Three Dimensional Face Recognition Using SVM Classifier
Mir Hashem Mousavi, Karim Faez, Amin Asghari · 2008
In this paper, we presented a novel approach for automated 3D face recognition using range data. An object recognition system generally consists of two main parts: data registration and data comparison. In first step, the nose tip was used as the reference point and 3D face shape was normalized to standard image size. The 2DPCA was applied to the resultant range data and the corresponding principal images were used as the feature vectors. Classification was carried out by calculating the similarity score between the feature vectors. The SVM classifier was used in choosing the closest match. Recognition rate of 97% rank-four was achieved.