Αντίληψη και αναγνώριση συναισθήματος σε εικόνες προσώπου με εφαρμογή σε συστήματα συναισθηματικής αλληλεπίδρασης ανθρώπου - υπολογιστή

Abstract

Faces provide a wide range of information about a person's identity, race, sex, age, and emotional state. In this thesis, we study the perception of facial expressions of emotion in our aim at developing a fully automated facial expression recognition system. Our studies begin with a research in the literature about the emotion perception from the scientific - psychological and medical - point of view. Based on these studies, we came up with the following assumptions: (1) a number of brain parts play a significant role in emotion perception and expression, (2) there are six `basic emotions', namely: `anger', `disgust', `fear', `happiness', `sadness' and `surprise' and, (3) there is cultural specificity in emotion perception and expression. The latter is further strengthened by our own empirical studies conducted to humans. Specifically, we developed two different questionnaires, wherein participants were shown face im- ages and were asked to classify the emotion. In the first questionnaire we used images gathered from the web, while in the second questionnaire we used images of Greeks forming an expression. The difference in the success rates further demonstrates that there is cultural specificity in the ways people understand and express the emotion. Moreover, from our empirical studies, we were able to identify the emotion classes that are present during a typical human-computer interaction session, which are namely: `neutral', `happiness', `sadness', `surprise', `anger', `disgust' and `boredom-sleepiness'. Towards building our facial expression recognition system, we constructed our own face image database, which consisted of a two different sets of face images in front and side view: low quality images which were acquired by using web cameras and high quality face images which were acquired by using digital cameras of high resolution. Finally, we developed our own facial expression recognition system, which consists of two modules: (1) a face detection subsystem and, (2) a facial expression recognition subsystem. Our face detection subsystem is based on neural network-based classifiers. For our facial expression recognition subsystem, we considered neural network-based and other classifiers, but we concluded that Support Vector Machine-based Classifiers demonstrated better results. The feature extraction process, performance evaluations and test results are demonstrated and analyzed.