Ear-to-ear Capture of Facial Intrinsics

We present a practical approach to capturing ear-to-ear face models comprising both 3D meshes and intrinsic textures (i.e. diffuse and specular albedo and normal maps). Our approach is a hybrid of geometric and photometric methods
and requires no geometric calibration. Photometric measurements made in a lightstage are used to estimate view dependent high resolution normal maps. We overcome problems of fixed photometric viewpoint by capturing in multiple poses. We use a multiview reconstruction pipeline of structure-from-motion followed by multiview stereo to compute a base mesh to which the photometric views are registered. We propose a novel approach to robustly stitching the normal maps and intrinsic textures into a seamless, complete and detailed face model. The resulting models provide photorealistic renderings in any view.