Audio-visual content-based violent scene characterization.pdf

Audio-visual content-based violent scene characterization.pdf

  1. 1、本文档共5页,可阅读全部内容。
  2. 2、原创力文档(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
  3. 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载
  4. 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
Audio-visual content-based violent scene characterization

I C I P - 9 8 AUDIO-VISUAL CONTENT-BASED VIOLENT SCENE CHARACTERIZATION Jeho Nam, Masoud Alghoniemy and Ahmed H. Tew k Department of Electrical and Computer Engineering University of Minnesota, Minneapolis, MN 55455, USA e-mail: fjnam,masoud,tew kg@ ABSTRACT We present a novel technique to characterize and index violent scenes in general TV drama and movies. Our goal is to identify violent signatures and localize vio- lent events within a movie to support \high-level video indexing. In particular, we exploit multiple \audio- visual signatures to create a perceptual relation for conceptually meaningful violent scene identi cation. Po- tential applications are automatic blocking of violence in movies watched by children, hiding violence using data hiding or information ltering and genre classi - cation of digital video database. 1. INTRODUCTION Currently, almost all reported video indexing schemes rely on low-level visual feature analysis. Regardless of conceptual meanings, those schemes allow searching for image/video containing visual characteristics similar to those de ned by a user, such as color, texture, motion, sketch and shape of video objects [1]. These approaches have achieved some success. However, they are limited to a narrow de nition of \content-based searching and matching. Higher level features must be used to ad- dress a general users desire to search for semantically meaningful information. For instance, one may desire to fast search and access time-localized speci c events of interest in a general lm such as \gun ghts scene in gangster or Western lms or \kissing scenes in ro- mantic movies. Unlike the domain-speci c applica- tions, these kinds of events can be commonly found in almost all lm genres. These semantic contents should be suitably described by composition of multiple low- level audio-visual features to link \high-level concepts [2]. In this paper, we introduce a novel high-level index- ing scheme to characterize and abstract violent c

文档评论(0)

l215322 + 关注
实名认证
内容提供者

该用户很懒,什么也没介绍

1亿VIP精品文档

相关文档