电子学报 ›› 2013, Vol. 41 ›› Issue (8): 1494-1499.DOI: 10.3969/j.issn.0372-2112.2013.08.007

• 学术论文 • 上一篇    下一篇

多表情人脸肖像的自动生成

宋红1, 黄小川2, 王树良1   

  1. 1. 北京理工大学软件学院, 北京 100081;
    2. 北京理工大学计算机学院, 北京 100081
  • 收稿日期:2012-10-25 修回日期:2013-04-26 出版日期:2013-08-25 发布日期:2013-08-25
  • 作者简介:宋 红 女,1977年生于陕西西安.博士,副教授.研究方向为计算机图形图像处理与模式识别. E-mail:anniesun@bit.edu.cn 黄小川 男,1988年生于湖南常德.硕士研究生,研究方向为计算机图形图像处理.
  • 基金资助:
    国家自然科学基金(No.60940024,No.61240010)

Automatic Generation of Portraits with Multiple Expressions

SONG Hong1, HUANG Xiao-chuan2, WANG Shu-liang1   

  1. 1. School of Software, Beijing Institute of Technology, Beijing 100081, China;
    2. School of Computer Science and Technology, Beijing Institute of Technology, Beijing 100081, China
  • Received:2012-10-25 Revised:2013-04-26 Online:2013-08-25 Published:2013-08-25

摘要: 肖像是一种能够抓住人物特征,又能隐藏细节,保留个人隐私的艺术表现形式,本文提出一种以中性表情人脸作为输入,自动生成不同表情的肖像算法.首先利用主动形状模型(Active Shape Model,ASM)提取人脸的关键特征点,然后根据统计学习获取的不同表情人脸的FAP(Facial Animation Parameter)规则,对中性人脸的特征点进行变形,生成表情三角网格,将中性人脸图像作为纹理映射到表情人脸网格,生成表情人脸图像,最后利用图像的梯度域信息和非真实感渲染技术,将带表情的人脸生成具有肖像风格的人脸图像.所生成的多表情人脸肖像效果良好,能够应用于网络和报刊杂志等非真实感图形学和数字娱乐等领域中.

关键词: 人脸肖像, 多表情, 特征形变

Abstract: Portrait is a kind of art form,which can catch the features of a person,hide the details and keep the personal privacy.In this paper,an algorithm for automatic generation of human portrait with multiple expressions is presented.The Active Shape Model is used to extract facial features,and the set of Facial Animation Parameters is used to synthesis the multiple expressions for a neutral face.And with the gradient domain of the image and the Non-Photorealistic Rendering,the portrait-liked face images with multiple expressions can be generated automatically.The generated portraits are effective and vivid.They can be applied in non-photorealistic computer graphics and entertainment,such as video,network,newspapers and magazines.

Key words: portrait, multiple expression, feature deformation

中图分类号: