An innovative hybrid approach to construct fuzzy-neural network for 3D face recognition system

Abstract

The 2D face recognition systems encounter difficulties in recognizing faces with illumination variations. The depth map of the 3D face data has the potential to handle the variation in illumination of face images. For feature matching an efficient fuzzy-neural technique is proposed. This paper presents a new approach in which the depth maps of the 3D face images, containing the depth information of the face image are used. Since the input images contain the depth information the input to the fuzzy neural network is illumination invariant. Using the normalized depth map and fuzzy-neural network, a fully automatic 3D face recognition system is developed. The system is evaluated on the 3D face databases; the CASIA database. The proposed system efficiently handles the varying lighting effects and provides significant recognition accuracy.

A Supervised Hybrid Methodology for Pose and Illumination Invariant 3D Face Recognition

“…The depth images have some advantages over 3D images [12]. The most important one is that the depth maps are robust to the change of illumination and color because the value on each point represents the depth value which does not depend on illumination or color [14]. The 3D face images contain highly accurate data but it is not feasible to process the large amount of facial data.…”

“…The use of depth images of 3D face images as input images has considerably handled the varying lighting effects. In [14] only frontal images were used and the experiments were carried out to study the performance of fuzzy neural network on depth map and it was found that the use of depth map effectively handles the illumination variations. In this paper extension to the work [14] is proposed and used moment invariants on frontal as well as non-frontal images to develop the rotation invariant face recognition system.…”

“…In [14] only frontal images were used and the experiments were carried out to study the performance of fuzzy neural network on depth map and it was found that the use of depth map effectively handles the illumination variations. In this paper extension to the work [14] is proposed and used moment invariants on frontal as well as non-frontal images to develop the rotation invariant face recognition system. Thus the hybrid intelligent system combines the positive traits of both the systems to incorporate the expressive power of the rules, transparency of view taken in the NN algorithm where they can mimic the behaviour of the human being.…”