摘要
在视频的检索和审核过程中,为了让人们快速形成对视频大概的认识,需要进行视频信息挖掘。通过Python从视频中分离出音频,利用讯飞开放平台的REST API将语音转为文字,将文字及视频基本信息存入视频信息库,通过Flask构建视频信息管理系统。对文字进行词频统计生成词云,抽取视频帧生成截图,利用词云、截图及该视频Web URL对应的二维码等合成视频的封面海报。通过海报,可以快速对视频有大概的了解,并可以扫码查看视频详情。
In the process of video retrieval and auditing,in order to make people quickly form a general understanding of video,video information mining is needed.The audio is separated from the video through Python.The REST API is used to turn the voice into text,and the basic in formation of the text and video is stored in the video information database,and the video information management system is constructed through Flask.Word frequency statistics is used to generate word clouds.Video frames are extracted to generate screenshots.Cover posters of video are synthesized by word clouds,screenshots and two-dimensional codes corresponding to the video Web URL.Through posters,you can quickly get a general understanding of the video,and you can scan the two-dimensional code to see the details of the video.
作者
宋永生
SONG Yong-sheng(Jiangsu College of Engineering and Technology,Nantong 226007)
出处
《现代计算机》
2019年第31期76-79,共4页
Modern Computer
基金
江苏省现代教育技术研究项目(No.2018-R-64928)
南通市科技计划项目(No.JCZ18065)
江苏省教育信息化研究课题(No.20172168)
江苏省现代教育技术研究项目(No.2017-R-53720)
江苏工程职业技术学院科研计划项目(No.GYKY/2016/13)
