哈尔滨商业大学学报（自然科学版）2019，Vol.35Issue(1)：77-80,4.
基于傅里叶描述子的图像检索技术研究
Research on image retrieval technology based on Fourier descriptors
摘要
Abstract
The Fourier descriptor is one of the most important methods for analyzing and recognizing the shape of an object. The normalized Fourier descriptor has the characteristics of rotation, translation, invariance of scaling, and nsensitivity to the selection of the starting point. This paper used Fourier descriptors to process binary image contours to obtain eigenvectors. After the eigenvectors are obtained, the Euclidean distance and correlation coefficient methods were used respectively to characterize the eigenvectors of the query image and each image in the database. Vectors were compared for similarities. Experiments were conducted on the P100 database. Through analysis of the experimental results, it was shown that the Fourier descriptor has a great relationship with the number of Fourier descriptors and the choice of similarity criteria in terms of image retrieval accuracy.关键词
图像检索/形状特征/傅里叶描述子/相似度/傅里叶系数/MatlabKey words
image retrieval/shape feature/Fourier descriptor/similarity/Fourier coefficient/Matlab分类
信息技术与安全科学引用本文复制引用
高泽,孙华东,范智鹏,任聪,赵志杰..基于傅里叶描述子的图像检索技术研究[J].哈尔滨商业大学学报（自然科学版）,2019,35(1):77-80,4.基金项目
哈尔滨商业大学研究生创新项目(YJSCX2017-441HSD) （YJSCX2017-441HSD）
哈尔滨市科技局优秀学科带头人基金项目(2017RAXXJ055) （2017RAXXJ055）
