肖像漫画绘制机器人技术研究

Research on Caricaturing Technology of Auto-Portrait Robot

通过计算机进行人脸的肖像漫画绘制是机器视觉研究的前沿之一。为了从一张摄像头拍摄到的数码照片经过图像处理得到可以绘画的漫画线条,从而让机器人在纸上画出肖像漫画,首先必须由计算机自动生成人脸的线条画。因此,需要采用基于机器视觉的方法,结合人脸检测和人脸特征提取等技术,研究如何把人脸中的特征轮廓线提取出来,同时根据特征进行漫画化处理,最后转换为矢量线,作为机器人移动绘画的轨迹。本文讨论了现有的肖像生成方法和漫画化研究,评述了当前机器人漫画绘制系统。在这些基础上,提出了技术发展前景和未来发展目标。

Utilizing computer to draw a caricaturing portrait is a front field of machine vision. In a digital photo that be shot by CCD, for obtaining some lines that can exhibit image with caricature feature and draw these lines by robot, an appropriate image processing algorithm should be involved. The algorithm should includc facc recognition and extracting feature of face. Obviously, the caricature will be generated according to some features of face. Finally, after vectoring these represent lines, an Auto-Portrait robot will be controlled to draw cartoon portrait on a paper. This paper discuss proposed method for portrait and caricature, comment at existing system of Auto-Portrait robot. Further, future prospect of development and target are presented.