Gender Classification Based on Face Recognition using Convolutional Neural Networks (CNNs)

Abstract

Biometrics are physical or behavior characteristics of human that can be used to identify someone. One form of physical characteristics possessed by humans is fingerprints, retinal scanning, face and hand geometry, while one form of behavioral characteristics possessed by humans is handwriting, signatures, mouse usage analysis, walking patterns, etc. Basically, physical characteristics are more easily observed than behavioral characteristics. Therefore, physical characteristics are more often used in many aspects of security. One of the most common physical characteristics is face. By seeing the face, we can find out or predict how old they are, their gender and even their expression. However, there are still many mistakes in predicting a gender through person's face. In fact, there are still many crimes in falsifying self-identity (such as gender). So, we need a method that is able to classify identity (gender) based on a person's face appropriately. One method that can be used is Convolutional Neural Networks (CNNs). Later, CNNs will classify a person's gender (male / female) based on a person's face image data. And based on.