By Topic

Neural Networks for Signal Processing IX, 1999. Proceedings of the 1999 IEEE Signal Processing Society Workshop.

Date 25-25 Aug. 1999

Filter Results

Displaying Results 1 - 25 of 61
  • Neural Networks for Signal Processing IX: Proceedings of the 1999 IEEE Signal Processing Society Workshop (Cat. No.98TH8468)

    Publication Year: 1999
    Save to Project icon | Request Permissions | PDF file iconPDF (218 KB)  
    Freely Available from IEEE
  • Author index

    Publication Year: 1999 , Page(s): 565 - 566
    Save to Project icon | Request Permissions | PDF file iconPDF (61 KB)  
    Freely Available from IEEE
  • Near-optimal flight load synthesis using neural nets

    Publication Year: 1999 , Page(s): 535 - 544
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (368 KB)  

    This paper describes the use of neural networks for near-optimal helicopter flight load synthesis (FLS), which is the process of estimating mechanical loads during helicopter flight, using cockpit measurements. First, modular neural networks are used to develop statistical signal models of the cockpit measurements as a function of the loads. Then Cramer-Rao maximum a-posteriori bounds on the mean-... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On condition monitoring of exhaust valves in marine diesel engines

    Publication Year: 1999 , Page(s): 554 - 563
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (524 KB)  

    The feasibility of noninvasive characterisation of exhaust valve conditions in large marine diesel engines were experimentally investigated on a four cylinder 500 mm bore 2-stroke marine diesel engine at MAN B&W Diesel's Research Center in Copenhagen, Denmark. The experiments comprised three different valve conditions, where two were concerned with artificially induced valve burn-through situa... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Subspace techniques in speech enhancement

    Publication Year: 1999 , Page(s): 449 - 458
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (456 KB)  

    This paper concerns speech enhancement. Speech is first modelled and then model parameters used to design a Kalman smoother to reduce the background noise. The 4SID (subspace state space system identification), DOA (direction-of-arrival) and polynomial techniques are described within a common subspace state space framework, and assumptions for each technique stated. It is shown that 4SID methods m... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Gaussianity measure for blind source separation insensitive to the sign of kurtosis

    Publication Year: 1999 , Page(s): 58 - 66
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (332 KB)  

    Various existing criteria to characterize the statistical independence are applied in blind source separation and independent component analysis. However, almost all of them are based on parametric models. The distribution model mismatch between the output PDF (probability density functions) and the chosen underlying distribution model is a serious problem in blind signal processing. Nonparametric... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Model combination and weight selection criteria for speaker verification

    Publication Year: 1999 , Page(s): 439 - 448
    Cited by:  Papers (1)  |  Patents (1)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (428 KB)  

    We focus on the score combination for three separate modeling approaches as applied to text-dependent speaker verification. The modeling methods that are evaluated consist of the neural tree network (NTN), hidden Markov model (HMM), and dynamic time warping (DTW). One of the main challenges in combining scores of several models is how to select the weight for each model. One method is to use equal... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some analytical results on critic-driven ensemble classification

    Publication Year: 1999 , Page(s): 253 - 262
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (428 KB)  

    We (1999) proposed a framework for ensemble classification wherein auxiliary networks, dubbed critics, are used to provide reliability information on the ensemble's individual classifiers/experts. We showed experimentally that critic-driven combining schemes extend the applicability of ensemble methods by overcoming the usual requirement that the individual classifier error rate p must be less tha... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Approximation by random networks with bounded number of layers

    Publication Year: 1999 , Page(s): 166 - 175
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (396 KB)  

    This paper discusses the function approximation properties of the Gelenbe random neural network (GNN). We use an extension of the basic model: the bipolar GNN (BGNN). We limit the networks to being feedforward and consider the case where the number of hidden layers does not exceed the number of input layers. We show that the feedforward BGNN with s hidden layers (total of s+2 layers) can uniformly... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multi-channel piecewise selective averaging of cognitive evoked potentials with variable latency

    Publication Year: 1999 , Page(s): 459 - 467
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (520 KB)  

    This work is about the development of an alternative way of averaging evoked potentials (EP) of cognitive activities. Since the main assumption of invariant waveforms time locked to the eliciting events does not hold for cognitive EPs, averaging results in distorted estimates. The authors' alternative selective averaging finds similar subsequences of fixed length with variable latency which are co... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A suboptimal Bayesian equalizer using an nonlinear multilayer combiner

    Publication Year: 1999 , Page(s): 353 - 362
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (468 KB)  

    In order to reduce the complexity and enhance the performance of the Bayesian equalizer (REQ) using the radial basis function (RBF) network, a new equalizer (RNEQ) using the RBF network with a nonlinear multilayer combiner is proposed. The proposed RNEQ produces the output using nonlinear multilayer combiner. The RNEQ is applied to a digital communication system and a nonlinear magnetic storage sy... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Allpass vs. unit-norm constraints in contrast-based blind deconvolution

    Publication Year: 1999 , Page(s): 273 - 282
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (416 KB)  

    Many contrast-based independent component analysis (ICA) algorithms can be directly applied to single-channel blind deconvolution tasks without modification. For prewhitened signals, however, an allpass constraint on the filter impulse response is more appropriate than a unit-norm constraint. We show how a finite-impulse-response filter can be adapted to maximize or minimize an arbitrary contrast ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Hierarchy of probabilistic principal component subspaces for data mining

    Publication Year: 1999 , Page(s): 497 - 506
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (612 KB)  

    Visual exploration has proven to be a powerful tool for multivariate data mining. Most visualization algorithms aim to find a projection from the data space down to a visually perceivable rendering space. To reveal all of the interesting aspects of complex data sets living in a high-dimensional space, a hierarchical visualization algorithm is introduced which allows the complete data set to be vis... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neuro Bayesian blind equalization with BER estimation in digital channels

    Publication Year: 1999 , Page(s): 333 - 342
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (508 KB)  

    The implementation of an optimal Bayesian algorithm for digital equalization is infeasible due to its computational complexity. We present a new approach to Bayesian blind equalization which is based on a hybrid architecture involving neural networks and evolutionary computation concepts. We develop a theoretical analysis which leads to recursive formulas to estimate the probability density functi... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • SUPANOVA: a sparse, transparent modelling approach

    Publication Year: 1999 , Page(s): 21 - 30
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (456 KB)  

    Traditional neural networks produce opaque models that are difficult to interpret. This work describes a transparent, non-linear, modelling approach that enables the constructed models to be visualised, enhancing their validation and interpretation. The technique combines the representational advantage of a sparse ANOVA decomposition, with the good generalisation ability of a support vector machin... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A comparison among feature selection methods based on trained networks

    Publication Year: 1999 , Page(s): 205 - 214
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (480 KB)  

    We present a review of feature selection methods, based on the analysis of a trained multilayer feedforward network, which have been applied to neural networks. Furthermore, a methodology that allows evaluating and comparing feature selection methods is carefully described. This methodology is applied to the 19 reviewed methods in a total of 15 different real world classification problems. We pres... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On training piecewise linear networks

    Publication Year: 1999 , Page(s): 122 - 131
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (380 KB)  

    Piecewise-linear (PWL) neural networks are networks with piecewise-linear node functions. They have attractive features, such as speed of training and amenability to digital VISI implementation. The paper presents an algorithm for training PWL networks. The algorithm is general in that it can be used for the optimization of general PWL functions. It is based on moving from one linear region to the... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Margin-like quantities and generalized approximate cross validation for support vector machines

    Publication Year: 1999 , Page(s): 12 - 20
    Cited by:  Papers (4)  |  Patents (2)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (304 KB)  

    We examine support vector machines (SVM) from the point of view of solutions to variational problems in a reproducing kernel Hilbert space. We discuss the generalized comparative Kullback-Leibler distance as a target for choosing tuning parameters in SVMs, and we propose that the generalized approximate cross validation estimate of them is a reasonable proxy for this target. We indicate an interes... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Sequential support vector machines

    Publication Year: 1999 , Page(s): 31 - 40
    Cited by:  Papers (8)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (420 KB)  

    We derive an algorithm to train support vector machines sequentially. The algorithm makes use of the Kalman filter and is optimal in a minimum variance framework. It extends the support vector machine paradigm to applications involving real-time and non-stationary signal processing. It also provides a computationally efficient alternative to the problem of quadratic optimisation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Discrimination of cylinders with different wall thicknesses using neural networks and simulated dolphin sonar signals

    Publication Year: 1999 , Page(s): 477 - 486
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (500 KB)  

    This paper describes a method integrating neural networks into a system for recognizing underwater objects. The system is based on a combination of simulated dolphin sonar signals, simulated auditory filters and artificial neural networks. The system is tested on a cylinder wall thickness difference experiment and demonstrates high accuracy for small wall thickness differences. Results from the ex... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On optimal data split for generalization estimation and model selection

    Publication Year: 1999 , Page(s): 225 - 234
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (456 KB)  

    The paper is concerned with studying the very different behavior of the two data splits using hold-out cross-validation, K-fold cross-validation and randomized permutation cross-validation. First we describe the theoretical basics of various cross-validation techniques with the purpose of reliably estimating the generalization error and optimizing the model structure. The paper deals with the simp... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Natural power method for fast subspace tracking

    Publication Year: 1999 , Page(s): 176 - 185
    Cited by:  Papers (7)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (352 KB)  

    Elaborates on a natural version of the power method for fast estimation and tracking of principal subspace or/and principal components of a vector sequence. The natural power method has the fastest convergence rate among a class of power-based methods such as the Oja method, the projection approximation subspace tracking (PAST) method, and the novel information criterion (NIC) method. Like the abo... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Geometrical structures of FIR manifold and their application to multichannel blind deconvolution

    Publication Year: 1999 , Page(s): 303 - 312
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (424 KB)  

    We study geometrical structures on the manifold of FIR filters and their application to multichannel blind deconvolution. First we introduce the Lie group and Riemannian metric to the manifold of FIR filters. Then we derive the natural gradient on the manifold using the isometry of the Riemannian metric. Using the natural gradient, we present a novel learning algorithm for blind deconvolution base... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Clustering approach to square and non-square blind source separation

    Publication Year: 1999 , Page(s): 315 - 323
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (380 KB)  

    Recently, a number of heuristic techniques, mostly based on topographic maps, have been introduced in order to overcome (some) of the limitations of the blind source separation (BSS) algorithms that are rooted in the theory of independent component analysis. Here, we introduce a new heuristic that relies on the tendency of the mixture samples to cluster around the source directions in mixture spac... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Rate-constrained self-organizing neural maps and efficient psychovisual methods for low bit rate video coding

    Publication Year: 1999 , Page(s): 390 - 399
    Save to Project icon | Request Permissions | Click to expandAbstract | PDF file iconPDF (452 KB)  

    The video coding problem is essentially an operational distortion-rate issue where the underlying input pixel data, probability distributions and dimensions are discrete, unknown and not smooth. In the low bit rate case the high resolution assumptions for vector quantization are not strictly valid and the problem is exacerbated. However, by considering the rate-constrained operational points on se... View full abstract»

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