首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于相对高度的曲线特征提取算法
引用本文:黄国华.基于相对高度的曲线特征提取算法[J].邵阳学院学报(自然科学版),2011,8(1):29-33.
作者姓名:黄国华
作者单位:邵阳学院理学与信息科学系,湖南邵阳,422000
基金项目:湖南省教育厅资助项目基于支持向量机的基因识别研究(09C888.)
摘    要:特征是物体匹配和识别的依据.为提高图形图像的识别率,对曲线的特征提取进行了探索,提出了一种基于曲线相对高度的特征提取算法.算法可大致描述如下:利用曲线的高分割曲线,再分别计算各段曲线的相对高度,不断重复前面两个操作,直至相对高度满足小于某个给定的常数.对于每一个相对高度,按照分割曲线的层次存储在二叉树相应的节点.实验结果表明,该算法能够满足曲线在平移、旋转、缩放和对称变换后,其特征不变,而且具有良好的灵敏度.该算法可用于手写汉字和人脸轮廓的识别.

关 键 词:特征提取  曲线  相对高度  二叉树  相似性度量

A New Algorithm of Extracting Feature from Curves Based on the Relative Height
HUANG Guo-hua.A New Algorithm of Extracting Feature from Curves Based on the Relative Height[J].Journal of Shaoyang University:Science and Technology,2011,8(1):29-33.
Authors:HUANG Guo-hua
Institution:HUANG Guo-hua(Department of science and information-science,Shaoyang University,Shaoyang Hunan 422000)
Abstract:Feature plays a decisive role in object matching and recognition.For the purpose of improving correctness of shape recognition,the question how to refine better feature from curves was investigated and a new algorithm based on relative heights of curves was introduced.The method was outlined as follows.First,Height partitions a curve into two parts.Secondly,relative heights are computed respectively in two small curves.Thirdly,if a relative height is less than a given constant,the subdivisions of the corres...
Keywords:feature extraction  curves  relative height  binary tree  similarity measure  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号