通信学报 ›› 2015, Vol. 36 ›› Issue (4): 179-184.doi: 10.11959/j.issn.1000-436x.2015089

• 学术通信 • 上一篇    下一篇

改进的图像局部特征区域描述方法

朱仁欢,高清维(),卢一相,孙冬   

  1. 安徽大学 电气工程与自动化学院,安徽 合肥 230601
  • 出版日期:2015-04-25 发布日期:2015-04-15
  • 基金资助:
    国家自然科学基金资助项目;国家自然科学青年基金资助项目;国家自然科学青年基金资助项目

Improved method for local image feature region description

Ren-huan ZHU,Qing-wei GAO(),Yi-xiang LU,Dong SUN   

  1. College of Electrical Engineering and Automation,Anhui University,Hefei 230601,China
  • Online:2015-04-25 Published:2015-04-15
  • Supported by:
    The National Natural Science Foundation of China

摘要:

针对描述子的性能与维数相矛盾的问题,提出了一种顽健的图像局部特征区域的描述方法。首先按照像素排序将局部特征区域分割为若干个子区域,然后利用基于阈值的分段局部描述子设计方法计算描述子,并采用纹理谱加权方法累加局部描述子得到子区域描述子,最后连接各部分子区域描述子得到最终的特征描述子。该方法综合了全局信息和局部信息,在保证描述子维数较小时对噪声具有一定的顽健性。实验结果表明该方法不仅对单调强度变化和旋转变化具有不变性,而且对其他几何和光学变换具有较好的顽健性。

关键词: 局部特征, 特征检测, 特征描述, 旋转不变性

Abstract:

A robust method for local image feature region description was proposed based on the problem of contradiction between performance and dimension of descriptor.First,the local feature region was divided into several subregions according to their intensity order.Then,the method of segmented local feature description based on threshold was used to compute the descriptor,and the method of the weighted texture spectrum was used to cumulate the local descriptor.At last,concatenate the descriptors of every subregion together to get the final descriptor of the interest region.This method combined overall information and local information together,and robust to noise when ensure the dimension of descriptor was small.The experimental results show that the proposed descriptor is not only invariant to monotonic intensity changes and image rotation but also robust to many other geometric and photometric transformations.

Key words: local feature, feature detection, feature description, rotation invariant

No Suggested Reading articles found!