检索结果-南通市图书馆

Modification ratio estimation for a category of adaptive steganography

在线全文

cnki会议

学校读者我要写书评

暂无评论

Science china(Information Sciences) 2010年第12期53卷 2472-2484页

作者： LUO XiangYang 1 , LIU FenLin 1 , YANG ChunFang 1 & lian shiguo 2 1 Zhengzhou Information Science and Technology Institute, Zhengzhou 450002, china 2 france telecom R&D beijing, beijing 100080, china 1. Zhengzhou Information Science and Technology Institute Zhengzhou 450002 China2. France Telecom R&D Beijing Beijing 100080 China

This paper investigates the detection of adaptive image steganography. Firstly, the sample pair analysis model of multiple least-significant bits (MLSB) replacement is analyzed, and the conditions under which the adaptive steganography can be detected via this model are given. For the category of adaptive steganography satisfying these conditions, a general quantitative steganalysis method is presented based on specific areas and sample pair analysis. Then, for a typical adaptive steganography, some concrete methods are proposed to select specific areas and trace sets of sample pairs, and estimate the stego modification ratios. Experimental results show that the proposed methods can estimate the stego modification ratio accurately. This verifies the validity of the proposed general quantitative steganalysis method.

关键词： adaptive steganography steganalysis MLSB (multiple least significant bits) sample pair analysis modification ratio

Novel encryption model for multimedia data

同方期刊数据库评论

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Journal of Systems Engineering and Electronics 2009年第5期20卷 1081-1085页

作者： Ye Dengpan~(1,2) & lian shiguo~3 1.The Key Laboratory of Aerospace Information Security and Trust Computing,Ministry of Education,Wuhan 430072,P.R.china 2.School of Computer Science,Wuhan Univ.,Wuhan 430072,P.R.china 3.france telecom R&D beijing,beijing 100080,P.R.china The Key Laboratory of Aerospace Information Security and Trust Computing Ministry of Education Wuhan 430072 P. R. China School of Computer Science Wuhan Univ. Wuhan 430072 P. R. China france telecom R&D beijing Beijing 100080 P. R. China

A novel encryption model is proposed. It combines encryption process with compression process, and realizes compression and encryption at the same time. The model＇s feasibility and security are analyzed in detail. An... 详细信息

关键词： multimedia encryption model compression feasibility security.

Using SVM for Text-Independent Speaker Verification

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Using SVM for Text-Independent Speaker Verification

第八届全国人机语音通讯学术会议

作者： Hao Yang, Yuan Dong, Xianyu Zhao, Xuefeng Zhang, Haila Wang beijing University of Posts and telecommunications france telecom research & development Center, beijing, china, 100083

Most of the state-of-the-art Text Independent Speaker Verification(SV) systems are based on Gaussian Mixture Models(GMM). Some literatures present that Support Vector Machines(SVM) could also give dramatic performance in Text-Independent SV. In this paper, both a GMM-UBM system and a SVM system are implemented on MIST 1999 SRE corpus. The SVM used in this paper is based on a Generalized Linear Discriminant Sequence Kernel. The DET curve of this SVM system seems equivalent to GMM-UBM baseline system.

关键词：

INSIGHT INTO THE ROLE OF PITCH INFORMATION IN TEXT-INDEPENDENT SPEAKER RECOGNITION

在线全文

cnki会议

学校读者我要写书评

暂无评论

INSIGHT INTO THE ROLE OF PITCH INFORMATION IN TEXT-INDEPENDE...

第八届全国人机语音通讯学术会议

作者： Xuefeng Zhang, Yuan Dong, Xianyu Zhao, Hao Yang, Haila Wang france telecom research & development Center, beijing, china, 100083 beijing University of Posts and telecommunications

This paper provides explanation and insight regarding the role of pitch information in text-independent speaker verification by experiments. The paper first brief a robust fundamental frequency extraction method. Then, a channel robust text-independent speaker verification based on GMM-UBM model which fuse the extracted pitch information and acoustic features is described in detail. Experiments are conducted using the NIST 1999 Speaker Recognition Evaluation corpus. Experimental results are presented show that the pitch features are more robust to handset variability and noise, but are not very robust to other factors like emotions. Therefore, how to effectively utilize the pitch information in speaker recognition is still a challenge for researchers.

关键词：

THE france telecom SYSTEM FOR THE TEXT-INDEPENDENT SPEAKER VERIFICATION

在线全文

cnki会议

学校读者我要写书评

暂无评论

THE FRANCE TELECOM SYSTEM FOR THE TEXT-INDEPENDENT SPEAKER V...

第八届全国人机语音通讯学术会议

作者： Yuan Dong, Xianyu Zhao, Xuefeng Zhang, Hao Yang, Haila Wang france telecom research & development Center, beijing, china, 100083 beijing University of Posts and telecommunications

This paper presents an overview of the architecture and algorithms implemented in france telecom’s text-independent speaker verification system. We describe individual components including the front-end processing, speaker modeling, test and score normalization steps. An effective combination of these components in the system has achieved a better performance in the operating region of NIST interest - the low false alarm operating region. Overall performance results of the system evaluation based on the NIST 1998 and NIST 1999 speaker recognition corpus are presented.

关键词：

SPORTS AUDIO CLASSIFICATION BASED ON MFCC AND GMM

在线全文

cnki会议

学校读者我要写书评

暂无评论

SPORTS AUDIO CLASSIFICATION BASED ON MFCC AND GMM

2009 2nd IEEE International Conference on Broadband Network & Multimedia Technology（IC-BNMT2009）

作者： Liu Jiqing 1 Dong Yuan 1 Huang Jun 1 Zhao Xianyu 2 Wang Haila 2 1 beijing University of Posts and telecommunications beijing 2 france telecom research & development Center beijing

Audio segmentation and classification can provide useful information for multimedia content *** this paper,we present a approach to segment and categorize the sports audio into speech,music and other environmental sounds for sports video classification and highlight *** investigate the performance of Mel Frequency Cepstral Coefficients （MFCC） in a Gaussian Mixture Model frame work,and compare it to traditional short-time energy and zero-crossing rate *** achieve a correct identification close to 90% on MFCC with its first and second derivatives.

关键词： audio classification MFCC GMM sports audio

Combination of the two dominant systems of SVM and GMM-UBM in text-independent speaker verification:SVM-based system using GMM super vector

在线全文

cnki会议

学校读者我要写书评

暂无评论

Combination of the two dominant systems of SVM and GMM-UBM i...

第九届全国人机语音通讯学术会议

作者： ZHAO Jian2, DONG Yuan1,2, ZHAO Xianyu1, YANG Hao2, LU liang2, WANG Haila1 (1. france telecom research & development Center, beijing 100080, china 2. beijing University of Posts and telecommunications, beijing 100876, china)

In the text-independent speaker verification field, Gaussian mixture model - universal background model strategy and support vector machines are the two most dominant methodologies. Just recently, a newly SVM-based speaker verification method using GMM super vector is proposed and arouses a hot current of research. In this paper, we are focused on the construction of this new speaker verification system and investigating the nuisance attribute projection and test normalization to enhance the performance further. Experiments are conducted on the core task of the 2006 NIST Speaker Recognition Evaluation corpus. The experimental results indicate that the SVM-based speaker verification sys-tem using GMM super vector, as a combination of the two dominant systems, can achieve better performance. With the use of nuisance attribute projection and test normalization, system performance can be significantly improved further.

关键词： support vector machines Gaussian mixture model super vector nuisance attribute projection test normaliza-tion speaker verification NIST 06 speaker recognition evaluation

在线全文

cnki会议

学校读者我要写书评

暂无评论

欧美人工智能监管模式及政策启示

中国行政管理 2024年第1期40卷 77-88页

作者：陈凤仙连雨璐王娜中国电信研究院战略发展研究所北京102209 北京大学政府管理学院北京100871

伴随以ChatGPT为代表的生成式人工智能的多场景应用,技术迭代升级背后的潜在异化风险带来全新的监管挑战。本文系统梳理当前主要经济体人工智能监管实践动态,分析典型监管模式路径及其现实困境,尝试在此基础上提出应对举措。研究发现,... 详细信息

伴随以ChatGPT为代表的生成式人工智能的多场景应用,技术迭代升级背后的潜在异化风险带来全新的监管挑战。本文系统梳理当前主要经济体人工智能监管实践动态,分析典型监管模式路径及其现实困境,尝试在此基础上提出应对举措。研究发现,伴随人工智能技术广泛应用,供应链安全、隐私保护、数字产权、数字伦理、数字鸿沟和算法偏见等深层问题逐渐凸显,给监管带来挑战,各国在人工智能监管的基础概念、价值理念及路径策略选择等多方面也存在诸多分歧,人工智能监管缺乏全球统一、协调一致的监管框架。当前主要经济体在人工智能监管方向上不断寻求创新和规制的平衡点,形成以欧盟和美英为代表的典型监管模式。未来人工智能监管将在“保持人的主体性”前提下,通过“限制”和“促进”双重手段,实现安全与发展之间的最佳平衡。我国应牢牢把握人工智能监管的规律和趋势,进一步统筹发展和安全,坚持问题导向和目标导向相结合,充分运用人工智能领域的新技术,适当借鉴美欧发达国家的经验做法,逐步构建起符合我国实际的人工智能监管模式。

关键词：人工智能监管困境监管创新

物联网安全防御体系分析及5G mMTC场景安全问题研究

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

通信与信息技术 2024年第2期 36-41页

作者：李博沈潋中国电信集团有限公司北京100032 国家电网有限公司大数据中心北京100052

随着智能家居、智慧医疗、车联网等技术的发展,物联网应用越发普及,在5G的牵引下,物联网有了新的发展动能,连接总数持续稳步增长,同时物联网安全攻击数量也空前增长,因此网络安全问题也引发了众多研究者的关注。首先介绍了物联网安全态... 详细信息

随着智能家居、智慧医疗、车联网等技术的发展,物联网应用越发普及,在5G的牵引下,物联网有了新的发展动能,连接总数持续稳步增长,同时物联网安全攻击数量也空前增长,因此网络安全问题也引发了众多研究者的关注。首先介绍了物联网安全态势,然后基于物联网的网络架构系统阐述了物联网纵深安全体系。最后针对国际电信联盟无线电通信部(International telecommunication Union-Radiocommunication Sector,ITU-R)中定义的大规模机器通信(Massive Machine Type Communication,mMTC)场景所面临的安全问题进行了阐述与分析。

关键词：物联网 5G 大规模机器通信