Classification and Prediction of Gender in Facial Images with CNN

Alvarado-Diaz W.; Meneses-Claudio B.; Roman-Gonzalez A. · Lecture Notes in Electrical Engineering (2021) · DOI: 10.1007/978-3-030-72208-1_5

Resumen

Computer technology development, the popularization of artificial intelligence, and facial recognition have become necessary for multiple applications. Both in the military and economic aspects, as it is gradually introduced into people’s lives, for example, in the use of facial recognition to unlock mobile phones. Since the 1990s, gender identification has begun to be studied through a photo of the face; it is worth mentioning that facial gender recognition is challenging in computer vision. This article is made to be applicable in marketing; in this way, it could offer differentiated products according to the clients’ gender. For this purpose, it has used public databases to classify the images of faces in men and women, with the implementation of a Convolutional Neural Network (CNN) model, which it obtained an efficiency in the classification of approximately 97%. It also carried out prediction tests in which the silver model achieved a hit rate of 86.25%.

Instituciones: Universidad de Ciencias y Humanidades