检索结果-内蒙古大学图书馆

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Zhenyu Xiong T.F. Zheng Zhanjiang Song Wenhu Wu Center for Speech Technology State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing China Beijing d-Ear Technologies Company Limited China

In this paper a new method of reducing the computational load for Gaussian mixture model universal background model (GMM-UBM) based speaker identification is proposed. In order to speed up the selection of N-best Gaussian mixtures in a UBM, a selection tree (ST) structure as well as relevant operations is proposed. Combined with the existing observation reordering pruning (ORP) method which was proposed for rapid pruning of unlikely speaker model candidates, the proposed method achieves a much larger computation reduction factor than any single individual method. Experimental results show that a GMM-UBM system used in a conjunction with ST and ORP can speed up the computation by a factor of about 16 with an error rate increase of only about 1% compared with a baseline GMM-UBM system.

关键词： Diversity reception Speech Laboratories intelligent systems Computational intelligence Computer science Error analysis Speaker recognition NIST Computer applications

来源：评论

学校读者我要写书评

暂无评论

A novel approach to fingerprint ridge line extraction

A novel approach to fingerprint ridge line extraction

引用

International Symposium on Communications and Information Technologies (ISCIT)

作者： R. Ma Yaxuan Qi Changshui Zhang Jiaxin Wang State Key Laboratory of Intelligent Technology and Systems Tsinghua University Beijing China Research Institute of Information Technologies Department of Automation Tsinghua University Beijing China

Extraction of fingerprint ridge lines is a critical pre-processing step in the identification of poor quality fingerprint images. This paper presents a new fingerprint ridge line extraction approach by way of ridge line tracing. In our research, the fingerprint ridge line in a gray scale image is viewed as a track of a ridge segment moving along the ridge. The curve tracing problem is solved by the target tracking technique in computer vision. We first formulate the model of fingerprint ridge line segments and then apply a target tracking method to trace each of the ridge lines. In addition, a feedback technique is adopted to correct the fingerprint directional image in each tracing step in order to improve tracing accuracy. By connecting all the traced ridge line segments, a polyline extraction of the ridge line can be obtained. Compared to conventional method, our approach exhibits superior performance.

关键词： Fingerprint recognition Feedback Data mining Image matching Target tracking Image segmentation Filtering Laboratories intelligent systems Information technology

来源：评论

学校读者我要写书评

暂无评论

Automatic content based title extraction for Chinese documents using support vector machine

Automatic content based title extraction for Chinese documen...

引用

IEEE International Conference on Natural Language Processing and Knowledge Engineering (NLP-KE)

作者： Zhengcao Zhang Maosong Sun Shaoming Liu The State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Haidian China Corporate Research Group Fuji Xerox Company Limited Kanagawa Japan

In this paper, a content-based and domain-independent method for automatically extracting titles from Chinese research papers is proposed. The information contained in the title itself and the similarity between the title and the body of the paper is exploited, under the condition that the experiment is carried out on plain texts in which no any format information such as font is used. A list of words only used in Chinese titles and a list of words never used in Chinese titles are further collected to facilitate the title extraction. We use the support vector machine classifier to perform a robust and more adaptable automatic title extraction. The method achieves good performance on a test set consisting of 2438 research papers which cover almost all of the academic disciplines.

关键词： Support vector machines Data mining Support vector machine classification Internet Sun Laboratories intelligent systems Machine intelligence Computer science Robustness

来源：评论

学校读者我要写书评

暂无评论

基于HMM的地形高度匹配算法

引用

自动化学报 2005年第6期31卷 960-964页

作者：冯庆堂沈林成常文森叶媛媛 State Key Laboratory of Intelligent Technology and Systems Tsinghua University Beijing 100084School of Electromechanical Engineering and Automation National University of Defense Technology Hunan 410073 School of Electromechanical Engineering and Automation National University of Defense Technology Hunan 410073

Terrain-aided navigation (TAN) uses terrain height variations under an aircraft to render the position estimate to bound the inertial navigation system (INS) error. This paper proposes a new terrain elevation matching(TEM) model, viz. Hidden-Markov-model(HMM) based TEM (HMMTEM) model. With the given model, an HMMTEM algorithm using Viterbi algorithm is designed and implemented to estimate the position error in INS. The simulation results show that HMMTEM algorithm can better improve the positioning precision of autonomous navigation than SITAN algorithm.

关键词： HMM 地形高度匹配算法半自动导航技术定位技术

来源：评论

学校读者我要写书评

暂无评论

An Overview of Temporal Information Extraction

引用

International Journal of Computer Processing of Languages 2005年第2期18卷 137-152页

作者： KAM-FAI WONG YUNQING XIA WENJIE LI CHUNFA YUAN Department of Systems Engineering and Engineering Management The Chinese University of Hong Kong Shatin Hong Kong China Department of Computing Hong Kong Polytechnic University Hong Hum Hong Kong China State Key Laboratory of Intelligent Technology and System Tsinghua University Beijing 100084 China

Research of temporal Information Extraction was regarded as a subtask of named entity recognition in 1990's. To date, the scope of this research is broadened, ranging from temporal expression extraction and annotation to temporal reasoning and understanding. This area of research is now a hot NLP topic and the results are applicable to question answering, information extraction, text summarization, etc. This paper presents the past, present and future research development in temporal information extraction.

关键词： Temporal information extraction Natural language processing Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Multi-view face detection based on real Adaboost algorithm

引用

Jisuanji Yanjiu yu Fazhan/Computer Research and Development 2005年第9期42卷 1612-1621页

作者： Wu, Bo Huang, Chang Ai, Haizhou Lao, Shihong State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing 100084 China Sensing Technology Laboratory Vision Sensing Technology Group Omron Corporation Kyoto 619-0283 Japan

In this paper, a multi-view face detection method based on real Adaboost algorithm is presented. Human faces are divided into several viewpoint categories according to their poses in 3D, and for each of these categories a form of weak classifiers in look-up-table (LUT) type is designed using Haar-like features that have confidences in real values as their outputs, and correspondingly its space of weak classifiers is constructed, from which the cascade face detector is learnt by using real Adaboost algorithm. For speed up, multi-resolution searching and pose prediction strategies are introduced. For frontal face detection, the experiments on CMU + MIT frontal face test set result in a correct rate of 94.5% with 57 false alarms;for multi-view face detection, the experiments on CMU profile face test set result in a correct rate of 89.8% with 221 false alarms. The average processing time on a PIV 2.4 GHz PC is about 80 ms for a 320 × 240-pixel image. It can be seen that the proposed method is very efficient and has significant value in application.

关键词： Haar-like feature LUT weak classifier Multi-view face detection Pose estimation Real Adaboost

来源：评论

学校读者我要写书评

暂无评论

A new improved method of motor speed identification for direct torque control of the asynchronous motor

A new improved method of motor speed identification for dire...

引用

International Conference on Electrical Machines and systems (ICEMS)

作者： Xiucheng Dong Haibin Wang Jung Wang Xiaoxiao Zhao School of Electrical and Information Engineering Xihua University Chengdu Sichuan China State Key Laboratory of Intelligent Technology and Systems Tsinghua University Beijing China Xihua University Chengdu Sichuan CN Sch. of Electr. & Inf. Eng. Xihua Univ. Sichuan China

Two speed estimators are designed respectively for direct torque control system of the asynchronous motor in this paper. One is MRAC identification model with high-pass and low-pass filter part based on traditional MRAC theory, besides, an improved u-n magnetic chain observer method has been adopted. We employ improved u-n magnetic chain observation method to get magnetic chain, enhance accuracy of the magnetic chain calculation, and overcome the limitation of pure integral magnetic chain calculation. Computer simulations are made in the high-speed section and low-speed section respectively. Simulation results prove that the improved method enhances identification precision and control performance of the whole system, and overcome fluctuates of the torque and motor speed. The other is the motor speed identification method based on neural network. This method not only needn't model on control objects, but also can overcome the effect of the changing parameters and improve the robustness of the system by ANN's learning and self-optimizing abilities. In addition, a new improved direct torque control (DTC) method based on the speed identification is proposed in this paper, and it has a better control performance, higher identification precision and practical significance. Finally, we adopt TMS320F240 digital signal processor to build relevant equipment system, and the validity of this method has been demonstrated by the experiment results.

关键词： Torque control Magnetic separation intelligent sensors Rotors Adaptive control Sensor systems and applications Voltage Neural networks Sensor phenomena and characterization Inductance

来源：评论

学校读者我要写书评

暂无评论

Novel Method Based on Ant Colony Optimization for Solving ill-conditioned Linear systems of Equations

引用

Journal of systems Engineering and Electronics 2005年第3期000卷 606-610页

作者： Duan Haibin Wang Daobo Zhu Jiaqiang School of Automation Science and Electrical Engineering，Beijing Univ.of Aeronautics and Astronautics，Beijing 100083，P.R.China Coll.of Automation Engineering，Nanjing Univ.of Aeronautics and Astronautics，Nanjing 210016，P.R.China Coll.of Automation Engineering，Nanjing Univ.of Aeronautics and Astronautics，Nanjing 210016，P.R.China State Key Laboratory of Intelligent Technology and Systems，Tsinghua Univ.Beijing 100084，P.R.China

A novel method based on ant colony optimization (ACO), algorithm for solving the ill-conditioned linear systems of equations is proposed. ACO is a parallelized bionic optimization algorithm which is inspired from the behavior of real ants. ACO algorithm is first introduced, a kind of positive feedback mechanism is adopted in ACO. Then, the solution problem of linear systems of equations was reformulated as an unconstrained optimization problem for solution by an ACO algorithm. Finally, the ACO with other traditional methods is applied to solve a kind of multi-dimensional Hilbert ill-conditioned linear equations. The numerical results demonstrate that ACO is effective, robust and recommendable in solving ill-conditioned linear systems of equations.

关键词： ill-conditioned linear systems of equations ant colony optimization condition number optimization

来源：评论

学校读者我要写书评

暂无评论

PCA-FA:Applying Supervised Learning to Analyze Gene Expression Data

引用

Tsinghua Science and technology 2004年第4期9卷 428-434页

作者：翁时锋张长水张学工 State Key Laboratory of Intelligent Technology and Systems Department of Automation Tsinghua UniversityBeijing 100084China

In previous gene expression data analyses, supervised learning has mainly focused on the clas-sification of attribute data, such as the different experimental conditions, different known classes of the same tumor and sex. However, supervised learning classification is not suitable for interval-scaled attributes, such as age and survival outcome of cancer patients. For this problem, this paper proposed a new method by combining two well-known methods: principal component analysis (PCA) and Fisher analysis (FA). The method, PCA-FA, realizes supervised learning with two types of attributes (nominal attributes and interval-scaled attributes). The fuzzy FA was introduced to model the interval-scaled attributes. In this paper, an ap-proximate linear relationship between gene expression data of lung adenocarcinoma patients and survival outcome is successfully revealed by PCA-TA.

关键词： supervised learning gene expression data principal component analysis Fisher analysis

来源：评论

学校读者我要写书评

暂无评论

Using Agents in Web-Based Constructivist Collaborative Learning System

引用

Tsinghua Science and technology 2004年第2期9卷 189-196页

作者：刘莹林福宗王雪 State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing 100084 China

Web-based learning systems are one of the most interesting topics in the area of the application of computers to education. Collaborative learning, as an important principle in constructivist learning theory, is an important instruction mode for open and distance learning systems. Through collaborative learning, students can greatly improve their creativity, exploration capability, and social cooperation. This paper used an agent-based coordination mechanism to respond to the requirements of an efficient and motivating learn-ing process. This coordination mechanism is based on a Web-based constructivist collaborative learning system, in which students can learn in groups and interact with each other by several kinds of communica-tion modes to achieve their learning objectives efficiently and actively. In this learning system, artificial agents represent an active part in the collaborative learning process; they can partially replace human in-structors during the multi-mode interaction of the students.

关键词： intelligent agents web-based education collaborative learning agent-based application

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：