In recent years the computational power of mobile devices has increased dramatically. Multi-core CPUs and graphical processing units are now standard components of most consumer mobile devices. It has become feasible to capture 3D data directly on mobile devices.

In this project we will verify the identity of a person by using a dense 3D point cloud (geometry of face with more than 30000 points) captured by mobile devices. The mobile app for capturing 3D point clouds of faces is provided by ETH Zurich, which is a collaboration partner of this project. A surface computed from a point cloud originally captured by the mobile device is used as identity (person) specific ground truth (gallery image). In this project we investigate which methods are applicable for test data (point sets) captured by mobile devices while still providing state of the art verification rates. Furthermore, the verification process should be solved in a time-critical environment, and processing time should be near real time. An additional challenge is the uncontrolled environment (e.g. lighting and background) under which the data is captured.