Automatic Age Classification of Prospective Voters Using Deep Convolutional Neural Network.

Abstract

Age estimation from images of the face, has gained more attention in recent as it is favorable in some realworld applications. In this work, we address the problem of Under-Age registration/voting in Nigeria Electoral system, which has been a major menace, hindering a free and fair election in the country since the assumption of the Democratic system of Government. To this end, a pretrained VGG-16 Deep Convolutional Neural Network while comparing two optimization algorithms without image preprocessing is employed to both extract features from image(s) of prospective voters and classify same under the established age classification group, as eligible or not to exercise their civil right. In light of this, a classification accuracy of 77.67% is achieved with the model.