A Study on the Face Recognition based on the Face Geometrical Characteristics

This paper has analyzed face characteristics and designed the face features, then it extracts the features and constructed the feature values, and at last, it realizes face recognition through matching of similarity of the feature values. According to the obvious difference of the five sense organs in color, using rgb color space to analyze the distribution range of the five sense organs, and in accordance with the statistical results, the five sense organs in the face are segmented and geometrical parameters of the relevant facial parts are obtained. Construction face feature values with the feature values formed with the proportion of these parameters. Face recognition is to match the feature values. As the extracted feature values have different stability and contribution to the recognition, this paper grants different coefficients to the components, and then uses weighted similarity calculation method to work out the similarity value. Finally, this paper set the maximum similarity matching as the basis for face recognition. For the selection of feature components, we have comprehensively considered the overall features and partial features of face and combined the face shape features and the features of the five sense organs. The experiment shows that these features can reflect the individual features of human face, which can be treated as an effective basis for face recognition, thus to realize face recognition.