摘要
为解决单纯描述形状轮廓时只能描述简单形状的缺点,提出了融合最小和统计链码与熵矩阵奇异值特征的形状分类和检索方法。从结构和统计的角度综合了基于形状轮廓和区域的形状描述,同时又兼具了形状的空间分布特征。对形状区域特征扩展空间分布特征,使形状描述具有鲁棒性,形状熵矩阵奇异值具有对噪声稳定的特点,弥补了链码描述形状特征易受噪声影响的局限性。设计并实现了基于MPEG-7标准形状数据库的形状检索平台,实验结果表明,该算法在形状进行检索和匹配时具有较好的旋转、平移和尺度不变性。
To address the shortcomings of the mere description of the shape of the contour can only describe a simple shape,a fusion of the smallest and statistical chain code and the entropy matrix singular value feature shape classification and retrieval methods, a combination of structure and statistical point of view based on the shape of the contour and regionalshape description,the same time, both the shape of the spatial distribution characteristics. Regional characteristics of the shape extends the spatial distribu- tion characteristics of shape description with the robustness, the shape of the entropy matrix singular value characteristics of the noise stability to make up the chain code to describe the shape characteristics susceptible to noise limitations. Designed and implemented a shape retrieval based on MPEG-7 shape da- tabase platform,and experimental results show that the algorithm has a better rotation, translation and scale invariance in shape retrieval and matching.
出处
《吉林大学学报(信息科学版)》
CAS
2012年第3期234-239,共6页
Journal of Jilin University(Information Science Edition)
基金
吉林省科技引导计划应用基础研究基金资助项目(20090505)
吉林大学自然科学基础研究基金资助项目(450060445135)
关键词
形状检索
链码
熵矩阵
奇异值
shape retrieval
chain code
entropy matrix
singular value
作者简介
康建玲(1987-),女,陕西武功人,吉林大学硕士研究生,主要从事图像理解研究,(Tel)86—18943929083(E—mail)kangjian—ling0926@sina.com;
郭树旭(1959-),男,长春人,吉林大学教授,博士生导师,主要从事信号检测与信号处理研究,(Tel)86-13089111921(E-mail)guosx@jlu.edu.cn。