摘要
提出了利用主曲线算法对人机交互中广泛使用的脱机手绘形状的识别。利用改进K段主曲线方法对学习样本点提取主干线,然后对被测试样本计算到主曲线的投影点的距离的均方差,识别是基于对目标主曲线的距离均方差最小。实验证明该算法也是比较有效的一种形状识别算法。
A hand-drawn shape recognition algorithm based on improved K-segment principal curve was presented.The algorithm was to generate a curve passing through the "middle" of the distribution.Recognition was based on the average distance from the sample points to the projected points on the generated principal curve.Experimental results show that the algorithm is feasible for hand-drawn shape recognition and exhibits good results.
出处
《计算机应用》
CSCD
北大核心
2009年第2期456-458,共3页
journal of Computer Applications
基金
国家863计划项目(2007AA12z156)
澳门科技发展基金资助项目(018/2005A)
北京师范大学珠海分校重点资助项目(Z06007)
关键词
主曲线
K段主曲线
形状识别
降维
principal curve
K-segment principal curve
shape recognition
dimension reduction
