计算机应用 ›› 2011, Vol. 31 ›› Issue (07): 1818-1821.DOI: 10.3724/SP.J.1087.2011.01818

• 图形图像技术 • 上一篇    下一篇

改进的CenSurE特征和基于相加图像梯度的快速描述符

陈方1,蒋云良1,许允喜1,2   

  1. 1. 湖州师范学院 信息与工程学院,浙江 湖州 313000
    2. 浙江大学 信息与电子工程系,杭州 310027
  • 收稿日期:2010-12-13 修回日期:2011-01-29 发布日期:2011-07-01 出版日期:2011-07-01
  • 通讯作者: 陈方
  • 作者简介:陈方(1987-),女,浙江诸暨人,助教,硕士,主要研究方向:图像信号处理、计算机视觉;蒋云良(1967-),男,浙江海宁人,教授,博士,主要研究方向:人工智能、GIS、数据融合;许允喜(1978),男,江苏句容人,讲师,博士研究生,主要研究方向:图像处理、计算机视觉、人工智能。
  • 基金资助:

    国家自然科学基金资助项目;浙江省自然科学基金资助项目

Improved CenSurE detector and a new rapid descriptor based on gradient of summed image patch

Fang CHEN1,Yun-liang JIANG1,Yun-xi XU1,2   

  1. 1. School of Information and Engineering, Huzhou Teachers College, Huzhou Zhejiang 313000, China
    2. Department of Information Science and Electronic Engineering, Zhejiang University, Hangzhou Zhejiang 310027,China
  • Received:2010-12-13 Revised:2011-01-29 Online:2011-07-01 Published:2011-07-01
  • Contact: Fang CHEN

摘要: CenSurE局部特征计算效率非常高,但是CenSurE特征的尺度采样是线性的,滤波器响应信号很稀疏,检测的特征重复率不高。采用对数尺度采样得到改进的CenSurE特征,获得了更高的检测性能。同时,提出基于相加图像梯度的快速描述符,称为GSIP。图像区域匹配和物体识别评价实验结果显示,和目前性能最优的SURF描述符相比,GSIP描述符独特性更强,速度更快,计算时间不到SURF描述符的1/2。

关键词: 局部特征描述符, CenSurE, 图像匹配, 目标识别

Abstract: This paper proposed a new, real-time and robust local feature and descriptor, which can be applied to computer vision field with high demands in real-time. Since CenSurE has extremely efficient computation,it has got wide attention. Due to its linear scales, the filter response signal is very sparse and cannot acquire high repeatability. Therefore, this paper modified the detector using logarithmic scale sampling, and obtained better performance. The new rapid descriptor was based on gradient of the summed image patch, called GSIP. The GSIP descriptor has superior performance. An extensive experimental evaluation was performed to show that the GSIP descriptor increases the distinctiveness of local image descriptors for image region matching and object recognition compared with the state-of-the-art SURF descriptor. Furthermore, compared with SURF, GSIP achieves a two-fold speed increase.

Key words: local feature descriptor, CenSurE, image match, object recognition