摘要
Hausdorff距离是一种有效的检测两个点集相似性的算法.为了有效的利用边缘信息对光照不敏感的特点,本文在原有SWHD(空间加权豪氏距离)算法的基础上,提出了一种利用局部距离映射图改进的hausdorff距离人脸识别算法.实验结果表明,此方法不仅可得到较好的识别率,而且能有效侧重于强调脸部的主要特征区域.
To compare two point sets, Hausdorff distance is an efficient measure that can determine the degree of their resemblance. In order to use the edge information effectively, In this paper, a new modified SWHD is proposed, which refers to 'local distance map'. Experimental results show that the new modified distance can achieve higher recognition rate, but more emphasize the crucial region of face effectively.
出处
《河北工业大学学报》
CAS
2004年第6期30-34,共5页
Journal of Hebei University of Technology
基金
河北省教育厅博士基金项目(B2003202)
