检索结果-内蒙古大学图书馆

International Conference on Information, communications and Signal Processing

作者： Yanhua Long Wu Guo Bin Ma Eng Siong Chng Donglai Zhu Lirong Dai Haizhou Li IFly Speech Laboratory EEIS University of Science and Technology China MOE-Microsoft Key Laboratory of Multimedia Computing and Communication USTC China Institute for Infocomm Research Singapore Nanyang Technological University Singapore

ISBN: (纸本)9781424446568

In this paper, we propose a subspace construction and selection strategy (SUBS) for speaker recognition with limited training and testing speech data. Based on the individual Gaussian distributions of Gaussian mixture model (GMM), each speaker's characteristic subspace is constructed by training an SVM using the corresponding Gaussian mean vectors from the GMms of both enrollment and imposter speakers. A subspace selection based on the structure risk criterion is used to select those subspaces with lower structure risks. The selected subspaces are then combined and used to evaluate the test utterances. We evaluate this subspace strategy on the 10sec-10sec test condition in 2008 NIST speaker recognition evaluations, achieving a relative 12.16% equal error rate reduction over the GMM supervector baseline system.

关键词： Speaker recognition Support vector machines Testing Gaussian distribution NIST Modular construction Speech Maximum likelihood linear regression Strontium Training data

来源：评论

学校读者我要写书评

暂无评论

Differentiated Bandwidth Allocation for Reducing Server Load in P2P VOD

Differentiated Bandwidth Allocation for Reducing Server Load...

引用

International Conference on Grid and Cooperative computing (GCC)

作者： Zhengjun Chena Kaiping Xue Peilin Hong Hancheng Lu University of Science and Technology of China Hefei Anhui CN MOE-MS Key Laboratory of Multimedia Computing and Communication and The Information Network Laboratory of EEIS Department University of Science and Technology Hefei China The Information Network Lab of EEIS Department University of Science and Technology of China (USTC) Hefei China The Information Network Laboratory of EEIS Department University of Science and Technology Hefei China

P2P technology offers a promising scalable solution for video-on-demand (VOD) service. Recent studies have found that departure misses are the major cause of server load in P2P-based VOD. Inspired by this finding, this paper addresses the design of bandwidth allocation policy to decrease departure misses and thereby reduce server load further. We first formulate the minimum departure misses problem. Then we propose a centralized algorithm which serves as our benchmark for all other schemes. The centralized algorithm makes use of surplus bandwidth to help peers prefetch chunks, thus reducing departure misses significantly. We then propose a distributed bandwidth allocation algorithm in which a stable peer with higher playback position is able to obtain a larger share of parentspsila upload bandwidth. A simple predictor is developed for stable node identification. Simulation results show that our distributed protocol outperforms traditional approach, and close to the centralized one.

关键词： Channel allocation Bandwidth Streaming media Peer to peer computing Network servers Web server Computer networks Grid computing multimedia computing Prefetching

来源：评论

学校读者我要写书评

暂无评论

Adaptive label-driven scaling for latent semantic indexing

Adaptive label-driven scaling for latent semantic indexing

引用

31st Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, ACM SIGIR 2008

作者： Xiaojun, Quan Enhong, Chen Qiming, Luo Hui, Xiong MOE-MS Key Laboratory of Multimedia Computing and Communication USTC MSIS Department Rutgers University

ISBN: (纸本)9781605581644

This paper targets on enhancing Latent Semantic Indexing (LSI) by exploiting category labels. Specifically, in the term-document matrix, the vector for each term either appearing in labels or semantically close to labels is scaled before performing Singular Value Decomposition (SVD) to boost its impact on the generated left singular vectors. As a result, the similarities among documents in the same category are increased. Furthermore, an adaptive scaling strategy is designed to better utilize the hierarchical structure of categories. Experimental results show that the proposed approach is able to significantly improve the performance of hierarchical text categorization.

关键词： Singular value decomposition

来源：评论

学校读者我要写书评

暂无评论

Image quality assessment metrics by using directional projection

引用

Chinese Optics Letters 2008年第7期6卷 491-494页

作者：庞建新张荣张晖黄轩刘政凯 MOE-Microsoft Key Laboratory of Multimedia Computing and Communication Department of Electronic Engineering and Information ScienceUniversity of Science and Technology of ChinaHefei 230027

Objective image quality measure, which is a fundamental and challenging job in image processing, evaluates the image quality consistently with human perception automatically. On the assumption that any image distortion could be modeled as the difference between the directional projection-based maps of reference and distortion images, we propose a new objective quality assessment method based on directional projection for full reference model. Experimental results show that the proposed metrics are well consistent with the subjective quality score.

关键词： PSNR SVD JPEG Image quality assessment metrics by using directional projection DMOS

来源：评论

学校读者我要写书评

暂无评论

Online topic-level user interest model for personalized recommendation

引用

Journal of Computational Information Systems 2008年第3期4卷 1257-1264页

作者： Lu, Ming Yu, Nenghai Rui, Xiaoguang MOE-Microsoft Key Laboratory of Multimedia Computing and Communication Dept. of Electronic Engineering and Information Science University Of Science Technology of China Hefei China

The rapid growth of the Internet has brought about the dramatic accumulation of data and the increasing possibility of information sharing. As the population on the Web grows, the analysis of user interests and behaviors will provide hints on how to improve the quality of service, such as recommendation systems. In this paper, we proposed a topic level user interest model based on the Latent Dirichlet Allocation (LDA) model, and an online supervised learning approach-Bayesian Online Perception is applied to construct and update the user interest model quickly and accurately. When recommending web pages to a certain user, both content-base and collaborative-based recommendation are performed based on our proposed user interest model. The experimental results showed that our approach was more effective than the other typical approaches to construct user interest model.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

Mining the structure of BBS virtual community

引用

Journal of Computational Information Systems 2008年第3期4卷 1249-1256页

作者： Chen, Jinrui Yu, Nenghai Mai, Lin Liu, Yijie MOE-Microsoft Key Laboratory of Multimedia Computing and Communication Dept. of Electronic Engineering and Information Science University of Science Technology of China Hefei China

In this paper, a user relationship model is proposed according to social network theory. Sub-group information and personal characters of users such as quantities of liveness, agency-level and reputation have been computed. The proposed user relationship model is applied to mine the structure of BBS virtual community. Experimental results on real BBS site demonstrate the effectiveness of the proposed model.

关键词： Social networking (online)

来源：评论

学校读者我要写书评

暂无评论

A ROBUST EYE LOCALIZATION ALGORITHM FOR FACE RECOGNITION

引用

Journal of Electronics(China) 2008年第3期25卷 337-342页

作者： Zhang Wencong Li Xin Yao Peng Li Bin Zhuang Zhenquan MOE-Microsoft Key Laboratory of Multimedia Computing and Communication University of Science and Technology of China Hefei 230026 China

The accuracy of face alignment affects greatly the performance of a face recognition system. Since the face alignment is usually conducted using eye positions, the algorithm for accurate eye lo- calization is essential for the accurate face recognition. In this paper, an algorithm is proposed for eye localization. First, the proper AdaBoost detection is adaptively trained to segment the region based on the special gray distribution in the region. After that, a fast radial symmetry operator is used to pre- cisely locate the center of eyes. Experimental results show that the method can accurately locate the eyes, and it is robust to the variations of face poses, illuminations, expressions, and accessories.

关键词： Eye localization AdaBoost Fast radial symmetry

来源：评论

学校读者我要写书评

暂无评论

ASYMBOOST-BASED FISHER LINEAR CLASSIFIER FOR FACE RECOGNITION

引用

Journal of Electronics(China) 2008年第3期25卷 352-357页

作者： Wang Xianji Ye Xueyi Li Bin Li Xin Zhuang Zhenquan MOE-Microsofl Key Laboratory of Multimedia Computing and Communication University of Science and Technology of China Hefei 230025 China

When using AdaBoost to select discriminant features from some feature space (e.g. Gabor feature space) for face recognition, cascade structure is usually adopted to leverage the asymmetry in the distribution of positive and negative samples. Each node in the cascade structure is a classifier trained by AdaBoost with an asymmetric learning goal of high recognition rate but only moderate low false positive rate. One limitation of AdaBoost arises in the context of skewed example distribution and cascade classifiers: AdaBoost minimizes the classification error, which is not guaranteed to achieve the asymmetric node learning goal. In this paper, we propose to use the asymmetric AdaBoost (Asym- Boost) as a mechanism to address the asymmetric node learning goal. Moreover, the two parts of the selecting features and forming ensemble classifiers are decoupled, both of which occur simultaneously in AsymBoost and AdaBoost. Fisher Linear Discriminant Analysis (FLDA) is used on the selected fea- tures to learn a linear discriminant function that maximizes the separability of data among the different classes, which we think can improve the recognition performance. The proposed algorithm is dem- onstrated with face recognition using a Gabor based representation on the FERET database. Ex- perimental results show that the proposed algorithm yields better recognition performance than AdaBoost itself.

关键词： AsymBoost AdaBoost Gabor feature Face recognition

来源：评论

学校读者我要写书评

暂无评论

Feature Annotation for visual concept detection in ImageCLEF 2008

Feature Annotation for visual concept detection in ImageCLEF...

引用

2008 Working Notes for CLEF Workshop, CLEF 2008 - Co-located with the 12th European Conference on Digital Libraries, ECDL 2008

作者： Jiang, Jingtian Rui, Xiaoguang Yu, Nenghai MOE-Microsoft Key Laboratory of Multimedia Computing and Communication Department of EEIS University of Science and Technology of China China

This paper shows our work on CLEF 2008. Our group joined the Visual Concept Detection Task of ImageCLEF 2008 this year. We submitted one run (run id: HJ-FA) for the evaluation. In the run, we applied a method called "Feature Annotation" to detect visual concept for the predefined concepts and we want to know how this information help in solving the photographic retrieval task. The applied method selected high level features for each concept from both local and global features, based on which the visual concepts are detected. The applied method consists of three procedures. First, feature extraction in which both local and global features are extracted from images. Then, a clustering algorithm is applied to "annotate the features". In this procedure, the features are affiliated with their corresponding concepts. Finally, we applied KNN algorithm to classify tests images according to the training images with the annotated features. The experiments were performed on the given training and test data on the 17 concepts. The paper concludes with an analysis of our results. Finally we identify the weaknesses in our approach and ways in which the algorithm could be optimized and improved.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

An improved density-sensitive semi-supervised clustering algorithm

An improved density-sensitive semi-supervised clustering alg...

引用

5th International Conference on Visual Information Engineering, VIE 2008

作者： Wu, Yulong Yuan, Pingbo Yu, Nenghai MOE-MS Key Laboratory of Multimedia Calculation and Communication University of Science and Technology of China Hefei 230027 China

ISBN: (纸本)9780863419140

This paper presents an improved density-sensitive distance measurement, which can effectively enlarge the distances among data points in different high density regions and shorten the distances among data points in the same high density region. Furthermore, a semi-supervised learning algorithm named improved density-sensitive semi-supervised clustering (IDS-SC) algorithm is introduced based on this distance measurement. The results demonstrate the superiority of IDS-SC in the application of Coral image set. ©2008 The Institution of Engineering and Technology.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：