Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (34): 10-12.

• 博士论坛 • Previous Articles     Next Articles

Color-based SIFT feature point detection and matching

GAO Jian,HUANG Xin-han,PENG Gang,WANG Min,WU Zu-yu   

  1. Department of Control Science and Engineering,Huazhong University of Science and Technology,Wuhan 430074,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-01 Published:2007-12-01
  • Contact: GAO Jian

基于彩色的SIFT特征点提取与匹配

高 健,黄心汉,彭 刚,王 敏,吴祖玉   

  1. 华中科技大学 控制科学与工程系,武汉 430074
  • 通讯作者: 高 健

Abstract: Because only the gray scale information is utilized,the SIFT method can’t differentiate the objects with similar shape but with different colors commendably.In order to solve such problem,presents a color-based SIFT feature point detecting method and analyzs the method’s performance in terms of different color models.The method detected interest points in the gray image scale space,and its eigenvectors are composed of the mean values of different color model components in each subregion and are computed based on the original color images.The experimental results have proved its validity.

Key words: feature point detection, feature matching, color model, SIFT method

摘要: 由于只是利用图像的灰度信息,SIFT算法不能很好地区分形状相似但颜色不同的物体。针对这一问题,提出了一种基于彩色的SIFT特征点提取算法,并着重分析了多种彩色模型对算法性能的影响。这种算法也是在图像的灰度尺度空间上检测特征点,但其特征向量由各描述子子区域的彩色模型分量的均值组成并在原始的彩色图像上进行计算。实验结果证明了该算法的有效性。

关键词: 特征点提取, 特征匹配, 彩色模型, SIFT方法