Automatic prediction of age-group from frontal facial images

Abstract

Methods to automatically assess the age group of a person using his/her frontal facial image are proposed in this paper. This work is done for three major ethnicities: African, American and Asian with five different age-groups such as (1-10 years), (11-30 years), (31-50years), (51-70 years), (71-100 years). The performances of the classifiers were tested with face images of African, American and Asian population belonging to both genders. For this, first the facial parts such as the left eye, right eye, nose, mouth etc., are detected using the well-known Viola Jones Object Detection technique.450 sample images of the FERET database were considered for this study. Histogram of Gradient (HoG) and face-structure features are extracted and modeled using ANN and SVM. The efficiency of the proposed methods was tested with the facial images of various races belonging to different age-group and gender. Artificial neural network gave an accuracy of 92.10% whereas support vector machine gave an improved accuracy of 94.60%.