Study On Synthesis For 3D Facial Cartoon Based On Samples

A facial cartoon is a portrait which has characteristics of both identified of a real person's face and exaggeration effect. In the early 20th century, facial cartoons became very popular with many newspapers and magazines preferring to have facial cartoons. In recently years, with constant development of multimedia technology and enhancement of computer processing capability, 3D facial cartoons entered into the virtual world and interested large number of people. However, most of 3D facial cartoons can be made only by professional artists. So how to synthesize 3D facial cartoon utilize computer became an interesting topic for many researchers in the fields of computer graphic, matching learning.The dissertation researched two key problems in 3D facial cartoon synthesis and texture generation. The following are the main achievement in this dissertation: Firstly, a synthesis method for 3D facial cartoon models based on samples is introduced. With this method, a 3D facial cartoon can rapidly be made according to a single facial cartoon picture.Secondly, a method of texture synthesis for 3D facial cartoon based on single picture is presented. Applying the generated texture into 3D facial cartoon model can make the synthesized 3D model more vivid and recognizable.Thirdly, a framework of synthesizing 3D facial cartoon from a single facial photo, with the help of the method of synthesizing 2D facial cartoon, is presented. The implementation of the framework is completely automatic.Lastly, a database for facial samples is created involving in the real face, the 2D facial cartoon, the 3D facial cartoon. And these samples can easily be used by the researchers engaging the study for facial cartoon.