IEEE Workshop on Statistical Signal Processing, 2003

28 Sept.-1 Oct. 2003

Filter Results

Displaying Results 1 - 25 of 178
  • A multiresolution approach to pattern recognition

    Publication Year: 2003, Page(s):417 - 420
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1453 KB) | HTML iconHTML

    This paper reports on a family of computationally practical classifiers called dyadic classification trees (DCTs). Like many multiresolution methods in other application areas, DCTs are formed by recursive dyadic partitioning of the input space, followed by pruning to avoid overfitting. We investigate three pruning rules, each motivated by statistical learning theory. These pruning rules involve p... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive nonlinear multigrid inversion with applications to Bayesian optical diffusion tomography

    Publication Year: 2003, Page(s):170 - 173
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1450 KB) | HTML iconHTML

    We previously proposed a general framework for nonlinear multi-grid inversion applicable to any inverse problem in which the forward model can be naturally represented at differing resolutions. The method has the potential for very large computational savings and robust convergence. In this paper, multigrid inversion is further extended to adaptively allocate computation to the scale at which the ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast adaptive Bayesian beamforming using the FFT

    Publication Year: 2003, Page(s):413 - 416
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1391 KB) | HTML iconHTML

    A fast algorithm is developed to implement a Bayesian beam-former that can estimate signals of unknown direction of arrival (DOA). In the Bayesian approach, the underlying DOA is assumed random and its a posteriori probability density function (PDF) is approximated by a discrete probability mass function. A Bayesian beamformer then balances a set of beamformers according to the associated weights.... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Levelset based segmentation using data driven shape prior on feature histograms

    Publication Year: 2003, Page(s):166 - 169
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1609 KB) | HTML iconHTML

    We develop a new data driven shape prior for use in image segmentation. This prior is designed to penalize the differences between the distributions of contour features obtained from training data shapes and those of a segmenting curve. We incorporate this prior into a level set segmentation framework and present an efficient method for its implementation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the accuracy of spectral density estimation

    Publication Year: 2003, Page(s):125 - 128
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1400 KB) | HTML iconHTML

    Recently, we proposed a method for estimating the distribution of a kernel spectral density estimator using the bootstrap. In this contribution, we address the issue of accuracy of the estimation. It is well known that the bootstrap is second order accurate. This holds whenever a bootstrap version of the population-sample relationship is constructed in such a way that each component that depends o... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Detection of an electron spin in a MRFM cantilever experiment

    Publication Year: 2003, Page(s):633 - 636
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1421 KB) | HTML iconHTML

    In the magnetic resonance force microscopy (MRFM) experiment that we consider, a cantilever is used to detect the presence of an electron spin. If an electron spin is present, the equation governing the motion of the cantilever is nonlinear. In the absence of a spin, the equation is linear. We propose a detector model that uses two Kalman filter (KF) state estimators. The estimator for the spin hy... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Efficient algorithms for inferences on Grassmann manifolds

    Publication Year: 2003, Page(s):315 - 318
    Cited by:  Papers (41)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1459 KB) | HTML iconHTML

    Linear representations and linear dimension reduction techniques are very common in signal and image processing. Many such applications reduce to solving problems of stochastic optimizations or statistical inferences on the set of all subspaces, i.e. a Grassmann manifold. Central to solving them is the computation of an "exponential" map (for constructing geodesies) and its inverse on a Grassmanni... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Joint estimation of image and deformation parameters in tomographic image reconstruction

    Publication Year: 2003, Page(s):162 - 165
    Cited by:  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1397 KB) | HTML iconHTML

    We consider an emission tomography reconstruction problem in which projection measurements from several successive time frames are available. Two strategies for doing motion-corrected image reconstruction are compared. In the first strategy, separate images are reconstructed from the measurements at each time frame. They are then consolidated by post-registration and averaging procedures. In the s... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Asymptotic efficiency of the BAD algorithm

    Publication Year: 2003, Page(s):86 - 89
    Cited by:  Papers (3)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1433 KB) | HTML iconHTML

    We study the high-SNR performance of the bi-directional arbitrated DFE (BAD) equalization algorithm. Using asymptotic efficiency analysis, we characterize the performance of BAD as the channel noise variance approaches zero. We consider the case in which the channel is perfectly known to the receiver, as well as the case in which the channel estimate is in error. Asymptotic efficiency analysis all... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bayesian inference of network loss characteristics with applications to TCP performance prediction

    Publication Year: 2003, Page(s):530 - 533
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1410 KB) | HTML iconHTML

    Network tomography, inferring internal network behavior based on the "external" end-to-end network measurements, is of particular interest when the network itself can not cooperate in characterizing its own behavior. In particular, it is impractical to directly measure packet losses or delays at every router. On the other hand, measuring end-to-end (from sources to receivers) losses is relatively ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimal sampling strategies for multiscale models with application to network traffic estimation

    Publication Year: 2003, Page(s):138 - 141
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1418 KB) | HTML iconHTML

    The paper considers the problem of determining which set of 2p leaf nodes on a binary multiscale tree model of depth N (N < p) gives the best linear minimum mean-squared estimator of the tree root. We find that the best-case and worst-case sampling choices depend on the correlation structure of the tree. This problem arises in Internet traffic estimation, where the goal is to estimate t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Hidden Markov chain identification

    Publication Year: 2003, Page(s):122 - 124
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1350 KB) | HTML iconHTML

    An identification scheme is developed for hidden Markov models (HMM). Unlike the realization problem, where one starts from exact probabilities, the identification problem makes a statistical inference from the pathwise output sequences. The basic principle in the identification of Markovian finite state systems from nonnumeric inputs and outputs is its lifting to a numerical representation via th... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Scheduling multiple sensors using particle filters in target tracking

    Publication Year: 2003, Page(s):549 - 552
    Cited by:  Papers (15)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1408 KB) | HTML iconHTML

    A critical component of a multi-sensor system is sensor scheduling to optimize system performance under constraints (e.g. power, bandwidth, and computation). In this paper, we apply particle filter sequential Monte Carlo methods to implement multiple sensor scheduling for target tracking. Under the constraint that only one sensor can be used at each time step, we select a sequence of sensor uses t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Parallel wireless mobile relays with space-time modulations

    Publication Year: 2003, Page(s):375 - 378
    Cited by:  Papers (30)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1416 KB) | HTML iconHTML

    We use Hurwitz-Radon matrices to construct a space-time code for parallel wireless mobile relays that are located anywhere between a source and a destination. Each relay receives a sequence of symbols (i.e., baseband signals corrupted by fading and noise) simultaneously from the source. These symbols are not decoded into information bits at the relays but are rearranged (i.e., space-time modulated... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bayesian learning for models of human speech perception

    Publication Year: 2003, Page(s):408 - 411
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1479 KB) | HTML iconHTML

    Human speech recognition error rates are 30 times lower than machine error rates. Psychophysical experiments have pinpointed a number of specific human behaviors that may contribute to accurate speech recognition, but previous attempts to incorporate such behaviors into automatic speech recognition have often failed because the resulting models could not be easily trained from data. This paper des... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Blind adaptive multiuser detection for uplink long-code CDMA systems

    Publication Year: 2003, Page(s):343 - 346
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1404 KB) | HTML iconHTML

    In the uplink of a long-code CDMA system, base station knows spreading codes of all serviced users. Given propagation delays, a blind adaptive CMA-based approach, with the aid of a set of MMSE-like constraints parameterized by channel-like vectors, is proposed to detect all users' symbols simultaneously. As by-products, the channel-like vectors are also obtained for all users. Since the constraint... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Equivalence of linear approaches in bioelectromagnetic inverse solutions

    Publication Year: 2003, Page(s):294 - 297
    Cited by:  Papers (10)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1444 KB) | HTML iconHTML

    Several different techniques have appeared in the magnetoencephalography (MEG) literature for solving the ambiguous inverse problem. We show that if the same second order statistics are uniformly applied in each case, then the resulting solutions are the same. Additionally, for the common assumption of independent sources, then the "noise significance " measure emphasises the need to select the da... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Theory and applications of orthogonal polynomials for Gaussian input

    Publication Year: 2003, Page(s):106 - 109
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1461 KB) | HTML iconHTML

    Power amplifiers are the major source of nonlinearity in communications systems. Such nonlinearity causes adjacent channel interference and increases bit error rate. Polynomials are often used to model the nonlinear power amplifier or its predistortion linearizer. In this paper, we present a novel set of orthogonal polynomials for baseband Gaussian inputs to replace the conventional polynomials, a... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Axial representations of 3D shapes

    Publication Year: 2003, Page(s):311 - 314
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1421 KB) | HTML iconHTML

    Many shapes can be described very well in terms of an axis. Examples are the stick figure drawings of man and animals from ancient times. We would like to give a natural (intuitive) definition of axis for cylindrical shapes. For 2D shapes, the medial axis is a good candidate, though it is sensitive to perturbations. For 3D shapes, the medial scaffold is often complicated surfaces and does not matc... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On capturing likelihood disparity for unsupervised image segmentation

    Publication Year: 2003, Page(s):158 - 161
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1568 KB) | HTML iconHTML

    In this paper, we study unsupervised Bayesian image segmentation approach which involves capturing model likelihood disparities among different texture features with respect to a global statistical model. Specifically, wavelet-domain hidden Markov models are used to characterize the global textural behavior of images in the wavelet-domain. Three clustering methods, i.e., the K-mean, a soft cluster... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Superimposed training for doubly selective channels

    Publication Year: 2003, Page(s):82 - 85
    Cited by:  Papers (3)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1480 KB) | HTML iconHTML

    We adopt a basis expansion approach to model the time and frequency (i.e., doubly) selective fading channel and employ a superimposed training technique to acquire the channel state information. Specifically, we add a known periodic sequence onto, instead of multiplexing it into, the information sequence. There is no reduction in transmission rate, and it is possible to estimate the time-varying c... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A multiphase level set method for tomographic reconstruction of dynamic objects

    Publication Year: 2003, Page(s):182 - 185
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1548 KB) | HTML iconHTML

    In this paper we present a novel variational method for tomographic reconstruction of dynamic objects with a multiphase level set method. Our method can reconstruct an image sequence composed of multiple regions from an extremely sparse and noisy projection data set. The number of regions in a scene can also change over time in our model. Temporal dynamics for region boundary and intensities are m... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Energy-based source localization via ad-hoc acoustic sensor network

    Publication Year: 2003, Page(s):387 - 390
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1420 KB) | HTML iconHTML

    In this paper, we investigate the performance characteristics of ad-hoc networks for acoustic sensors for source localization and design resource efficient signal processing algorithms that trade-off performance and complexity. We discuss the performance of location estimators that can be implemented in a distributed setting and compare their performance to the associated Cramer-Rao bound (CRB). A... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Causal and stable FIR-IIR Wiener filters

    Publication Year: 2003, Page(s):118 - 121
    Cited by:  Papers (4)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1389 KB) | HTML iconHTML

    IIR Wiener filters are the optimum filters with minimum MSE. In general, these filters are noncausal and their causal counterparts have larger MSE. In this paper, we propose causal and stable FIR-IIR Wiener filters which have the same MSE performance when the FIR part is sufficiently long. FIR-IIR Wiener filter employs the best delay LS inverse filter to model the noncausal part of the IIR Wiener ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The embedded triangles algorithm for distributed estimation in sensor networks

    Publication Year: 2003, Page(s):371 - 374
    Cited by:  Papers (6)
    Request permission for commercial reuse | Click to expandAbstract |PDF file iconPDF (1464 KB) | HTML iconHTML

    We propose a new iterative distributed estimation algorithm for Gaussian hidden Markov graphical models with loops. We decompose a loopy graph into a number of linked embedded triangles and then apply a parallel block-Jacobi iteration comprising local linear minimum mean-square-error estimation on each triangle (involving a simple 3x3 matrix inverse computation) followed by an information exchange... View full abstract»

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