Based On Rapid Detection Of Thick Granularity Video Fingerprint System Design And Implementation

Along with the development and interation of digital technology，videoinformation takes more and more place in people’s daily life.Every day a large numberof digital videoes are producted、stored and spreaded.The video information accupiesa significant proportion in data stream of internet and this trend is becominggradually larger.But seriouse problems are coming with the video information spreads inlarge-scale.For example,the copyright of video and illegal videoes have threatensecurity of society and economic interest.So how to monitor video informationeffectivly has become inevitable in today’s society.At the base of using video fingerprint as the main mean of video detectiontechnology,I developed a system mathes fasting vedeo fingerprint detection for thepurpose of real time monitoring in this subject.First of all I give an overview of videofingerprint technologies that currently exisited and analysis representative ones ofthem.After the analysis of the development of video fingerprint technology,I turn to thestudy of video fingerprint extraction algorithm.With given an overview of theadvantages and disadvantages of the representative video fingerprint extractionalgorithms and directional guidance of video fingerprint extraction for fingerprintdetection system.At last I give an video fingerprint extraciton on the base of the timeslice of key frames.In the meantime a detailed account is shown in this subject.In view of designing the mode of fingerprint before studying the fingerprint matchalgorithm in chapter V,I present an fingerprint mode with tree structure after extractingfingerprint of the video successfully.And a necessary prerequisite is prepared forfingprint matching fastly.The fingerprint match algorithm based on generous andcareful granularity are raised after chosing tree structure as fingerprint’s model.In orderto reduce fingerprint matching scale,first of all,get bucket mapping number byhuckethash method and then compute Hamming distance by conducting FFM methodand then we can contract fingerprint gradually. I design and realize the database with a view of information of video and soon.After giving sufficient analysis for service logic of video fingerprint matchingsystem,video information is stored by presenting a blue print conbining file system anddatabase.And I have stored and realized database for video information saving andthe blue print can save complex video information well.I complete designment and realizability of the system after analysised chaptersbefore and presented algorithm.Scalability and decouple are fully taken into accountwhen building the the framework of the system,and I combine Struts2,Hibernate andSpring which are all the current mainstream framework,then based on which I use thejava language implement the whole system. System can be divided into:the actionlayer,service layer,the model layer,Dao layer and the corresponding toolkit on the basisof hierarchical structure,based on which can make the development be moreindependent.The visual effects shots are obtained after completing the development ofthe system and executing the fingerprint detection with the actual video,and all basedon the actual system which is developped by myself.Chapter VII gives a full analysis for the deficiencies and perfection of thesystem,which can provide direction guide for follow-up work.