6th International Conference on Signal Processing, 2002.

26-30 Aug. 2002

Go

Filter Results

Displaying Results 1 - 25 of 228
  • 2002 6th International Conference on Signal Processing Proceedings [front matter]

    Publication Year: 2002, Page(s):0_1 - xxiv
    Request permission for commercial reuse | PDF file iconPDF (1033 KB)
    Freely Available from IEEE
  • Research on cooperation and learning in multi-agent system

    Publication Year: 2002, Page(s):1159 - 1162 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (229 KB) | HTML iconHTML

    Cooperation and learning in multi-agent systems (MAS) is of special interest in DAI. This paper presents a cooperation model called MACM that provides a flexible coordination mechanism to support cooperation and learning in MAS. The learning agent adopts model-free distributed Q-learning. By using projection method, the distributed Q-learning algorithm needs less storage space for the Q-table than... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Author index

    Publication Year: 2002, Page(s):1855 - 1865
    Request permission for commercial reuse | PDF file iconPDF (284 KB)
    Freely Available from IEEE
  • An adaptive interference cancellation receiver

    Publication Year: 2002, Page(s):1271 - 1274 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (287 KB) | HTML iconHTML

    The paper proposes an adaptive interference cancellation receiver for code-division multiple access (CDMA). In the receiver, we first introduce a set of reliable factors based on the output signal of the ith user for its mth bit, then we can calculate MAI according to reliable factors and finally the symbol decision is performed by interference cancellation. By simulation in ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quantum neural network in speech recognition

    Publication Year: 2002, Page(s):1267 - 1270 vol.2
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (280 KB) | HTML iconHTML

    This paper describes a new kind of neural network - quantum neural network (QNN) and its application to recognition of continuous digits. QNN combines the advantages of neural modeling and fuzzy theoretic principles. Experiment results show that more than 15% error reduction is achieved on a speaker-independent continuous digits recognition task compared with the backpropagation (BP) network. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A kind of method for mining classification rules based on fuzzy sets

    Publication Year: 2002, Page(s):1263 - 1266 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (322 KB) | HTML iconHTML

    Fuzziness is one of the general characteristics of human thinking and objective things. This paper presents a kind of methods for mining classification rules based on fuzzy sets. This method first obtain thematic data set using fuzzy retrieval techniques, then sums up abstractly by means of fuzzy clustering, and thus the gains necessary classification rules. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dynamic grouping multi-classes face recognition method

    Publication Year: 2002, Page(s):1259 - 1262 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (353 KB) | HTML iconHTML

    The paper presents a dynamic grouping multi-class face recognition method, which has knowledge-increasable ability and can solve the problems of large classes face recognition and pattern classes dynamic extension. By adopting multiple classifiers parallel working in the process of training and dynamic grouping recognition, the method can not only speed up calculation and improve the recognition r... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Interactive gradient algorithm for radial basis function networks

    Publication Year: 2002, Page(s):1187 - 1190 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (260 KB) | HTML iconHTML

    In this paper the radial basis function neural network is divided into two parts: (1) the input and the hidden layer, (2) the output layer, and the parameters of the two parts are trained through an interactive gradient learning algorithm. Experimental results in function approximation are more attractive, which show that the algorithm not only avoids the slow rate of the conventional gradient alg... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A HME neural network knowledge-increasable model

    Publication Year: 2002, Page(s):1255 - 1258 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (308 KB) | HTML iconHTML

    The HME network divides a task into small tasks by the principle of divide and conquer to improve the performance of a single network. This approach often brings simple, elegant and efficient algorithms. By studying the dual manifold architecture for mixtures of neural networks and analyzing the probability of knowledge-increasable model based on information geometry, the paper proposes a new meth... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An output coding approach for knowledge increasable artificial neural network

    Publication Year: 2002, Page(s):1183 - 1186 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (321 KB) | HTML iconHTML

    How to inherit the learned knowledge of existing neural networks without destroying their structure and functionality is a difficult problem. In this paper, we propose an output coding approach for building such a system, which fully utilizes the information gained from the component neural units. By coding the neural outputs, a neural network becomes a self-contained system. For a given pattern, ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Wavelet neural networks for adaptive equalization

    Publication Year: 2002, Page(s):1251 - 1254 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (322 KB) | HTML iconHTML

    A structure based on the wavelet neural networks is proposed for nonlinear channel equalization in a digital communication system, the minimum error probability (MEP) is applied as performance criterion to update the weighting matrix of wavelet networks. Our experimental results show that performance of the proposed wavelet networks based on equalizer can significantly improve the neural modeling ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A dynamic denoising natural image compression

    Publication Year: 2002, Page(s):1179 - 1182 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (278 KB) | HTML iconHTML

    This paper provides a new compression of natural images, which transforms image data by independent component analysis (ICA), estimates every component based on the sparseness properties related to simple-cell receptive fields closely and compresses the transformed data making use of runlength coding. This method can not only compress natural images efficiently, but also remove the noise of the im... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The noise elimination system using the amplitude-locked loop with cochannel FM interference

    Publication Year: 2002, Page(s):1668 - 1671 vol.2
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (299 KB) | HTML iconHTML

    The frequency modulation (FM) signal was conceived to minimise the effect of additive noise - since additive noise is the principal source of signal corruption in amplitude modulation (AM), the pre-cursor of frequency modulation. When the FM signal is received the additive noise is removed by limiter action. This process is a necessary part of FM demodulation. Without hard limiting, the signal is ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive particle swarm optimization on individual level

    Publication Year: 2002, Page(s):1215 - 1218 vol.2
    Cited by:  Papers (19)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (311 KB) | HTML iconHTML

    An adaptive particle swarm optimization (PSO) on individual level is presented. By analyzing the social model of PSO, a replacement criterion, based on the diversity of fitness between the current particle and the best historical experience, is introduced to maintain the social attribution of swarm adaptively by taking off inactive particles. The testing of three benchmark functions indicates that... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An architecture of active learning SVMs for spam

    Publication Year: 2002, Page(s):1247 - 1250 vol.2
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (342 KB) | HTML iconHTML

    We propose a new method for spam categorization based on support vector machines (SVMs) using active learning strategy. We study the use of support vector machines in classifying e-mail as spam or nonspam. It analyzes the particular properties of our special task and identifies why SVMs are appropriate for dealing with spam. Instead of using a randomly selected training set, the learner has access... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improvement of brain lesions detection using information fusion approach

    Publication Year: 2002, Page(s):1104 - 1107 vol.2
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (367 KB) | HTML iconHTML

    Automatic segmentation of brain lesions, such as multiple sclerosis in MRI images, is a complex operation. One of the main difficulties is to optimize the dilemma between the false positives and false negatives present in the segmented image. We propose here a new approach to this problem. The idea is to exploit the complementary results from different segmentation algorithms as well as a priori k... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast block-based true motion estimation using distance dependent thresholds (DTS)

    Publication Year: 2002, Page(s):937 - 940 vol.2
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (344 KB) | HTML iconHTML

    A new fast motion estimation algorithm, called distance dependent thresholding search (DTS), is presented for block-based true motion estimation applications, and introduces the novel concept of variable distance dependent thresholds. The performance of the DTS algorithm is analyzed and quantitatively compared with both the traditional and exhaustive full-search (FS) technique, and the computation... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The quantization effects of different probability distribution on multilayer feedforward neural networks

    Publication Year: 2002, Page(s):1175 - 1178 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (343 KB)

    A statistical model of quantization was used to analyze the effects of quantization in digital implementation, and the performance degradation caused by number of quantized bits in multilayer feedforward neural networks (MLFNN) of different probability distribution. The performance of the training was compared with and without clipping weights for MLFNN. We established and analyzed the relationshi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The adaptive algorithms of the smart antenna system in 3G wireless communication systems

    Publication Year: 2002, Page(s):1664 - 1667 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (344 KB) | HTML iconHTML

    In a CDMA system, the smart antenna system is mainly applied to the base station. The smart antenna can adjust the direction pattern adaptively and reduce the interference signals using some adaptive interference ing algorithms, and thus enhancing the performance of the communication systems. The article provides a description and analysis of some basic adaptive algorithms and their modifications.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research and implementation of an improved Reed decoding algorithm

    Publication Year: 2002, Page(s):1762 - 1765 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (289 KB) | HTML iconHTML

    Reed-Muller code is a kind of decomposable code which can be decoded with multi-stage decoding. In this paper, we first analyze the theory of Reed-Muller coding and introduce a decoding Reed algorithm, which can recover information from received codes without obvious computation of syndrome. An improved Reed decoding algorithm is proposed, which can greatly reduce the complexity of producing a che... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generalized kernel function Fisher discriminant for pattern recognition

    Publication Year: 2002, Page(s):1075 - 1078 vol.2
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (302 KB) | HTML iconHTML

    In this paper, according to the concept of generalized Fisher (1938) discriminant (GFD) presented by Foley and Sammon (1975) , the generalized kernel function Fisher discriminant (GKFD) is investigated and proved based on the linear Fisher discriminant (LFD) and kernel function Fisher discriminant (KFD). It generalizes the solution of two-class pattern recognition nonlinearly, and the decision fun... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of recursive orthogonal least squares algorithms to training and the structure optimization of radial basis probabilistic neural networks

    Publication Year: 2002, Page(s):1211 - 1214 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (323 KB) | HTML iconHTML

    The paper introduces applying recursive orthogonal least squares algorithm (ROLSA) to training radial basis probabilistic neural networks (RBPNN) and selecting their hidden centers. First, ROLSA is used to solve the weights between the second layer and the output layer of RBPNN. Second, we interpret the basic principle of selecting hidden centers and give a detailed selection procedure. In additio... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of genetic algorithms to the structure optimization of radial basis probabilistic neural networks

    Publication Year: 2002, Page(s):1243 - 1246 vol.2
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (320 KB) | HTML iconHTML

    The genetic algorithm (GA) is applied in this paper to select hidden centers of radial basis probabilistic neural networks (RBPNN). The encoding method of individuals for GA, proposed in this paper, embodies not only the number but also the positions of selected centers. In addition, precision control is integrated into definition of the fitness function. Finally, we use the two-dimensional Gaussi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An advanced cluster analysis method based on statistical test

    Publication Year: 2002, Page(s):1100 - 1103 vol.2
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (332 KB) | HTML iconHTML

    The paper presents a novel clustering algorithm supervised by statistical tests. It deals simultaneously with three key data analysis problems: cluster tendency; cluster analysis; cluster validity. It provides an effective tool to analyze the validity and reasonableness of unsupervised pattern classification, especially in the case of a large number of samples. The experimental result illustrates ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A low bit-rate video-coding algorithm based upon variable pattern selection

    Publication Year: 2002, Page(s):933 - 936 vol.2
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (359 KB) | HTML iconHTML

    Recent research into pattern representation of moving regions in block-based motion estimation and compensation in video sequences has focused mainly upon using a fixed number of regularly shaped patterns. These are used to match the macroblocks in a frame that have two distinct regions involving static background and moving objects. A new variable pattern selection (VPS) algorithm is presented wh... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.