摘要
根据视觉信息处理的多通道滤波理论,研究了基于多通道Gabor滤波的纹理分割方法.该方法以覆盖全部频率空间、8个滤波方向的Gabor滤波器组构造滤波通道,通过随机样本方差聚类分析,结合最小距离判别法,实现了图像纹理的准确分割.
According to the theories of multi-channel filtering in visual information process, a method based on multi-channel Gabor filters for texture segmentation was studied. In this method, filtering channels are constructed by Gabor filters witch cover the whole frequency domain and include eight different directions. By using random sample square error clustering algorithm and minimum distance discrimination, accurate segmentation of image textures was realized.
出处
《湖南文理学院学报（自然科学版）》
CAS
2006年第3期75-77,共3页
Journal of Hunan University of Arts and Science(Science and Technology)
关键词
多通道GABOR滤波
方差聚类算法
最小距离判别
纹理分割
Multi-channel Gabor filter
Square error clustering algorithm
Minimum distance discrimination
Texture segmentation
