Voice-based Gender and Age Recognition System

Abstract

The ability to detect gender and age from voice is a valuable tool in a variety of applications, like voice-based biometric identification, natural language processing, and speech recognition. Recent advances in Deep Learning have enabled the development of highly accurate gender and age detection models. In this paper, the discussion is about the Machine Learning based gender and age detection model using voice. The various approaches used to extract features from speech, and the data-set used for model evaluation and classification are obtained using different Machine Learning algorithms. The discussion is about the opportunities and challenges in this area of research. It is concluded by highlighting some of the open challenges and future directions in this field. Age prediction from voice using a grid search pipeline is a Machine Learning technique that uses a range of algorithms to detect the age of a person using their voice. In the proposed model, RobustScalar, Principal component analysis (PCA), and Logistic Regression algorithms are used. The grid search pipeline uses a combination of models to identify the best age prediction algorithm for a given data-set. For Gender prediction sequential model with 5 hidden layers has been used. The results were obtained based on the trained model for the common voice data-set with an accuracy of around 91% for gender and 59% for age.