Different face regions detection based facial expression recognition

Human computer interaction (HCI) is an important element of design of automation systems. Of late, facial expression recognition (FER) has been accepted to be an integral component of upcoming HCI systems. In this paper, we propose a method to detect a portion of a face which is a part of human image. We also propose a method to detect selected face regions automatically where a nominal deformation of face muscle is observed. Selected face regions includes mouth, eyebrow and eye which are detected automatically based on their different individual properties. Next, we extract facial features from the face regions which are based on relative displacement of face muscles. Then we employ Hidden Markov Model (HMM) for expression classification. HMM classifier is subjected to give better reliability. We have found satisfactory recognition accuracy with our proposed method in expression recognition and this might be an effective addition to the research area of FER.