基于边界曲线演化模型的生长骨架算法

被引:8
作者
刘文予
白翔
朱光喜
机构
[1] 华中科技大学电子与信息工程系
[2] 华中科技大学电子与信息工程系 武汉
关键词
骨架; 曲线演化; 边界; 多尺度; 视觉重要成分;
D O I
10.16383/j.aas.2006.02.013
中图分类号
TP391.41 [];
学科分类号
080203 ;
摘要
基于距离变换的骨架算法往往不能直接用于骨架识别,且骨架的连通性难以保证.本文提出一种新型的骨架算法,由一个初始骨架点开始逐点生长出各骨架分支,同时在骨架生长过程中用离散曲线演化模型消除造成信息冗余的骨架枝,保留视觉上重要的骨架枝,实现了骨架的多尺度控制.实验证明本算法复杂度低,得到的骨架连通性得到保证,能较好地表示图形中视觉重要成分,符合人类视觉习惯,可直接用于图形识别和形状度量.
引用
收藏
页码:255 / 262
页数:8
相关论文
共 10 条
[1]  
Euclidean skeletons. Gregorie Malandain,Sara Fernandez-Vidal. Image and Vision Computing . 1998
[2]  
Distance transformations in digital images. Borgefors G. Computer Vision Graphics and Image Processing . 1986
[3]  
Hierarchic Voronoi skeletons. Ogniewicz R L,KSubler O. Pattern Recognition . 1995
[4]  
Extraction of the Euclidean skeleton based on a connectivity criterion. Choi W P,Lam K M,Siu W C. Pattern Recognition . 2003
[5]  
Shape similarity measure based on correspondence of visual parts. Latecki L J,Lak’amper R. IEEE Transactions on Pattern Analysis and Machine Intelligence . 2000
[6]  
Ligature instabilities in the perceptual organization of shape. August J,Siddiqi K,Zucker S. Computer Vision and Image Understanding . 1999
[7]  
Thompson, Renato Perucchio.A topology-preserving parallel 3D thinning algorithm for extracting the curve skeleton. Xie W J,Robert P. Pattern Recognition . 2003
[8]  
Biological shape and visual science (part I). Blum H. Journal of Theoretical Biology . 1973
[9]  
Simulating the grassfire transaction form using an active contour model. Leymarie F,Levine M. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1992
[10]  
A dynamic approach to skeletonization. Che W J,Yang X N,Wang G Z. Journal of Software . 2003