Facial component extraction and face recognition with support vector machines

A method for face recognition is proposed which uses a two-step approach: first, a number of facial components are found, which are then glued together, and the resulting face vector is recognized as representing one of the possible persons. During the extraction step, a wavelet statistics subsystem provides the possible locations of the eyes and mouth, which are used by a support vector machine (SVM) subsystem to extract the facial components. The use of a wavelet statistics subsystem speeds up the recognition process markedly. Both the feature detection SVMs and the wavelet statistics subsystem are trained on a small number of actual images with marked features. Afterwards, a large number of face vectors are constructed, which are then classified with another set of SVM machines