Gender Recognition from Facial Images using Convolutional Neural Network
Shubham Mittal, Shiva Mittal · 2019
Gender is one of the facial attributes whose identification serves crucial step in systems including smart billboards, emotion recognition and surveillance systems. This paper explores a deep learning based solution for automatic detection of gender from face images of a well-balanced dataset. The solution involves a transfer learning framework where the knowledge is reused out of a deep learning model which performed quite well in classification task of other domain. We investigate the reusability of Visual Geometry Group-16 (VGG16), an off-the-shelf Convolutional Neural Network (CNN) model which is pre-trained upon large dataset of natural images. Fine-tuning a segment of the CNN upon the moderately sized dataset yields better performance than obtained from the state-of-the-art approaches when implemented upon the public LFW-Gender dataset.