| Face information include the basic features of the human face,human face identity featureinformation and the expression of the face feature information etc. face information recognition istrying to make the computer, liking humans, have the same vision and the ability to analyze,whichcontains new multidisciplinary area of research. The face information identification is using thecomputer to analyze the image and detect the face area and then to determine its identity and itsexpressions in order to achieve a better intelligent interaction between man and machine. Acomplete information face recognition system consists of the following four components, including:face image acquisition, face detection, face identity and expression feature extraction,face identityand expression recognition and so on.This thesis raise a new idea for face detecting, which combine based on skin color and based onfacial features algorithm, firstly,we use the algorithm based on skin color to detect the area of faceskin, and then checking the area of face using the algorithm based on facial features. ultimately wecould achieve the effect of a high-precision face inspection.In this thesis, raise a new idea for face information identification suitable for fixed users, Theidea of the method is to test image identification, then after identification specific object carriedexpression classification experiments show that the two-step recognition law not only caneffectively improve the pet game recognition speed can be significant with increase expressionrecognition stage of the recognitionrate.This thesis was first put forward the concept of face information recognition, and this is the firsttime use to the human face information recognition system based on Double Step Identificationdeveloping a smart virtual pet game, the smart virtual pet game could Collection the gamer’s faceimages of real-time using the capture camera and recognize the gamer’s Identity and expression. |
PDF Full Text Request