基于人物的相关网络视频挖掘PAGE\*MERGEFORMAT33摘要随着越来越多的视频文件出现在人们的生活中，如何有效处理它们成为亟需解决的问题。本论文以基于人脸的视频挖掘为研究方向，目的是通过分析大量视频文件，得到视频内部隐含的深层信息。具体的原理是：首先对视频集进行帧提取和预处理，得到较为清晰的含人脸图像；然后进行面部特征提取，得到眼睛、鼻子、嘴、眉的相对位置；最后对得到的特征值进行处理，根据加权的欧式距离计算出相似度，从而实现基于人脸的视频挖掘。本系统已可以初步完成基于人物的相关网络视频挖掘，然其准确度有待提高。关键词：视频挖掘，面部特征提取，几何特征识别AbstractAsmoreandmorevideofilesappearinpeople'slife,howtoeffectivelydealwiththemasaneedtosolvetheproblem.Inthispaper,basedonhumanfacevideominingastheresearchdirection,purposeisthroughtheanalysisofalargenumberofvideofiles,getvideohiddendeepinsideinformation.Specificprincipleis:thefirstframeofvideosetsareextractedandpreprocessing,getaclearfaceimage;Facialfeatureextraction,andthentotherelativepositionoftheeyes,nose,mouth,eyebrows;Finallytogetthecharacteristicvaluesofprocessing,accordingtotheweightedEuclideandistancetocalculatethesimilaritydegree,soastorealizevideominingbasedonhumanface.Thissystemhasbeendonecanpreliminarilybasedonthecharactersofnetworkvideomining,butneedstoimproveitsaccuracy.Keywords:videomining,facialfeatureextraction,geometricfeaturerecognition目录TOC\o"1-3"\h\z\uHYPERLINK\l"_Toc420571080"第一章绪论PAGEREF_Toc420571080\h5HYPERLINK\l"_Toc420571081"1.1背景PAGEREF_Toc420571081\h5HYPERLINK\l"_Toc420571082"1.2算法简介PAGEREF_Toc420571082\h6HYPERLINK\l"_Toc420571083"1.3论文主要内容PAGEREF_Toc420571083\h6HYPERLINK\l"_Toc420571084"1.4论文结构安排PAGEREF_Toc420571084\h7HYPERLINK\l"_Toc420571085"第二章系统的概要设计PAGEREF_Toc420571085\h7HYPERLINK\l"_Toc420571086"2.1系统的解决方案PAGEREF_Toc420571086\h7HYPERLINK\l"_Toc420571087"2.2系统的流程图PAGEREF_Toc420571087\h7HYPERLINK\l"_Toc420571088"2.3视频预处理方案PAGEREF_Toc420571088\h8HYPERLINK\l"_Toc420571089"2.4人脸检测方案PAGEREF_Toc420571089\h8HYPERLINK\l"_Toc420571090"2.5视频挖掘方案PAGEREF_Toc420571090\h9HYPERLINK\l"_Toc420571091"第三章视频预处理PAGEREF_Toc420571091\h9HYPERLINK\l"_Toc420571092"3.1视频预处理PAGEREF_Toc420571092\h9HYPERLINK\l"_Toc420571093"3.2帧