3D FACE RECOGNITION WITH JOINT TRANSFORM CORRELATORS
Pages 129-132
Alin C. Teușdea
Abstract
Technological development acquiring 3D object spatial shapes generates the problem of discriminating them in three dimensions. There are many methods to generate or to acquire 3D faces. Here we consider a 3D scanned face database with thirteen models without using a mesh rendering. Only the 3D spatial distribution of the points is retained from these models and converted in 256 grey levels, representing a 2D projection image. These 2D projection images are the reference and non-reference images (scene image) of the input joint image for the joint transform correlators. This paper presents the computer simulations results for the optical correlation with three joint transform correlators: modified amplitude joint transform correlator, (MAJTC), modified phase input joint transform correlator, (MPJTC), and preprocessed modified phase input joint transform correlator (preMPJTC).