Facial Expression Synthesis Based On Image Morphing

Human face is one of the most important organs in the human external characteristics, facial expression is a fundamental method to convey emotion information between different people; it is a special means in addition to verbal communication. The psychology research indicates that humanity’s majority of expressions and communications are based on through the expression. Along with the rapid development of computer vision, image processing and computer graphics, facial expression synthesis has been a hot research topic in these fields. Human has complex and abundant face expressions; it is a very challenging topic to synthesize a human’s facial image that is of special expression naturally and realistically. Facial expression synthesis plays an important role in film and television entertainment, electronic games, network communications, human-computer interfaces, virtual host, medical treatment and so on.This thesis mainly studies the facial expression synthesis technology based on image morphing; the purpose is to vividly realize face organ deformation, which enables a neutral face image to be synthesized in a specific facial expression image. The research involves face recognition, image morphing and facial expression synthesis.This thesis proposes a feature points based on face morphing algorithm. The algorithm first realizes face recognition through ASM, its main purpose is to automatically acquire the facial organs feature points which is closely related to facial expression action. And then do image morphing according to these points. This thesis realizes two types of expression synthesis:one kind is with a known neutral face and a selected another face of neutral image and expression image to synthesize the facial expression image, the other kind is with a known neutral face, through several other pieces of neutral and facial expression images to synthesize the corresponding expression face images. The traditional image morphing techniques need to manually mark feature points, which is time consuming. This thesis uses the ASM algorithm to automatically extract feature points and improves efficiency. Our facial expression synthesis method has the advantages as having simple and direct idea, convenient steps and easy implementation and so on.