基于有向图理论模型的网络新闻图像检索算法  

Network News Image Retrieval Algorithm Based on Directed Graph Theory Model

作　　者：战涛 姚璐 Zhan Tao;Yao Lu(Xi′an Mingde Institute of Technology,Xi′an 710124,China)

机构地区：[1]西安明德理工学院,西安710124

出　　处：《科技通报》2022年第8期35-40,共6页Bulletin of Science and Technology

摘　　要：针对当前网络新闻图像检索算法识别率较差、检索准确率和查全率较低、检索耗时长的问题,设计基于有向图理论模型的网络新闻图像检索算法。分析用户需求信息特征得到网络新闻图像特点,对图像进行分类及去噪处理,采用均衡滤波的方式处理图像的细节,根据网络新闻图像特征标识来定义有向图理论模型,通过有向图理论模型描述网络新闻图像数据融合的关联性,实现网络新闻图像的检索。实验结果表明,本文算法对网络新闻图像的识别率为90%,检索准确率和查全率始终在90%以上,检索耗时最高为2.36 s。Aiming at the problems of poor recognition rate of current network news image retrieval algorithms,low retrieval accuracy and recall rate,and long retrieval time,a network news image retrieval algorithm based on the directed graph theory model is designed.Analyze the information characteristics of user needs to obtain the characteristics of the network news image,classify and denoise the image,process the details of the image by means of equalization filtering,define the directed graph theoretical model according to the network news image feature identification,and use the directed graph theoretical model Describe the relevance of network news image data fusion,and realize the retrieval of network news images.The experimental results show that the algorithm in this paper has a recognition rate of 90% for network news images,the retrieval accuracy and recall rate are always above 90%,and the retrieval time is up to 2.36 s.

关 键 词：有向图理论模型 网络新闻 图像检索 准确率 查全率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]