3D Facial Landmark Detection & Face Registration A 3D Facial Landmark Model & 3D Local Shape Descriptors Approach

In this Technical Report a novel method for 3D landmark detection and pose estimation suitable for both frontal and side 3D facial scans is presented. It utilizes 3D information by using 3D local shape descriptors to extract candidate interest points that are subsequently identified and labeled as anatomical landmarks. The shape descriptors include the shape index, a continuous map of principal curvature values of 3D objects, the extrusion map, a measure of the extruded areas of a 3D object and the spin images, local descriptors of the object’s 3D point distribution. However, feature detection methods which use general purpose shape descriptors cannot identify and label the detected candidate landmarks. Therefore, the topological properties of the human face need to be taken into consideration. To this end, we use a Facial Landmark Model (FLM) of facial anatomical landmarks. Candidate landmarks, irrespectively of the way they are generated, can be identified and labeled by matching them with the corresponding FLM. The proposed method is evaluated using an extensive 3D facial database, and achieves high accuracy even in challenging scenarios.