L.K. Kashapova, A.A. Khrulev, I.N. Spiridonov

When a psychiatrist makes the diagnosis, he uses patient's anamnesis firstly. But the mental status is of ever greater importance. Mental status can be analysed by using tests and simultaneous emotion analysis. At the present day emotion analysis is made as the sum total of the examiner's observations that unfortunately results in the subjectivity of outcome. To eliminate this disadvantage what we need is automatizing of emotion analysis systems. There are a variety of emotional state indicators. They are classified into 3 levels, which are psychologic, physiologic and behavioristic. In psychiatry using of behavior-related indicators are much more significant than the others. Facial mimics is one of the most informative and investigated behavior indicators. It is characterized by the intensity of facial movements, muscle tension, the proportion between separate facial movements, direction of movements and time duration. To make use of the information afforded by facial expressions reliable, valid and efficient methods of measurement are critical. The Facial Action Coding System (FACS) (Ekman & Friesen, 1978) is the most comprehensive, psychometrically rigorous and widely used system for describing facial expression. Actions units used as measurement units are the smallest visually discriminable facial movements. By comparison, other systems are less thorough and often assume a one-to-one mapping between facial expression and emotion. Because of its descriptive power, FACS is regarded by many as the standard measure for facial behavior and it used widely in diverse fields. FACS consists of 44 action units. Thirty are anatomically related to contraction of a specific set of facial muscles. The anatomic basis of the remaining 14 is unspecified. These 14 are reffered to in FACS as miscellaneous actions. For action units that vary in intensity, a 5-point ordinal scale is used to measure the degree of muscle contraction. Prototypic expressions (disgust, fear, joy, surprise, sadness, anger) are coded in the emotional facial action system (EMFACS) made the same authors. By converting FACS codes to EMFACS face images may be coded for emotion-specified expressions. Thus we can recognize emotion by using facial expression, but for enabling rigorous, efficient and quantitative measurement of facial expressions developing an automatic method of facial expression recognition is needed. Firstly, facial gesture parameters set must be developed. Secondly, automatic image processing method for mimics parameters set computation is required. We developed mimics parameters set (MPS) based on facial fiducial points determined in ISO/IEC 19794-5 - Information technology. Biometric data interchange formats. Part 5: Face image data?. 14 dynamic fiducial points and 5 constant fiducial points were used. For quantitative measurement of points deplacement the features labeled as fi (i=1...14) have been implemented. This feature set consists of distances, noted as s(x,y), between fiducial points. s(x,y) is the Eclidean distance between the facial points. Constant points are used as reference points. Distances between these points are used for normalization purposes. After points displacement measurement features are normalized by facial features sizes. This normalization promotes unified measurement scale for different people. So we need automatic image processing for facial expression recognition. Firstly, using video is required. Secondly, the procedure of facial feature extracting consists of a few steps that include face detection, face points location, facial gesture parameters set extraction, facial feature extraction using FACS and facial expression recognition using EMFACS. At the present day we have developed the first three steps.