Search

The paper presents a face recognition system based on 3D features with the aim of verifying the identity of subjects accessing a controlled Ambient Intelligence Environment and customizing all the services accordingly. The proposed approach relies on stereoscopic face acquisition and 3D mesh reconstruction avoiding non-automated and expensive 3D scanners, unsuitable for real time applications in general. A bidimensional feature descriptor is extracted from each 3D mesh. It consists in a color image transferring face's 3D features in a 2D space. An automatic weighting mask of each authorized person improves the robustness of recognition in presence of diverse facial expressions and beard. The experiments conducted show high average recognition rate and a measurable effectiveness of both flesh mask and expression weighting mask. [ABSTRACT FROM AUTHOR]