-art approaches terms sketches. also demonstrate use our easy editing on models.