[1]高建坡,王煜坚,杨浩,等.基于均值移动和椭圆拟合的人脸跟踪算法[J].东南大学学报(自然科学版),2006,36(6):897-902.[doi:10.3969/j.issn.1001-0505.2006.06.005]
 Gao Jianpo,Wang Yujian,Yang Hao,et al.Face tracking algorithm based on mean shift and ellipse fitting[J].Journal of Southeast University (Natural Science Edition),2006,36(6):897-902.[doi:10.3969/j.issn.1001-0505.2006.06.005]
点击复制

基于均值移动和椭圆拟合的人脸跟踪算法()
分享到:

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
36
期数:
2006年第6期
页码:
897-902
栏目:
计算机科学与工程
出版日期:
2006-11-20

文章信息/Info

Title:
Face tracking algorithm based on mean shift and ellipse fitting
作者:
高建坡 王煜坚 杨浩 吴镇扬
东南大学信息科学与工程学院, 南京 210096
Author(s):
Gao Jianpo Wang Yujian Yang Hao Wu Zhenyang
School of Information Science and Engineering, Southeast University, Nanjing 210096, China
关键词:
人脸跟踪 均值移动 椭圆拟合 形状模型
Keywords:
face tracking mean shift ellipse fitting shape model
分类号:
TP391
DOI:
10.3969/j.issn.1001-0505.2006.06.005
摘要:
为了解决均值移动跟踪算法对目标的尺度变化自适应能力差的缺点,针对人脸跟踪具体问题,提出了一种基于均值移动和椭圆拟合的人脸跟踪算法.根据当选择较大核窗进行均值移动跟踪时一般可以准确定位目标的事实,以及人脸形状椭圆近似性的特征,利用一个较大核窗的均值移动跟踪器对人脸目标进行粗略定位,在此基础上再用一种高效鲁棒的直接最小二乘椭圆拟合方法来自动调整人脸尺度的大小.实验表明,该改进算法能有效地解决均值移动人脸跟踪中的目标尺度自适应调整问题,其跟踪效果明显优于原均值移动目标跟踪算法.
Abstract:
To overcome the shortcoming that mean shift tracker can’t adjust scale with object during tracking process, a face tracking algorithm based on mean shift and ellipse fitting is proposed. In accordance with the fact that mean shift tracker can always give the right target location when a larger kernel bandwidth is chosen and the facial shape can be appropriate to an ellipse, firstly, the coarse location of face target is attained via a mean shift tracker with bigger kernel bandwidth, then a robust and efficient direct least square ellipse fitting method is used to adjust the facial scale. The experimental results demonstrate the efficiency of this algorithm. Its performance has been proven superior to the current mean shift tracking algorithm.

参考文献/References:

[1] Greenspan H,Goldberger J,Eshet I.Mixture model for face-color modeling and segmentation [J].Pattern Recognition Letters,2001,22(14):1525-1536.
[2] Bojic N,Pang K K.Adaptive skin segmentation for head and shoulder video sequences [J]. Visual Communications and Image Processing,SPIE,2000,4067:704-711.
[3] Nummiaro K,Koller-Meier E,van Gool L.An adaptive color-based particle filter [J]. Image and Vision Computing,2003,21(1):99-110.
[4] Perez P,Hue C,Vermaak J,et al.Color-based probabilistic tracking [C] //European Conference on Computer Vision LNCS 2550.Berlin:Spring-verlag,2002:661-675.
[5] Comaniciu D,Ramesh V,Meer P.Kernel-based object tracking [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003,25(5):564-577.
[6] Comaniciu D,Meer P.Mean shift:a robust approach toward feature space analysis [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619.
[7] Chen Yizong.Mean shift,mode seeking,and clustering [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1995,17(8):790-799.
[8] Fitzgibbon A,Pilu M,Fisher R B.Direct least square fitting of ellipse [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1999,21(5):476-480.

备注/Memo

备注/Memo:
基金项目: 国家重点基础研究发展计划(973计划)资助项目(2002CB312102).
作者简介: 高建坡(1975—),男,博士生; 吴镇扬(联系人),男,教授,博士生导师, zhenyang@seu.edu.cn.
更新日期/Last Update: 2006-11-20