计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (35): 190-193.

• 图形、图像、模式识别 • 上一篇    下一篇

基于局部方向纹理谱描述符的图像检索

毋小省1,孙君顶1,2   

  1. 1.河南理工大学 计算机科学与技术学院,河南 焦作 454000
    2.河南理工大学 控制工程重点学科开放实验室,河南 焦作 454000
  • 出版日期:2012-12-11 发布日期:2012-12-21

Image retrieval based on local direction texture spectrum descriptors

WU Xiaosheng1, SUN Junding1,2   

  1. 1.School of Computer Science and Technology, Henan Polytechnic University, Jiaozuo, Henan 454000, China
    2.Provincial Open Lab for Control Engineering Key Disciplines, Henan Polytechnic University, Jiaozuo, Henan 454000, China
  • Online:2012-12-11 Published:2012-12-21

摘要: 提出了一种改进的局部方向纹理谱描述符。根据图像局部邻域中心像素及与其相对像素间的灰度关系及灰度差异关系来定义局部区域纹理方向,在此基础上结合中心对称局部二值模式的定义给出了新的局部方向纹理谱的定义。新描述符在不增加特征维数的基础上,更好地融入了局部纹理信息。采用不同图像库及不同评价准则进行图像检索对比实验,结果表明,该方法取得了较好的检索效果。

关键词: 纹理谱, 局部二值模式, 中心对称局部二值模式, 方向局部二值模式

Abstract: To improve the performance of region texture spectrum descriptors Center-Symmetric Local Binary Pattern(CS-LBP) and Direction Local Binary Pattern(D-LBP), a new region operator, variance and Direction Local Binary Pattern(vD-LBP) is proposed in the paper. Based on the further definition of texture direction in a local region, both the gray-value changes and the gray-value variance changes of the central pixel and its center-symmetric pixels are considered together in the new descriptor. The descriptors mentioned in the paper are tested and evaluated respectively on three commonly used image databases and the experimental results show that vD-LBP is an very effective method for texture description.

Key words: texture spectrum, local binary pattern, center-symmetric local binary pattern, direction local binary pattern