DESIGNING A CONVOLUTIONAL NEURAL NETWORK FOR GENDER IDENTIFICATION FROM FACIAL IMAGES

Abstract

Gender is among the most important and useful information that can be identified from human facial images. Recent techniques for gender classification problem have been mostly based on deep learning methods and have gained higher results than conventional approaches which are relied on local features extracted from input pictures. In this paper, we focus on building a convolutional neural network and combine several data augmentation methods to build up a gender classification system. Obtained experimental results upon public face image database LFW show that our system achieves high accuracies (97.5%) and is compared with published works in the literature.