CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleMultimedia Message Monitoring System-Design and Implementation
Thesis Advisor胡卫明 ; 谭铁牛
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword多媒体短消息 监控系统 敏感图像识别 特殊标志检测 Multimedia Message Monitoring System Objectionable Image Recognition Symbol Detection
Abstract随着中国移动2002年10月开通彩信业务以来,短消息业务告别了文本时代,进入了多媒体的新纪元。多媒体短消息业务促进了人与人之间的交流,丰富了人们的沟通方式,但 也不可避免地带来一些负面效应。敏感图像和反动标志等不健康的媒体内容在短消息中的 传播成为可能。为了避免这些不健康内容的泛滥,辅助网络安全人员对这些内容进行分 类,我们迫切需要这样一个监控系统,能够根据图像内容自动检测其中的敏感信息。 本文针对短消息监控系统的发计与开发,从算法层面到系统层面都进行了一些有益的 尝试,算法涉及到了许多图像过滤的基本问题,包括统计肤色模型、敏感图像识别与特殊 标志检测等等,系统开发涉及整个系统框架的设计、每个模块的具体设计以及模块的整合 等等。本文的主要工作和贡献有: (1)提出了多媒体短消息监控系统的一个基本框架,能够整合到多媒体短消息业务中心 的现有运行体系中,以增值业务的方式为业务中心提供图像过滤服务。 (2)比较了直方图模型和混合高斯模型用于肤色建模的效果,研究了不同的高斯核数目 对于肤色建模的影响。 (3)研究了基于肤色连通区域和基于躯干轮廓两种敏感图像识别算法,比较了它们的性 能 (4)研究了基于相关图的特殊标志检测算法 (5)实现了监控系统的各个子模块,并搭建了3个演示系统。 总的说来,本文在图像过滤算法作了一些有益的探索,并且实现了一个实用的监控系统原型。
Other AbstractSince October 2002, when China Mobile first started Multimedia Message Service (MMS), short message service has entered a new era. The introduction of MMS has provided abundant means of communication and better information exchange among its clients, while it has also induced some negative effects. Propagation of unhealthy media contents chanracterized by adult images and sensitive symbols was made possible by MMS. To avoid the proliferation of such contents and assist security engineers in screening them, we cai1 for monitoring systems which can automatically tell the existence of unhealthy contents in short messages based on image content analysis. Focusing on the design and implementation of such a message mornitoring system, this paper studied many topics from algorithm to system design, including algorithms for skin detection, adult image recognition and symbol detection, as well as the design of system framework, implemenation of system modules and integration of different system modules into the framework. The contribution of this paper includes, (1) proposed the framework of multimedia message mornitoring system, which can be integrated into the current architecture of Multimedia Message Service Center(MMSC) and provides value added service on image filtering for MMSC. (2) compared the effect of the histogram model and the Gaussian Mixture Model on skin detection, and studied the influence of different component number for GMM on detection results. (3) studied two algorithms for adult image recognition based on connected skin color components and extracted body contours respectively, and compared their performance on our large image database (4) studied correlogram based symbol detection algorithm (5) implemented each system module and built three demo systems Generally speaking, the author has made some positive attempt on image analysis algorithms in this paper and implemented a working system.
Other Identifier770
Document Type学位论文
Recommended Citation
GB/T 7714
傅周宇. 多媒体短消息监控系统—设计与实现[D]. 中国科学院自动化研究所. 中国科学院研究生院,2004.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[傅周宇]'s Articles
Baidu academic
Similar articles in Baidu academic
[傅周宇]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[傅周宇]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.