人物肖像画及其卡通动画仿真系统的建立

Creating a System with Realism of Portrait and Cartoon

文章主要目的是利用一台普通的个人电脑和一张人脸的正面照片来自动生成具有写实风格或者夸张效果的人物肖像漫画及其动画。AAM模型是一种统计模型,可以很好地应用于人脸的特征点定位。但是AAM模型在信息的处理上忽略了颜色信息的作用,而肤色信息在人脸检测时往往起着非常大的作用。文中主要采用的方法是先利用肤色信息减小AAM搜索的搜索范围,然后基于这种改进的AAM模型,定位照片上的人脸特征,并以此为基础设计了一个人物肖像画生成系统。该算法对近百幅人脸的正面照片进行实验,实验结果令人满意。该方法能够快速生成特定人脸的肖像画并能较好地应用于卡通动画制作向导、多媒体智能娱乐、游戏和网络即时通讯软件等领域中。

The paper aims at creating an automated portrait generation system with realism or exaggerator and cartoon on the front image through a common personal computer. Active Appearance Model is a statistic model , it can be widely used in facial features extraction. But it still does not use all the information available, such as the skin color information, which has been widely used in face detection and preprocessing of digital images. Making full use of the chromatic information of skin will greatly reduce the scale of AAM searching, Based on this improved AAM ,we can accurately locate features of image and design an automated portrait generation system. This approach has been proved correct by experiments with hundreds of images and the results are almost the same as those expected. The experiment results show that the fast approach is very efficient in creating portrait generation and it will be widely used in Cartoon Wizard , multimedia amusement , game and network communicational software and so on.