网络视频数据获取与后处理技术综述  

A review on the network video data acquisition and post processing

作　　者：张昆 张峰 张德 王惠峰 王子玮 白立飞 熊荔 ZHANG Kun;ZHANG Feng;ZHANG De;WANG Huifeng;WANG Ziwei;BAI Lifei;XIONG Li(Information Science Academy,CETC,Beijing 100086,China)

机构地区：[1]中国电子科技集团公司信息科学研究院

出　　处：《电视技术》2019年第6期24-30,44,共8页Video Engineering

摘　　要：视频在内容表达上具有直观性和丰富性的优点,其作为信息和知识的载体更容易被人们接受。近年来,随着互联网尤其是移动互联网技术的成熟,网络视频行业呈现出繁荣发展的态势。另外,以深度学习为代表的人工智能技术与视频数据相结合,推动了视频内容理解领域的核心技术突破,促成多项相关技术落地。可以预见,未来的信息检索系统将承载视频数据,而网络视频数据获取与后处理技术,作为视频信息检索系统数据来源的提供者,将发挥巨大的作用。本文综述了网络视频数据获取技术的现状,并对该方向的未来发展作了展望。Video has the advantage of being intuitive and rich in content expression,and it is more easily accepted as a carrier of information and knowledge.In recent years,with the maturity of the internet,especially mobile internet,the online video industry has shown a prosperous development.In addition,the combination of artificial intelligence technology and video data represented by deep learning has promoted the core technological breakthroughs in the field of video content understanding,and has led to the emergence of a number of related technologies.It is foreseeable that future information retrieval systems will carry video data,and network video data acquisition technology,as a provider of data sources for video information retrieval systems,will play a huge role.This paper reviews the current status of network video data acquisition technology and looks forward to the future development of this direction.

关 键 词：信息检索 网络视频 爬虫 镜头边界检测 关键帧 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]