The Reseach On Simulating Face Aging Process

Human face is one of the most universal model of human vision, it is a facial structure that is complex and delicate, such as portable identity, age, sex, wealth of information. The information is the important characteristics and the primary basis for exchange and mutual recognition in people's daily lives. As people age, facial features have changed significantly, which make the human face recognition automatic is more difficult, and now it has become the technology bottleneck of current face recognition . Face aging process simulation system have wide application prospect for improving the performance of face recognition, criminal investigation, as well as film and television entertainment search, human-computer interaction, multimedia and other special effects.As one of the most common patterns of human vision, face is a complex and delicate facial structure that carries many human features such as social status, age, gender, etc. Face is an important feature and primitive basis for exchange and mutual recognition in people's daily lives. As people age, facial features have changed significantly, which make the human face recognition automatic more difficult. Now facial feature has become a technological bottleneck for the current face recognition. Face aging process simulation system has wide application prospect for improving the performance of face recognition, criminal investigation, as well as film and television entertainment search, human-computer interaction, multimedia and other special effects.The main motivation of this thesis is to combine the similarity principle to: (1) study the law of changes in facial aging process; (2) establish the statistical model of face and design algorithm for aging simulation; (3) simulate face aging process automated using computer to predict the future appearance of face The detailed research works includes the following aspects: it is adopted a hierarchical structure model to sample a library of more than 400 pieces black male of 15 to 45-year-old. It was found that the rules of the facial aging process changed non-linearly. Then it is carried out piecewise-linear simulation of the aging process, combined with the principle of similarity of human faces at all levels and finally synthesis the results faces of the aging reversely to achieve the forecast looks.In order to verify the validity of the algorithm, three methods are adopted which are simulation based on the original portrait of the aging process, linear simulation of the aging process and the piecewise-linear simulation based on local similarity, the compared experimental results show that: the method based on local similar simulation of the aging process can more accurately reflect the link with the local features and personality than the method transforming the common characteristics of wrinkles as a compensation. Compared with linear, piecewise-linear simulation of the aging process is more in line with the aging face of the laws of nature. In a certain number of samples, the algorithm can be subjective and objective evaluation.