Building a 3-D Appearance Model of the Human Face
Karl Sjöstrand, Rasmus Larsen, Brian Lading · 2003
This paper describes a method for building an appearance model from three-dimensional data of human faces. The data consists of 3-D vertices, polygons and a texture map. The method uses a set of nine manually placed landmarks to automatically form a dense correspondence of thousands of points. This makes sure the model is able to capture the subtle details of a face. The model can be used for face segmentation and fully automated face registration.