基于部件子图的面部图象合成系统

Facial Image Synthesis System Based on Components of Face Photograph

提出了一种基于部件子图的面部图象合成方法，借鉴塑膜重叠法的思想，把面部分为几个部件，从部件库中提取基本部件，经过处理之后合成面部图象。在计算机上应用数字图象处理和图形学的有关知识，设计实现成一种集部件获取、部件数据库管理和面部合成为一体的面部合成系统。对各类部件实现了按描述的分类和检索，针对不同部件的特点，采用不同的获取手段；在原始合成的基础上，实现优化合成。

This paper proposes a facial image synthesis method based on combination of face photograph components The combination is similar to the model of“plastic film overlapping',by which a face framework consists of several components,and each original component to be found from the component database will be pasted together after a reform processing A facial image synthesis system is designed and implemented by the proposed method as well as by methods of image processing and computer graphics The functions of the implemented system include component segmentation and registration,component database management,and synthesis operation The original components can be classified and retrieved by word description in the database Different component classes can be generated or obtained by different approaches adaptively The smoothness and optimization processing are involved in the synthesis