Automatic age detection based on facial images

Abstract

In this paper, the proposed implementation of a soft-biometric system for automatic age detection from facial images is described. In order to do this, the method followed was that of a classical biometric system. The first step is preprocessing, to enhance the feature extraction. The next step is the parameterization, where techniques like wavelet transformed, discrete cosine transformed or local binary patterns were used. And finally, the last step is the classification system, implemented by Support Vector Machines.