Author : Ritu Chaturvedi and Prof. K. C. Waghmare

Sentiment Analysis is an intellectual process of extricating user’s feelings and emotions. Automatic recognition of face and identifying emotions of individual plays significant role in the development of human being- computing machine interaction (HCI).  Facial expression recognition in real time should be accurate and efficient. The proposed approach consists of two modules. The first module detects the face in the group of people and in the second module the classification of emotion is performed. To detect face Haar feature based cascade classifier is used,  since it is real time and performs better after compared to other face detectors. Support Vector Machine, KNN and  Neural Network classifiers are used for training and classification of emotion. Model is trained on Indian Spontaneous Emotion Dataset (ISED). Model is trained to detect 7 emotions i.e. Angry, Fear, Disgust, Sad, Happy, Surprise,  and Neutral.

Keywords: Face Detection, Person Identification, Emotion Recognition, Neural Network, Local Binary Pattern Histogram, Sentiment Analysis