By Topic

Neural Networks and Signal Processing, 2008 International Conference on

Date 7-11 June 2008

Filter Results

Displaying Results 1 - 25 of 148
  • Asymptotic analysis of nonlinear electronic circuits

    Publication Year: 2008 , Page(s): 1 - 4
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (819 KB) |  | HTML iconHTML  

    It is very important to analyze mixers and modulators for designing communication circuits. They are driven by multiple frequencies, one of which is usually very high carrier frequency compared to the other. To know the transient behaviors, we need to calculate many carrier waveforms, so that the transient analysis is very time-consuming. Hence, we propose an efficient envelope analysis for calculating the asymptotic behaviors of the amplitudes, which is based on the harmonic balance (HB) method with the slowly varying coefficients. In order to develop the Spice-oriented simulators, the Fourier expansions of nonlinear devices such as bipolar transistors are executed with MATLAB, and the Fourier modules should be stored in our computer library. Thus, we can easily formulate the determining equations of HB method called sine-cosine circuit. We found from many examples that the envelope method is several ten times faster than the transient analysis. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Back-propagation with chaos

    Publication Year: 2008 , Page(s): 5 - 8
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (450 KB) |  | HTML iconHTML  

    Multilayer feed-forward neural networks are widely used based on minimization of an error function. Back-propagation is a famous training method used in the multilayer networks but it often suffers from a local minima problem. To avoid this problem, we propose a new back-propagation training based on chaos. We investigate whether randomicity and ergodicity property of chaos can enable the learning algorithm to escape from local minima. Validity of the proposed method is examined by performing simulations on three real classification tasks, namely, the Ionosphere, the Wincson Breast Cancer (WBC), and the credit-screening datasets. The algorithm is shown to work better than the original back-propagation and is comparable with the Levenberg-Marquardt algorithm, but simpler and easier to implement comparing to Levenberg-Marquardt algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A closed-form semi-blind solution to MIMO-OFDM channel estimation

    Publication Year: 2008 , Page(s): 9 - 13
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (339 KB) |  | HTML iconHTML  

    Semi-blind channel estimation as a combination of the training-based or pilot-assisted method and a pure blind approach is considered to be a feasible solution for practical wireless systems due to its better estimation accuracy as well as spectral efficiency. However, in the existing semi-blind channel estimation techniques, the weighting factor employed to trade off the training-based and the blind criteria has not been appropriately determined. In this paper, a closed-form solution is developed for semi-blind channel estimation of MIMO-OFDM systems. An appealing scheme for the computation of the weighting factor is proposed, leading to an analytical expression for the weighting factor in terms the MSE (mean square error) of the training-based criterion and that of the blind part. A number of computer simulation-based experiments are conducted, confirming the effectiveness of the derived semi-blind solution. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dynamic hand gesture recognition using a CNN model with 3D receptive fields

    Publication Year: 2008 , Page(s): 14 - 19
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (471 KB) |  | HTML iconHTML  

    In this paper, a pattern recognition model for dynamic hand gesture recognition is proposed. The proposed model combines a convolutional neural network (CNN) with a weighted fuzzy min-max (WFMM) neural network; each module performs feature extraction and feature analysis, respectively. The data representation proposed in this research is a spatiotemporal template which is based on the motion information of the target object. To process the data, we develop a modified CNN model by extending the receptive field to a three-dimensional structure. To increase the efficiency of the pattern classifier, we use a feature analysis technique utilizing the WFMM algorithm. The experimental results show that the proposed method can minimize the influence caused by the spatial and temporal variation of the feature points. The recognition performance using only the selected features for the classification process is evaluated. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research on finding community structure based on filtration network model

    Publication Year: 2008 , Page(s): 20 - 23
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (321 KB) |  | HTML iconHTML  

    By defining community recursive coefficient M, we propose a new efficient algorithm called filtration split algorithm for discovering community structure in complex networks. By optimizing the M of child-networks based on dynamic recursive principle, the local communities are discovered automatically. Theoretical analysis and experiment results show that the algorithm can filtrate more than one edge once and make the networks split in parallel. For a network with n vertices, m edges, and c communities, the computation complexity is less than O((c+1)m+(c+1)). For many real-world networks are sparse m~n and c+1 Ltn, our algorithm can run in essentially linear time O((c+1)n). View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Exploring complex networks by thermal flux spreading

    Publication Year: 2008 , Page(s): 24 - 27
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (308 KB) |  | HTML iconHTML  

    Thermal conduction physical method based on finite element model (FEM) is utilized in searching for the shortest path between two nodes in complex networks. The thermal flux matrix of nodes is constructed, and the element with the maximal value in each node thermal flux vector indicates one section of the shortest path to the source node. Theoretical analysis and experimental results show that the strategy can avoid the problem of bringing high flux of inquiring data packet into network and the correct shortest path form source node to any other nodes can be found simultaneously. The shortest path from the source node to any other nodes can be found in time O(mt/Deltat), where m is the number of edges, t is the conduction time, Deltat is the integral interval, which is essentially in linear time with m, and all the shortest path between any two nodes can be found in time O(nmt/2Deltat). View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Symbolic vector dynamics for processing chaotic signals II: Noise reduction

    Publication Year: 2008 , Page(s): 32 - 36
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (302 KB) |  | HTML iconHTML  

    The estimation precision of symbolic vector dynamic method is determined by the symbol error of symbolic sequence. If we get the original symbolic sequences from the noisy signal, then we can recover the original sequence without any error. The aim of this paper is to further develop symbolic vector dynamical estimation method which has been proposed [K. Wang, Phys. Lett. A 367 (2007) 316-321]. We will prove that any ML methods using MMSE criterion can not correct the symbolic error, because they mistakenly take the error generated by symbolic error as the noise. We can use SVD to develop a novel estimation method, which will correct symbolic error in high SNR. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Symbolic vector dynamics for processing chaotic signals I: Communication

    Publication Year: 2008 , Page(s): 28 - 31
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (279 KB) |  | HTML iconHTML  

    The idea of using chaotic signals in the field of communication attracts the attention of researchers as well as engineers. The aim of this paper is to develop data transmission schemes in which the demodulation process is merged with SVD based noise estimation method utilizing CMLs. Because the estimation method of SVD only utilizes the symbol vector sequence and does not care the practical value, our modulation-demodulation scheme based on the estimation method of SVD can be expected to give accurate estimates for information symbol vector at low SNR. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An option-based empirical investigation of Chinese corporate liquidity value

    Publication Year: 2008 , Page(s): 37 - 42
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (336 KB) |  | HTML iconHTML  

    Corporate liquidity pricing is a new topic. With the discussion of the essence of liquidity, this paper established an exchange-option-based corporate liquidity pricing model which combines the investment option and insurance option taking into considerations of their execution probabilities. And we apply the model to investigate the liquidity value of Chinese firms with the data of the listed companies in Shanghai stock market. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Estimation of induction motor speed based on artificial neural networks inversion system

    Publication Year: 2008 , Page(s): 43 - 47
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (363 KB) |  | HTML iconHTML  

    As rotation speed is necessary for high-performance induction motor control, how to estimate the speed quickly and accurately is concerned by most scholars. On the analysis of theoretic invertibility of the induction motorpsilas mathematic model, a speed estimation based on neural networks inversion is proposed. The structure of multi-layer feed-forward neural network (MFNN) is trained by advanced backpropagation arithmetic. Also the achievement method and experiment results were given. The results show that the responses based on ANN inversion method can track the rotation speed quickly and accurately. The method proposed is effective in application. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Orientation and damage inspection of insulators based on Tchebichef moment invariants

    Publication Year: 2008 , Page(s): 48 - 52
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (406 KB) |  | HTML iconHTML  

    Based on the Tchebichef moment invariants, methods of orientation and damage inspection of insulators which was applied in the vision system of inspection robot on the power transmission lines was proposed in the paper. The image of insulators was first subjected to a normalization process to obtain rotation, scale and translation invariance. As feature vector, the Zernike moment invariants were then extracted from the normalized images. The recognition was realized by nearest neighbor feature matching. At last, damage of insulators was inspected by the gray level change rate of longitudinal tangent. Compared with Zernike moments and Hu moments, the accuracy of Tchebichef moment invariants performed much better. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A fractal network model with tunable fractal dimension

    Publication Year: 2008 , Page(s): 53 - 57
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (303 KB) |  | HTML iconHTML  

    We propose a model for growing fractal networks based on the mechanisms learned from the diffusion-limited aggregation (DLA) model in fractal geometries in the viewpoint of network. By studying the DLA network, our model introduces multiplicative growth, aging and geographical preferential attachment mechanisms, whereby featuring topological self-similar property and hierarchical modularity. According to the results of theoretical analysis and simulation, the degree distribution of the proposed model shows a mixed degree distribution (i.e., exponential and algebraic degree distribution) and the fractal dimension and clustering coefficient can be tuned by changing values of parameters. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Error and attack tolerance of the fractal network modelwith tunable fractal dimension

    Publication Year: 2008 , Page(s): 58 - 61
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (346 KB) |  | HTML iconHTML  

    Inspired by the diffusion-limited aggregation (DLA) model, we propose a new fractal network model with tunable fractal dimension. By introducing multiplicative growth, aging and geographical preferential attachment mechanisms, our model not only has a fractal topological structure but also hierarchical modularity. This paper focuses on the error and attack tolerance of the model. By changing the values of the model parameters, we investigate the robustness of the model by observing the response of the networks reproduced by our model under the fragmentation process. Simulation results show that networks produced by our model have a significant higher robustness than other non-fractal networks, such as Internet. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Chaotic Cyclic Attractors Shift Keying

    Publication Year: 2008 , Page(s): 62 - 66
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (272 KB) |  | HTML iconHTML  

    In this paper, an extension of the existing chaotic digital modulation named chaotic cyclic attractors shift keying (CCASK) is proposed, which uses different orders of chaotic cyclic attractors to modulate the information. The new scheme is analysed under a discrete-time domain, and the noise performance in AWGN channel is simulated, and also calculated in a mixed analytic-numerical way, under some reasonable hypothesis. As a noncoherent auto-correlation receiver, CCASK doesn't require chaotic synchronization contrary to CSK, neither the instable threshold values contrary to COOK. Furthermore, the analysis and simulation noise performance results prove that CCASK gains 3.5 dB better noise performance than DCSK. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Solving permutation problem in frequency-domain blind source separation using microphone sub-arrays

    Publication Year: 2008 , Page(s): 67 - 72
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (281 KB) |  | HTML iconHTML  

    Blind source separation for convolutive mixtures can be solved effectively in the frequency domain where independent component analysis is performed in each frequency independently. However, the permutation problem arises: the permutation ambiguity of ICA in each frequency bin should be aligned so that a separated signal in the time-domain contains frequency components of the same source signal. In this paper, we present a new method for solving the permutation problem using microphone sub-arrays. It is based on the combination of two approaches: direction of arrival (DOA) estimation for sources and the inter-frequency correlation of signal envelopes. First, DOA estimation is performed using microphone sub-arrays so that the permutation problem is solved more robustly in low frequencies. Second, we exploit the correlation between the adjacent bins to fix the permutation for the remaining frequencies. Experimental results show that the proposed method provided a more robust solution to the permutation problem in a real acoustic environment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Novel image watermarking scheme based on ICA

    Publication Year: 2008 , Page(s): 73 - 77
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (772 KB) |  | HTML iconHTML  

    In this paper, a novel image watermarking scheme based on independent component analysis (ICA) is proposed. In the proposed scheme, the original image is sub-sampled firstly to obtain four sub-images, and the features of the original image are extracted by performing ICA to these sub-images, and then the feature, which has the largest variance, is selected out to be watermarked on its block DCT coefficients. During the embedding, human visual system (HVS) is used to determine which regions should be watermarked in order to balance the imperceptibility and the robustness. Experiments show that the watermark can be extracted blindly, and it is robust to common watermark attack and image processing, such as JPEG compression, noising, scaling, filtering, histogram equalization, and so on. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Study on high capability intelligent algotrithm in load flow of power system

    Publication Year: 2008 , Page(s): 78 - 82
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (314 KB) |  | HTML iconHTML  

    Power flow calculation is the most important and basic calculation in power system. This calculation is not only a very important calculation to seize power system planning and operation of the system state, but also a very important calculation for the analytical systems such as the stability index. The algorithm is usually Newton-Larfson , the P-Q algorithm, etc. However, the traditional methods of power system, with unknown parameters expended in Exponential increase becomes complex and make the Power system more difficult to create a mathematical model and solution. In this paper, a kind of traditional BP neural network algorithm with the parallel computing function was firstly recommended for the simulation of the flow calculation ;and then another kind of high-order BP neural network was recommended for the simulation, and the result shows that the algorithm can effectively resolve the difficulties of traditional modeling algorithm, and high calculating cost, and the poor in real-time. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Shipboard target location based on blind source separation

    Publication Year: 2008 , Page(s): 83 - 86
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (353 KB) |  | HTML iconHTML  

    Passive acoustic source location and tracking implemented by microphone array has been applied broadly to target detection and fault diagnosis etc. The conventional relative acoustic source location methods often obtain poor location results in complicated environments, especially in impulse noise environments. The proposed paper applies blind source separation to shipboard target location and thoroughly analyzed this new location algorithm. The experiments in this paper show that a more stable and accurate location results may be obtained by this new location way. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Coarse-to-fine based matching for audio commercial recognition

    Publication Year: 2008 , Page(s): 87 - 90
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (383 KB) |  | HTML iconHTML  

    In this paper, an automatic audio commercial recognition based on coarse-to-fine matching strategy is proposed. First, to index audio commercials in the database efficiently, the locality sensitive hash (LSH) is applied to accelerating the initial retrieval procedure. Second, with respect to the collision occurred in the initial LSH matching, a fine matching procedure based on Fine Granularity Successive Elimination (FGSE) is triggered out to eliminate rapidly those irrelevant candidates which have passed the initial coarse matching. At last, an elaborate post processing with correlation in time domain is presented for further decreasing false alarms following fine matching. The promising experiment results show that satisfactory precision and recall related to audio commercial recognition are obtained, achieving 99.4% and 97.5% respectively. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Regularization super-resolution image fusion considering inaccurate image registration and observation noise

    Publication Year: 2008 , Page(s): 91 - 94
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (617 KB) |  | HTML iconHTML  

    In this paper, a kind of super-resolution image fusion algorithm is proposed to regularize the distortion of the reconstructed high-resolution (HR) image caused by the inaccurate image registration and the observation noise. For this purpose, the registration error, caused by inaccurate image registration, is considered as the noise mean added in the observation noise known as additive white Gaussian noise (AWGN). Based on this consideration, two constraints are regulated pixel by pixel within the framework of Millerpsilas regularization, and combined with regularization parameters to construct one cost function. The regularization parameters are adaptively estimated in each pixel in terms of the registration error, as well as in each observation channel in terms of the AWGN. Simulation shows that the proposed regularized SR algorithm can fuse the information from multiple LR images effectively and achieve the reconstructed HR images with much sharper edges and higher PSNR. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A new approach to image compression using vector quantization of wavelet coefficients

    Publication Year: 2008 , Page(s): 95 - 98
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (520 KB) |  | HTML iconHTML  

    Traditional image coding methods, such as vector quantization (VQ), discrete cosine transform (DCT) based coding, and entropy coding of subband, have been designed to eliminate statistical redundancy within still images. In this paper, a combined approach utilizing both transform coding and vector quantization techniques is used, hoping to achieve the best result in terms of compression ratio with acceptable recovery quality. The transform coding used is 2-D wavelet transform and the key is to tap the correlation between wavelet coefficients of different subbands in the same spatial location rather than only in the same orientation. Performance comparisons are made with three other VQ-based compression models. The result shows the strength of this novel approach in that it has the best reconstructed image quality in terms of its signal to noise ratio for a fixed compression ratio. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A blind super-resolution framework considering the sensor PSF

    Publication Year: 2008 , Page(s): 99 - 103
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (398 KB) |  | HTML iconHTML  

    Blind super-resolution, which incorporates blur identification into super-resolution, has been proposed to restore the degraded image with either partially known or totally unknown blur. In the conventional framework for blind super-resolution, the sensor PSF caused by camera lens/CCD has always been ignored. Therefore, the identified blur and restored HR image are affected by the sensor PSF, especially when the support of identified blur is relatively small. In this paper, we propose a blind super-resolution framework considering the sensor PSF to solve this problem. In the proposed framework, the sensor PSF and the identified external blur are considered separately, and then Error-Parameter-Analysis algorithm is applied to implement the blind super-resolution. Experiments show that the proposed framework can enhance the accuracy of the blur identification and the quality of the restored HR image. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research and implementation of channel estimation of digital radio receiver based on OFDM

    Publication Year: 2008 , Page(s): 104 - 108
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (315 KB) |  | HTML iconHTML  

    OFDM system can improve frequency band utility, decrease symbols interference caused by multi-path effect. It also can obtain reliable data transmission with high speed. The paper studies the channel estimation technique in a digital radio broadcast receiver based on OFDM-DRM. Due to the time and frequency selective variance of the channel encountered during OFDM transmission, the channel response has to be estimated in order to get exact data. This is done by using the known scattered pilots transmitted in OFDM symbols. OFDM based digital radio system adopts linear interpolation, windowed FFT interpolation and Wiener interpolation, in which Wiener interpolation is the best but requires a large quantity of computational consumption. The paper analyzes the Wiener algorithm..and simulates the MSE versus SNR for different Wiener filtering methods, and presents the flow charts of Wiener algorithm and realizes them in the computer. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A chaotic SS code generating method and its application to a DS-UWB system

    Publication Year: 2008 , Page(s): 109 - 113
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (424 KB) |  | HTML iconHTML  

    The presently existing methods for generating chaotic binary sequence under finite precision are first studied in this paper, the short cyclic behavior of chaotic sequences is analyzed in detail. At the same time, a chaotic similar function is defined for presenting the cyclic characteristics of the sequences. Based on these efforts, an improved method with scrambling control for generating chaotic binary sequences is proposed. To quantitatively describe the improvement of periodic performance of the sequences, an orthogonal estimator is then defined. Some simulating experiments are also given in this paper. We also apply the short chaotic binary sequence, which is generated by our proposed method, to a multiuser DS-UWB system as a SS code, and compare the application performance with that of the conventional Kasami code. All the results in this paper demonstrate that our proposed method can effectively increase the period of the short chaotic binary code, and more evidently improve the performance of the DS-UWB system. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A cepstral domain algorithm for formant frequency estimation from noise-corrupted speech

    Publication Year: 2008 , Page(s): 114 - 119
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (996 KB) |  | HTML iconHTML  

    A new scheme for the estimation of formant frequencies from noise-corrupted speech signals is presented in this paper. In order to overcome the effect of noise, first, instead of conventional autocorrelation function (ACF), a once-repeated ACF of the observed data is employed. A ramp cosine cepstrum model of the ORACF of speech signal is developed, followed by a model-fitting based least-square optimization to extract the formants. For the purpose of implementation, the discrete cosine transform (DCT) is used which offers computational advantages for real signals and solves the phase unwrapping problem. Synthetic and natural vowels as well as some naturally spoken sentences in noisy environments are tested. The experimental results demonstrate a better performance obtained by the proposed scheme in comparison to some of the existing methods at low levels of signal-to-noise ratio. View full abstract»

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