A Robust Approach for Gender Recognition Using Deep Learning
Shefali Arora, M. P. S. Bhatia · 2018
Gender information using facial patterns serves an important use in various user interaction applications. This paper proposes an approach based on Convolutional Neural Networks to identify gender from faces. The network is trained using back-propagation and Adam optimization. Using convolution operations, performance of proposed CNN network is evaluated on publicly available CASIA face recognition dataset. The proposed network is able to process 640 ×480 pixel face images in a very less time. A combination of convolutional and max pooling layers is used and classification accuracy of 98.5 percent is achieved in just 50 epochs. The results on CASIA benchmark prove that a superior classification performance can be attained by using Convolutional Neural Networks for gender recognition. The accuracy attained is more than earlier methods used for face recognition.