High-quality hair modeling from a single portrait photo

Abstract

We propose a novel system to reconstruct a high-quality hair depth map from a single portrait photo with minimal user input. We achieve this by combining depth cues such as occlusions, silhouettes, and shading, with a novel 3D helical structural prior for hair reconstruction. We fit a parametric morphable face model to the input photo and construct a base shape in the face, hair and body regions using occlusion and silhouette constraints. We then estimate the normals in the hair region via a Shape-from-Shading-based optimization that uses the lighting inferred from the face model and enforces an adaptive albedo prior that models the typical color and occlusion variations of hair. We introduce a 3D helical hair prior that captures the geometric structure of hair, and show that it can be robustly recovered from the input photo in an automatic manner. Our system combines the base shape, the normals estimated by Shape from Shading, and the 3D helical hair prior to reconstruct high-quality 3D hair models. Our single-image reconstruction closely matches the results of a state-of-the-art multi-view stereo applied on a multi-view stereo dataset. Our technique can reconstruct a wide variety of hairstyles ranging from short to long and from straight to messy, and we demonstrate the use of our 3D hair models for high-quality portrait relighting, novel view synthesis and 3D-printed portrait reliefs.

Head Reconstruction from Internet Photos

“…Visualization of the reprojection error for 3 methods. methods that can use our reconstructed shape as initialization, e.g., in [26] the template was produced manually.…”

Hair-GAN: Recovering 3D hair structure from a single image using generative adversarial networks

“…While single-view hair modeling methods are becoming increasingly popular and important as sing-view, un-calibrated images are widely available on the Internet. Chai et al [16] [17] first introduce the technique of single-view hair modeling by utilizing different kinds of prior knowledge, including layer boundary and occlusion, and shading cues [3]. A major problem of their methods is the lack of control over the geometry at views distant from the input image.…”

“…However, hair modeling, one of the most crucial tasks, still remains challenging due to the complexity and variety of hairstyles in the real world. Methods of single-view hair modeling [1] [2] [3] are considered as a much more user-friendly in comparison with those of multi-view modeling methods [4] [5] [6] [7] [8] [9], which usually require specialized equipments in controlled studio environment and long processing cycles.…”

Game‐ready 3D hair model from a small set of images

“…Input images of a long hairstyle, reconstructed guide hair strands, and our result FIGURE 13 We segment out the hair region within an image and construct a two-dimensional orientation field. Eight such fields are combined to generate a three-dimensional orientation field.…”