A Novel Architecture combining convolutional neural network and support vector machine for expression recognition in driving environment

Abstract Emotion monitoring is one of the key parameters in our day-to-day lives. The person's state of mind can be readable from their emotions. Expressing emotions usually happens in two ways of communication, namely verbally and non-verbally. Verbal communication is easy to communicate and understand between people in most situations, whereas nonverbal communication, like showing emotions, is difficult to understand in some cases. These mental emotions can control the person to go through either good situations or bad situations, especially if the emotions of the driver are considered, the good emotions like happiness, neutral can make the driver be in a good mental state can drive the vehicle safely however emotions like sad, angry, disgust, afraid are the emotions that influence the driver capabilities can cause accidents. To avoid this, advanced driver assistance systems are introduced in automotive vehicles to assist the driver for various functions for safety purposes. In addition to that, Emotion monitoring in advanced driver assistance systems can be accomplished by using facial expression recognition technology, which is evolved by training a convolutional neural network with applying machine learning and deep learning approaches to detect faces and predict emotions from the feature obtained from the networks. To achieve this, we proposed a novel architecture combining a convolutional neural network and support vector machine for expression recognition in the driving environment. The experimental results demonstrate the effectiveness of the proposed architecture by achieving remarkable accuracy.