Using the Keras Model for Accurate and Rapid Gender Identification through Detection of Facial Features

As the need for person-aligned, effective, and ethical structures emerges, automatic identification of gender is gaining interest in the era of machine-man interactions. Most of the systems for gender detection were analysed using textual or audiovisual sources. Far too many suggested different approaches for automatic identification of gender using the characteristics acquired from people’s bodies and/or behaviours. But the accuracy has always been a question or a drawback in automated gender detection. In the proposed research work, First, the faces and the facial features which includes eyes, mouth, and nose are detected using Haar Cascade based on Viola-Jones face detection algorithm. Before, the gender detection, the noise is reduced by applying adaptive filters, thereby increasing the accuracy. The obtained facial features are given as the input or test data to the neural network. The neural network is designed to obtain the features and acts as a classifier to detect the genders. Simply, the feature extraction and gender detection is performed using the open source neural network called keras.