Gender Recognition using in-built Inertial Sensors of Smartphone

The research work on user authentication using biometric parameters like face, iris, fingerprint, and gait, have been vastly carried out over the last few decades. Also, the researchers are putting great efforts in the field of gender recognition of a person. Predicting the gender of a person by considering the biometric parameters described above has been the topic of interest in many areas like security, forensic, e-commerce applications, investigation, and some smart automation systems. In this paper, we have presented an analytical approach for gender recognition based on the information of human gait obtained from the inertial sensors of a smartphone, and the comparative analysis of various machine learning classification and regression algorithms on the recognition performance. The overall aim of our research is to provide an insight and to improve the preciseness of the gender recognition system. Moreover, the proposed system can enhance the processing speed of the biometric recognition systems by determining the gender of the person at the initial stage to reduce the search area. We have achieved the state of the art performance and obtained an accuracy of 96.3% using the proposed method.