基于SIFT特征的多媒体图像信息检索方法研究  

Research on Information Retrieval of Multimedia Images Based on SIFT Features

作　　者：孙英 刘忠利 Sun Ying;Liu Zhongli(School of Information Engineering,Jincheng College of Nanjing University of Aeronautics and Astronautics,Nanjing Jiangsu 210000,China;Basic Teaching Department,Jincheng College of Nanjing University of Aeronautics and Astronautics,Nanjing Jiangsu 210000,China)

机构地区：[1]南京航空航天大学金城学院信息工程学院,江苏南京210000 [2]南京航空航天大学金城学院基础教学部,江苏南京210000

出　　处：《信息与电脑》2020年第7期179-180,共2页Information & Computer

摘　　要：针对传统方法查全率较低的问题,笔者提出基于SIFT特征的多媒体图像信息检索方法。该方法采用数字信息描述方法描述多媒体图像信息,提取图像关键特征;将数据进行XML格式转换,并以XML数据库方式存储数据;引入SIFT特征检索算法,找出多媒体图像与XML数据库中欧式距离最近的前两个点,并经过比例阈值对比,输出检索结果,实现研究。经实验证明,提出的方法平均查全率远高于传统方法。Aiming at the problem of low recall rate of traditional methods,a research on multimedia image information retrieval methods based on SIFT features is proposed.Digital information description method is used to describe multimedia image information,and key features of the image are extracted.The data is converted into XML format and stored in XML database.The SIFT feature retrieval algorithm is introduced to find the top two European-style distances between the multimedia image and the XML database.Points,and after the ratio threshold comparison,the retrieval results are output for research.The experiment proves that the average recall of the proposed method is much higher than the traditional method.

关 键 词：SIFT特征 多媒体图像信息 检索方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]