3 D Face Acquisition , Modelling and Recognition

This paper presents a novel integrated 3D face verification system. The system comprises of a Structured Light Scanner acquisition stage, with 3D surface fitting and filtering. A-priori information of the human face is utilised to select and extract a region of interest from the face surface, whilst the Iterative Closest Point (ICP) algorithm is employed to establish correspondence between test and target and to compensate for the non-rigid nature of the surfaces. The verification process uses a Euclidean distance classifier across combinations of features extracted from the surfaces and results from a database of 21 subjects are presented.