基于特征变形的人脸肖像漫画生成.pdfVIP

Vo l. 38 No.3 浙江工业大学学报 第 38 卷第 3 期 2010 年 6 月 JOURNAL OF ZHE]lANG UNIVERSITY OF TECHNOLOGY ]un. 2010 基于特征变形的人脸肖像漫回生成 冯晓噩1 ，潘翔2 (1.浙江工商大学信息学院，浙江杭州 310018; 2. 浙江工业大学软件学院.浙江杭州 310032) 摘要:漫画以极度夸张和变形的手法来表现人物，使其人脸特征更为突出.通过分析传统人脸肖像 漫画的作画规律，提出使用基于特征线段和鱼眼放大挤压算法的人脸肖像漫画生成算法.实现交互 拾取肖像人脸的关键特征点、整体的拉伸压缩变形方式以及局部的鱼眼变形方式，使得普通用户也 能生成个性化的人脸肖像漫画，大大减少了乎绘的工作量.所生成的漫画风格人脸肖像效果良好， 能够应用于电视动画节目、网络和报刊杂志等非真实感罔形学和数字娱乐等领域中. 关键词:人脸肖像;漫画;特征变形;鱼眼变形 文章编号:1006-4303(2010)03-0278-04 中团分类号:TP391 )c 献析、i只码 :A Geoeratiog algorithm of portrait cartooo based 00 feature deformatioo FENG Xiao-{ei 1 , PAN Xiang2 (1. College of Information Engineerin日， Zhejiang Gongshang University , Hangzhou 310018 , China; 2. College of Software. Zhejiang University of Technology , Hangzhou 310032 , China) Abstract: The portrait cartoon is used to express a personality using exaggerated and deformable method , which can make the {acial features more attractive. A generating algorithm of portrait cartoon based on eigenvector and fish-eye algorithm is presented by analyzing the sketching law. It can be realized through obtaining key feature points interactively , stretching or compressing the whol巳 portrait , and using fish-巳ye deformation algorithm partly. The general users also can generate personalized portrait cartoon using this method , which can reduce the hand-drawing workload. 丁he generated portrait cartoons are effective and vivid. They can be applied in non-photorealistic computer graphics and entertainment , suc