Classification of the Driver’s Emotions Using a Convolutional Neural Network

Abstract

The emotional and mental state of the driver is very important in the driving process for safety and security reasons. In fact, there are several factors that affect driving safety, namely fatigue, stress, nervousness, sadness and anger at the wheel. Hence, the need to detect and understand the emotional state of the driver is primordial to promote driving skills such as attention, good judgment, correct decision making and quick reaction time. This paper presents an approach based on a convolutional neural network (CNN) model for the recognition of emotional expression with an accuracy of 66.14% on the public facial expression database FER2013. Afterwards, we realized a real-time emotion recognition system by transferring the skills acquired on static images, which continuously detects the driver's face based on a video camera and then classified the emotion state shown by the driver.