Implementation of a System for 3D Face Detection and Recognition

This paper provides a method for detection and recognition face of man using 3D face extracting and applying 3D features. In this paper, it is used from the Bio-ID package contained 1520 images of man face. A number of these images were used for 2D face modeling; another number was used for 3D face modeling, and another number was used for test. At first, the landmarks on the face image are determined, semi automatically. Then the shape, texture and appearance model of face images is constructed. Using these models and the fast Active Appearance Model search, the landmarks on the test image are determined. More ever, from 24 3D images, obtained by a 3D scanner, the variations of shape, texture and appearance are modeled. Using the 3D models, 2D landmarks and an 3D Initialized Active Appearance Model Search method (3D IAAMS), the 3D frame of face (this frame is described by 3D landmarks) in an image is constructed. These 3D frames with the texture are used for face recognition.