The Design And Implementation Of Children Face Aging Simulation System

Face aging is the inevitable, irreversible and non-artificially-controllable slow changing process of facial appearance occurring along with the age increases, which is the co-consequence of genetic heredity and exterior living environment. Computer digital simulation of human face aging process can be applied in many fields, in which children face aging simulation technology plays a vital role in predicting the growth of children, searching for missing children and providing entertainment value. However, so far there’s no public-shared human face aging simulation system based on specific age for children nationwide or worldwide.Facing this problem, this thesis researched on the algorithm of children face aging simulation, children face age estimation and other related issues, completed the design of children face aging simulation system, and achieved the corresponding client-end application for cellphone and web. The main work and contribution of this thesis are summarized as the following aspects:(1) collected face pictures of Asian children and preliminary established an Asia children face image database;(2) considered different classification algorithms, carried out comparative experiment on the main-stream face feature extraction algorithm and found out the children face age estimation method suitable for this thesis, which utilized active appearance models to present face features and used support vector machine to estimate the age by group;(3) used prototype-based face aging simulation algorithm and achieved the age-associated children face aging simulation;(4) used open-source Web framework Flask to encapsulate the core algorithm such as children face age estimation algorithm and children face aging simulation algorithm, and then realized children face aging simulation system;(5) designed client-side cross-platform solution, achieved client-side application of face aging simulation for Android cellphone and web, which provided the functions such as locating feature points of children face image, estimating age and simulating human face aging based on specific age.(6) This thesis also tries to give evaluation method for face aging simulation performance, and this method evaluated the human face aging performance indirectly by assessing whether the human face identity information can be remained and whether the face appearance characteristic feature belongs to the target age in the aging simulation result.Based on the product developing procedure for software engineering, this thesis provided detail analysis, design and realization of the system. Android cellphone client-side application of this system is already released to the cellphone app market such as Snap Pea. A web-based app of children face aging simulation has also been released online.