The Research And Realization Of Facial Expression Recognition

Facial expression recognition is a developing hot research topic in artificial field. It is one of most challenging problems in the fields of biometric identification, image procession, machine vision, movement tracking, pattern recognition, physiology and psychology. Facial expression recognition is aiming at letting artificial intelligent products recognize human facial expressions and analyze emotions automatically. It is an important part of affective computing and intelligent human-machine interactive, which has a wide range of applications and potential market value.This facial expression recognition system was composed with the following parts: detecting face image from complex background, locating eye position, segmenting and normalizing pure facial expression image, wavelet transformation to reduce the dimension , extracting facial features, classing facial expression .In this thesis ,feature extraction, feature selection and expression classification are studied. Several improved algorithms and methods for these tasks are developed. The performances of our methods are illustrated by simulation experiment results. The major contributions of this paper are as follows:1.Using SMO algorithm to detect face and improved Hough algorithm is studied in order to locate eye position .Geometric preprocessing including rotation, clipping, and scaling , photometric preprocessing including histogram equalization are adopted.2. Comparison of current facial expression recognition algorithm is discussed. The method of using wavelet transformation to reduce the dimension of the features and using the transformation image to substitute for the originality picture, which can reduce the sensitivity to variations of lighting and position.Using Weighted PCA, improved fisher discriminant and PCA/LDA algorithm to train a feature subspace.3. Based on the strategy of making use of the differences between the forms of one expression, one expression is divided into some subexpressions. An algorithm of using dynamic C-Means clustering to generate expression templates and using Parzen algorithm to classify image which needed classify.4.An automatic facial expression recognition prototype system is developed based on Labview and the algorithm studied in this thesis. The system gets images from database. It consists of such modules as face detection, feature extraction and expression classification. At the same time it recognizes 7 expression of happiness, fear, sadness, disgust, anger, surprise, and neutral.