• 综合
  • 标题
  • 关键词
  • 摘要
  • 学者
  • 期刊-刊名
  • 期刊-ISSN
  • 会议名称
搜索

作者:

Huang, W. (Huang, W..) | Yin, B. (Yin, B..) | Jiang, C. (Jiang, C..) | Miao, J. (Miao, J..)

收录:

Scopus

摘要:

In this paper, we propose a new approach for eye features extraction based on 3D deformable eye template that can extract the precise eye features of both front faces and the faces with different orientation. Firstly, the hair is separated to estimate the pose of face by a method based on seven hair templates, from which the rough facial orientation can be obtained. Then, one of the three simple eye templates is chosen to locate the eyeball according to the facial orientation. Finally the 3D deformable eye template is used to extract the precise feature of the eyes. Some experiment results demonstrate the efficiency of the approach for both front faces and the faces with different orientation.

关键词:

3D deformable eye template; Energy function; Eye localization; Pose estimation

作者机构:

  • [ 1 ] [Huang, W.]College of Computer Science, Beijing Polytechnic University, Beijing 100022, China Digital Media Lab., Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, China
  • [ 2 ] [Yin, B.]College of Computer Science, Beijing Polytechnic University, Beijing 100022, China Digital Media Lab., Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, China
  • [ 3 ] [Jiang, C.]College of Computer Science, Beijing Polytechnic University, Beijing 100022, China Digital Media Lab., Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, China
  • [ 4 ] [Miao, J.]College of Computer Science, Beijing Polytechnic University, Beijing 100022, China Digital Media Lab., Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100080, China

通讯作者信息:

电子邮件地址:

查看成果更多字段

相关关键词:

相关文章:

来源 :

Proceedings of 2001 International Symposium on Intelligent Multimedia, Video and Speech Processing, ISIMP 2001

年份: 2001

页码: 340-343

语种: 英文

被引次数:

WoS核心集被引频次:

SCOPUS被引频次:

ESI高被引论文在榜: 0 展开所有

万方被引频次:

中文被引频次:

近30日浏览量: 1

在线人数/总访问数:533/4948352
地址:北京工业大学图书馆(北京市朝阳区平乐园100号 邮编:100124) 联系我们:010-67392185
版权所有:北京工业大学图书馆 站点建设与维护:北京爱琴海乐之技术有限公司