Global facial recognition using gabor wavelet, support vector machines and 3d face models

The present research is aimed to develop an optimal method for face recognition based on wavelet Gabor filtering, feature extraction, and Support Vector Machine (SVM) using the BU-3DFE database containing 3D face models. Process for working with 350 models corresponding to 50 persons, i.e. 10 models per person divided seven for training and three for testing. The proposed technique involves projecting the face models obtained from the BU-3DFE database to the three planes using Matlab 2015a functions, and then, treating them as 2D images for recognition. The aim of this work is to achieve efficient 3D facial recognition with acceptable performance. As a result, the highest obtained value was 97.3% for SVM (kernel cubical). The results obtained for the proposed approach were compared with those of other recent 3D facial recognition methods to evaluate the potential of the former. Contribution of the present research is to facilitate urban security through providing a more efficient way for recognition of people who threaten the peace and tranquility of society, public or private institution, etc.