计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (28): 51-54.

• 学术探讨 • 上一篇    下一篇

基于Quaternion-Gabor滤波器的彩色纹理分割算法研究

马 瑜1,2,王晓辉2   

  1. 1.宁夏大学 物理电气信息学院,银川 750021
    2.上海交通大学 自动化系图像处理与模式识别研究所,上海 200240
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-10-01 发布日期:2007-10-01
  • 通讯作者: 马 瑜

Color texture segmentation algorithm using quaternion-gabor filters

MA Yu1,2,WANG Xiao-hui2   

  1. 1.School of Physics and Electrical Information Science,Ningxia University,Yinchuan 750021,China
    2.Institute of Image Processing & Pattern Recognition,Dept. of Automation,Shanghai Jiaotong University,Shanghai 200240,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-10-01 Published:2007-10-01
  • Contact: MA Yu

摘要: 提出一种的简单快速的多通道Gabor滤波技术对彩色纹理进行分割。首先,通过DRBFT和IDRBFT对彩色纹理的进行多通道Gabor滤波,再运用PCA对滤波得到的特征向量进行降维,对降维后的特征向量进行k-mean聚类,最后再对聚类后的区域用mean shift进行平滑,通过对平滑后的区域进行边缘检测就可以得到不同纹理的边界。最后给出几种分割算法的实验结果比较,表明该算法对于分割彩色纹理还是非常有效的。

关键词: DRBFT 简约双四元数, 多通道滤波器, k-means mean shift

Abstract: In this paper a simple and fast multichannel Gabor filtering of color texture is introduced for color texture segmentation.Firstly,using multichannel filtering filters the input color texture via DRBFT and IDRBFT.Then the PCA is used to reduce the feature space’s dimension and the k-means method is adopted to cluster the dimension reduced feature vector.At last the clustering region is smoothed by mean shift algorithm.From the smoothed region we can get the texture edge and superpose on the original image.Experimental results of the proposed approach and other conventional methods are compared and analyzed to demonstrate that the proposed approach is very effective in segmenting color texture.

Key words: DRBFT, RBs, multichannel filtering, k-means, mean shift