Recognition of Facial Expressions through Live Camera

Abstract

Facial expressions of humans are an important and easy way of displaying sentiments. In this computer vision, automatic analysis of unspoken sentiments has become a very challenging and interesting task that has numerous applications in product automation and marketing, psychology, etc. Since, expressing one’s emotions through expression varies from person to person, this task is very difficult. The various applications of computer vision have become doable due to advancements in machine learning. This work proposes a method to detect the emotions of a person through a live camera where the system does not capture any images and does not store any images. It only compares facial features with a data set and shows one expression in text format that is happy, sad, angry, etc. To determine the emotions of a person, we use convolutional neural network. In this proposed work, the problem of facial sentiment analysis will be solved by an architecture known as the Convolutional neural network (CNN). The Facial expression recognition (FER-2013) public data set will be used for training and testing. To accomplish this task, we will first pre-process the data, then extract the features and finally classify them by well-trained model network.