Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (16): 201-203.DOI: 10.3778/j.issn.1002-8331.2009.16.059

• 图形、图像、模式识别 • Previous Articles     Next Articles

Image feature extraction based on BEMD and gray level co-occurrence matrix

LONG Peng-fei,HE Liang,LV Hui,ZHANG Chun   

  1. College of Computer and Communication Engineering,Changsha University of Science and Technology,Changsha 410076,China
  • Received:2008-04-03 Revised:2008-06-16 Online:2009-06-01 Published:2009-06-01
  • Contact: LONG Peng-fei

基于BEMD和灰度共生矩阵的图像特征提取

龙鹏飞,贺 亮,吕 回,张 纯   

  1. 长沙理工大学 计算机与通信工程学院,长沙 410076
  • 通讯作者: 龙鹏飞

Abstract: A new method is proposed for image feature extraction with Bidimensional Empirical Mode Decomposition(BEMD) and Gray Level Co-occurrence Matrix.Namely the image is disassembled into the Intrinsic Mode Function(IMF) domain,then the feature of each IMF is extracted with Gray Level Co-occurrence Matrix method.To validate the availability of the arithmetic,authors extract the feature of the complex texture and the SAR image on pixel level.And achieving the image segmentation with KFCM,the segmentation result of images validates that the method is effective.

Key words: Bidimensional Empirical Mode Decomposition(BEMD), Intrinsic Mode Function(IMF), Gray Level Co-occurrence Matrix, complex texture, SAR image, Kernel-based Fuzzy C-Means(KFCM), image segmentation

摘要: 提出了一种新的图像特征提取方法,用二维经验模式分解将图像分解到固有模态函数(Intrinsic Mode Functions,IMF)域,即将图像分解成一系列的IMF和一个残差。并结合灰度共生矩阵对所提取到的各IMF图像和残差图像进行特征提取。为了验证算法的有效性,将其推广到像素级,对合成纹理和遥感图像进行了特征提取,并结合核模糊聚类(KFCM)算法对提取的特征向量做聚类分析,实现了图像的有效分割。

关键词: 二维经验模式分解, 固有模态函数, 灰度共生矩阵, 合成纹理, 遥感图像, 基于核的模糊C-均值聚类, 图像分割, ,