文章基本信息
- 标题：Content based Video Retrieval and Recommendation Using Speech and Text information
- 本地全文：下载
- 作者：Sarang S. Ajnadkar ; Swati Patil
- 期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
- 印刷版ISSN：2320-9798
- 电子版ISSN：2320-9801
- 出版年度：2016
- 卷号：4
- 期号：1
- 页码：744
- DOI：10.15680/IJIRCCE.2016.0401161
- 出版社：S&S Publications
- 摘要：Creatingvideorecordings ofevents such as lecturesor meetings is increasinglyless expensiveandeasy.Thus theVideodata isincreasinginagreatdealonWorldWideWeb(www)andsothustheneedofmoreefficientandcorrectly functioningmethodofvideo indexing,groupingandvideoretrieval inWWWorLargevideo archivesisnecessary.Thispaperpresentsaspeechandtextbasedvideo retrievalandVideosearch systemusingOptimalCharacterRecognition (OCR)andAutomatedSpeech Recognition (ASR).First,we convertthevideointokey-framesandextracttheAudioandTextusingOCRandASR.Followingstep istoproduceasummary presentingkeypointsofthevideo,bymakinguseoftextandaudioextractedfrom theVideo.Thissummarywillthenbeusedfor groupingandIndexingofvideos.Thisinturnwillimprovetheuser'saptitudetoquicklyreviewthismaterial.Thiswillmake usergo throughonlyinformationthattheyneeded.However,thetextinthevideomayvaryindimension,orientation,style,background, contrastandvariationsinrhythm,volumeofandnoiseinspeechandthedifferentiatingbetweenthekey-speechesandunnecessary other sounds usedduring therecording as well,makesdataextractionextremelychallenging
- 关键词：VideoIndexing; OCR;ASR;key-frames;data extraction
