摘要
本文在Bayesianshapemodel(BSM)算法的基础上提出了一种新的二维图像特征提取算法。新算法为形状矩阵定义了一种新的表示方法,并提出了一种根据物体形状某些点(轮廓点)的坐标来估计剩余点(控制点)的坐标的新的方法。实验结果表明,在与传统算法计算量等同的条件下,该算法在精确度方面获得了很大的提高。
Based on Bayesian shape model（BSM）, We proposed a new two-dimensional picture feature extraction algorithm. The new algorithm give a new definition of the shape matrix, and proposed a new method to estimate coordination of some points （control points） of the shape from the others points （contour points）. The experiment results shows the new algorithm is much more accurate than traditional algorithm, while costs the same computing time.
出处
《微计算机信息》
北大核心
2006年第05S期189-192,共4页
Control & Automation
基金
国家自然科学基金(No.90207008)
