Research And Implement Of Video Information Collection And Evaluation System

With the rapid development of the Internet technology in the world,the Internet Protocol TV(IPTV)is increasingly mature,and also becomes more and more popular.People can find thousands of videos which they are interested to in IPTV system,so the quality and quantity of videos in IPTV system directly affect the development of IPTV system.With the explosive growth of videos,it is a big problem that how to select high quality videos which users are interested to.In this paper,it through the information collection technology to get video information from video sites on the Internet,through the similarity matching technology to integrate the video information,and according to the different evaluation criteria and evaluation method,it can provides the data support for the selection of video resources in IPTV system.The main content of this paper is as follows:Firstly,this paper uses the breadth-first traversal algorithm,the multithreading processing method,and a series of sifting strategies which include the Internet addresses crawling range,Internet addresses crawling depth and Internet addresses regular expression,to get video information from video sites on the Internet.This method improves the efficiency of information collection.For information extraction technology,this paper uses a template-based approach for Information Extraction,which can turn the unstructured video information on the Internet to structured video information.Secondly,collected video information is integrated into standard data by using text similarity and string similarity.In the process of text similarity calculation,TF*IDF algorithm is used to extract text features,and the similarity is calculated by using the spatial vector model.Thirdly,a comprehensive video evaluation method is adopted to analyze the characteristics of the video,and to provide accurate data for the IPTV system.Finally,this paper designs and realizes the video information collection and evaluation system,uses the web crawler technology and string,text similarity technology to collect and integrate video information.It evaluates video information depending on evaluating indicators.