| 基于特征加权的自然纹理FCM聚类分割算法 |
| |
| 引用本文： | 朱映辉,江玉珍,欧阳春娟.基于特征加权的自然纹理FCM聚类分割算法[J].微机发展,2010(1):108-110,115. |
| |
| 作者姓名： | 朱映辉 江玉珍 欧阳春娟 |
| |
| 作者单位： | 韩山师范学院数学与信息技术系;井冈山大学信息科学与传媒学院; |
| |
| 基金项目： | 江西省教育厅2008科技项目(GJJ08417) |
| |
| 摘 要： | 为了实现自然材质的纹理分割,根据自然纹理的弱规则性特点,提出一种由图像灰度值、灰度分布统计及图像纹理能量统计作为纹理表征的特征参数,并组成三维特征矢量以实现自然纹理分割的算法。考虑到样本不同特征值对分类的不同影响,算法中引入了特征加权的FCM模糊聚类方法以提高各特征参数在聚类约束力上的可控制性,从而实现纹理图像的更有效分割效果。实验证明,该方法简单高效、可控性强,对各种自然纹理图像具有较好的纹理分割效果。
|
| 关 键 词： | 自然纹理 纹理分割 特征加权 FCM 纹理能量 |
Natural Texture Segmentation Algorithm Based on Weighted Features FCM Clustering |
| |
| ZHU Ying-hui,JIANG Yu-zhen,OUYANG Chun-juan.Natural Texture Segmentation Algorithm Based on Weighted Features FCM Clustering[J].Microcomputer Development,2010(1):108-110,115. |
| |
| Authors: | ZHU Ying-hui JIANG Yu-zhen OUYANG Chun-juan |
| |
| Affiliation: | ZHU Ying-hui1,JIANG Yu-zhen1,OUYANG Chun-juan2 (1.College of Mathematics , Information Technology,Hanshan Normal University,Chaozhou 521041,China,2.Information Science , Medium College,Jinggangshan University,Ji\'an 343009,China) |
| |
| Abstract: | In order to carry out the texture segmentation of natural material,aiming at the weak regularity of natural texture,a natural texture segmentation algorithm combined with image pixel value,statistical analysis of grey distribution and texture energy is proposed.Considering the different function of the different sample features in classifying,the FCM clustering with weighted features is adopted in the algorithm to improve the clustering restraint controllability of the sample features,and get the better tex... |
| |
| Keywords: | natural texture texture segmentation weighted features FCM texture energy |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
