By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 9 • Date Sep 1990

Filter Results

Displaying Results 1 - 22 of 22
  • The segmental K-means algorithm for estimating parameters of hidden Markov models

    Page(s): 1639 - 1641
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (284 KB)  

    The authors discuss and document a parameter estimation algorithm for data sequence modeling involving hidden Markov models. The algorithm, called the segmental K-means method, uses the state-optimized joint likelihood for the observation data and the underlying Markovian state sequence as the objective function for estimation. The authors prove the convergence of the algorithm and compare it with the traditional Baum-Welch reestimation method. They also print out the increased flexibility this algorithm offers in the general speech modeling framework View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • FIR filtering by the modified Fermat number transform

    Page(s): 1641 - 1645
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (316 KB)  

    Right-angle circular convolution (RCC) and the modified Fermat number transform (MFNT) are introduced. It is shown that a linear convolution of two N-point sequences can be obtained by a corresponding N-point RCC. It is also shown that the MFNT supports RCC so that a linear convolution can be computed by an N -point MFNT and its inverse plus N multiplies View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Convergence behavior and N-roots of stack filters

    Page(s): 1529 - 1544
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1408 KB)  

    The convergence behavior of two types of stack filters is investigated. Both types are shown to possess the convergence property and to exhibit nontrivial behavior. The first type of stack filter has the erosive property; it erodes any input signal to a root after a sufficient number of passes. The second type of stack filter has the dilative property; it dilates any input signal to a root after a sufficient number of passes. For each type of stack filter, an algorithm is presented which can determine a filter that has any specific signal or set of signals as roots. These two algorithms are efficient in that their execution time is a linear function of the length of the input signal, the width of the filter window, and the number of signals to be preserved. Since some stack filters have the phenomenon of oscillations when they filter some input signals successively, a partial ordering is defined over the set of stack filters which makes it possible to determine upper and lower bounds for these oscillations View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A stability analysis of two-dimensional nonlinear digital state-space filters

    Page(s): 1578 - 1586
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (632 KB)  

    Two different methods for a stability analysis of 2-D nonlinear discrete state-space systems under zero input conditions are provided. The first method reduces the task of testing a 2-D discrete nonlinear or shift-varying system to a single 2-D linear stability test of a system matrix with nonnegative system matrix entries. The second method is based on a number of norm tests for products of extreme matrices, and can be considered the 2-D counterpart of the method for 1-D systems described by K.T. Erickson and A.N. Michel (1985). Both of the introduced methods are based on the sector description of the nonlinearity and can be used to analyze digital filter stability under finite-word-length effects View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Resolution threshold of beamspace MUSIC for two closely spaced emitters

    Page(s): 1545 - 1559
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (992 KB)  

    An analysis of a beamspace version of the MUSIC algorithm applicable to two closely spaced emitters in diverse scenarios is presented. Specifically, the analysis is applicable to uncorrelated far-field emitters of any relative power level, confined to a known plane, and observed by an arbitrary array of omnidirectional sensors. An expression for the threshold array signal-to-noise ratio at which beamspace MUSIC is able to resolve the emitters is obtained. The preprocessor that minimizes the resolution threshold is identified. It is demonstrated that the resolution threshold is proportional to the dimension of the noise subspace; therefore, the threshold can be reduced substantially by utilizing an appropriate beamformer to reduce the dimension of the noise subspace. It is also demonstrated that MUSIC in conjunction with a suitable preprocessor can provide a resolution threshold lower than conventional (sensor-space) MinNorm View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Synthesis of 2-D state-space digital filters with low sensitivity based on the Fornasini-Marchesini model

    Page(s): 1587 - 1594
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (520 KB)  

    The balanced realization and low-sensitivity structure of two-dimensional recursive digital filters are considered in the framework of the Fornasini-Marchesini local state-space model. A procedure is introduced for the balanced realization of 2-D recursive digital filters. The sensitivities of a 2-D transfer function are investigated with respect to the coefficients in the local state-space model. The overall sensitivity is evaluated using the 2-D controllability and observability Gramians. The filter structure reducing the overall sensitivity is synthesized for two cases: one free from l2 scaling constraints on the state variables and the other under the scaling constraints. These filter structures are shown to be closely related to the balanced realization. An example is given to illustrate the utility of the proposed technique View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analytical signal processing for pattern recognition

    Page(s): 1645 - 1649
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (328 KB)  

    The development of a novel technique for the alignment of signals distorted by linear or nonlinear perturbation of the sampling instants is described. This technique considers the analytic signals of the given signals and uses the fact that, if a signal is distorted by a linear time-scale distortion, the contour of the analytic signal in the complex plane is unaltered. It is found that the analytic signal contours of signals subject to nonlinear time-scale distortion also match very closely. Constraints are derived on the allowed distortion characteristics of the sample instants in order to obtain analytic contours which are essentially coincident. Analytic contours which contain loops are also considered, and methods of overcoming the problems of comparing dissimilar signals are described View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Huffman-type code generator with order-N complexity

    Page(s): 1619 - 1626
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (716 KB)  

    A new method with order-N complexity is presented to construct Huffman-type minimum-redundancy codes for N distinguished symbols in the source alphabets. This method includes a contraction process as well as an expansion process. The contraction process has (N-3) contraction stages. To reduce the data-transfer operations, (N-5) vacancies are reserved beforehand in the array which is used to store the probabilities of symbols. At each contraction stage, the number of the symbols whose probabilities are equal to or less than the sum of the two least probabilities is stored and named as the expansion index. Under the contraction process, the code lengths of the original symbols and those at each contraction stage increase monotonically. Since the proposed method gives a monotonically increasing code both in code length and in code value, its implementation becomes easy in VLSI technology, in microprocessor-based systems, or in software programming View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The Hermite transform-theory

    Page(s): 1595 - 1606
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (900 KB)  

    The author introduces a scheme for the local processing of visual information, called the Hermite transform. The problem is addressed from the point of view of image coding, and therefore the scheme is presented as an analysis/resynthesis system. The objectives of the present work, however, are not restricted to coding. The analysis part is designed so that it can also serve applications in the area of computer vision. Indeed, derivatives of Gaussians, which have found widespread application in feature detection over the past few years, play a central role in the Hermite analysis. It is also argued that the proposed processing scheme is in close agreement with current insight into the image processing that is carried out by the human visual system. In particular, it is demonstrated that the Hermite transform is in better agreement with human visual modeling than Gabor expansions View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dependence of processing gain and interference immunity of adaptive narrow-band beam-formers on angular locations of interference sources

    Page(s): 1633 - 1637
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (332 KB)  

    The dependence of processing gain and interference immunity of adaptive narrowband linear beamformers on angular locations of single, double, and L interference sources is investigated. Corresponding expressions of these figures of merit are derived. It is found that different angular locations may correspond to the same processing gain and interference immunity. The former varies from unity to a maximum value which under certain conditions is independent of both the array geometry and the number of interference sources. The null directions of the quiescent beam correspond to unity processing gain and 100% interference immunity. However, there are always other directions where the processing gain is unity but the interference immunity is extremely small. The maximum processing gain is always associated with 50% immunity View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Efficient sinc function interpolation technique for center padded data

    Page(s): 1512 - 1517
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (764 KB)  

    An efficient zooming FFT (fast Fourier transform) algorithm that allows center padding sinc function interpolation of 2-D images is presented. This algorithm avoids the phase shifts that would be introduced if the efficient Skinner interpolation method is used. Output pruning is incorporated to allow efficient determination of a zoomed subimage. Time savings of more than 50% can be achieved. Example images illustrating the use of the algorithm in conjunction with zooming and ARMA (autoregressive moving average) modeling of data are given View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A note on the third-order intermodulation due to quantization

    Page(s): 1627 - 1628
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (144 KB)  

    A simple closed-form approximation is derived for the amplitude of the most disturbing third-order intermodulation product resulting from the quantization of one large and one weak sinusoid. From this expression, results can be obtained by hand computation which, otherwise, can be only obtained by computer simulation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design and performance of an analysis-by-synthesis class of predictive speech coders

    Page(s): 1489 - 1503
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1444 KB)  

    The performance of a broad class of analysis-by-synthesis linear predictive speech coders is quantified experimentally. The class of coders includes a number of well-known techniques as well as a very large number of speech coders which have not been named or studied. A general formulation for deriving the parametric representation used in all of the coders in the class is presented. A new coder, named the self-excited vocoder, is discussed because of its good performance with low complexity, and because of the insight this coder gives to analysis-by-synthesis coders in general. The results of a study comparing the performances of different members of this class are presented. The study takes the form of a series of formal subjective and objective speech quality tests performed on selected coders. The results of this study lead to some interesting and important observations concerning the controlling parameters for analysis-by-synthesis speech coders View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the measure of the set of factorizable polynomial bispectra

    Page(s): 1637 - 1639
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (244 KB)  

    In a recent work (1989), the authors have shown that factorization of bispectrum is not always possible. In the present work, they show that the subset of factorizable bispectra has Lebesgue measure zero in the set of polynomial bispectra, i.e. those that are obtained from finite-support bicumulants. Hence, a polynomial bispectrum cannot almost always be exactly realized as that of the output of a linear model driven by a third-order white input. This result can be generalized to multidimensional polynomial bispectra. Although it follows that a linear model driven by a third-order white input cannot almost always realize a given bispectrum or a bicumulant sequence, the use of a linear model as an approximation in certain applications can be justified if the computed bispectrum has an index value close to unity View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Efficient address generation for prime factor algorithms [digital signal processing]

    Page(s): 1518 - 1528
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (876 KB)  

    An attempt is made to explain as clearly as possible the problem of address generation and how the prime factor mapping technique is used in this class of algorithms. Two novel address generation schemes are proposed to improve efficiency. The first scheme reduces the computation required for unscrambling data in an in-place realization of the PFA (prime factor algorithm) by reducing the number of variables used to calculate the data addresses. The second scheme is to be used in an in-place in-order realization of PFA. It achieves high efficiency by replacing complicated modulo operations of conventional approaches by simple indirect addressing techniques. Making use of this scheme, software packages have been written for the computation of DFTs (discrete Fourier transforms) using a high-level language and two low-level languages (the 80286/287 and TMS330C25 assembly languages). Results of these realizations show that a reduction of 50% in address generation time is achievable, giving a saving of 30% in total computation time. A hardware address generator is also developed, which may provide clues to improving digital signal processor architectures in the future View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Sequential algorithms for parameter estimation based on the Kullback-Leibler information measure

    Page(s): 1652 - 1654
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (256 KB)  

    Methods of stochastic approximation are used to convert iterative algorithms for maximizing the Kullback-Leibler information measure into sequential algorithms. Special attention is given to the case of incomplete data, and several algorithms are presented to deal with situations of this kind. The application of these algorithms to the identification of finite impulse response systems is considered View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • New insights into the high-order Yule-Walker equations

    Page(s): 1649 - 1651
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (244 KB)  

    Some new analytical results on the high-order Yule-Walker (HOYW) equations are presented. A simple singularity condition for the high-order autocorrelation matrix is found which makes it possible to determine situations where application of the HOYW equations could be especially unsuitable. It is also shown that using the HOYW equations is equivalent to performing conventional noise compensation in the main diagonal of the low-order autocorrelation matrix. This is in connection with the possibility of matching a given set of (the first) 2M+1 autocorrelation estimates to a noisy autoregressive process of order M View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Array based design of digital filters

    Page(s): 1628 - 1632
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (368 KB)  

    It is noted that signal processing designs for real-time large-scale systems are increasingly confronted with two conflicting objectives. The traditional objective of optimal design in low signal-to-noise ratio environments is confronted with the need for simplicity in implementation and speed of computation. The inclusion of high throughput and efficient hardware utilization as constraints on digital filter designs is considered. In particular, implementation of the design via an array processor is introduced. The concept of fast processing becomes synonymous with high throughout and efficient implementation on such a device. Using an array interpretation of the FFT structure, the retention of this highly efficient structure in a general design setting is demonstrated. For a typical signal extraction design, a constrained least-squares minimization is introduced to determine optimal enhancing filters with highly efficient array implementation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Algorithms meeting the lower bounds on the multiplicative complexity of length-2n DFTs and their connection with practical algorithms

    Page(s): 1504 - 1511
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (588 KB)  

    In a previous work (see Electron. Lett., vol.20, no.17, p.690, 1984), the author described an algorithm that computes a length-2n discrete Fourier transform using 22+1-2n2+4n-8 nontrivial (i.e. ≠±j=±√-1) complex multiplications. In the present work, it is first shown that this algorithm actually provides the attainable lower bound on the number of complex multiplications. A slight modification of the last step of this algorithm is also shown to provide the attainable lower bound on the number of real multiplications. A connection with the split-radix FFT algorithm (SRFFT) is then explained, showing that SRFFT is another variation of these optimal algorithms, where the last step is computed recursively from shorter FFTs in a suboptimal manner. Finally, once the connection between the minimal complexity and SRFFT (which is the best known practical algorithm) is understood, it provides useful information on the possibility of further improvements of the SRFFT View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The Hermite transform-applications

    Page(s): 1607 - 1618
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1380 KB)  

    It is demonstrated how the Hermite transform can be used for image coding and analysis. Hierarchical coding structures based on increasingly specified basic patterns, i.e. general 2-D patterns, general 1-D patterns, and specific 1-D patterns such as edges and corners, are presented. In the image coding application, the relation with existing pyramid coders is described. A new coding scheme, based on local one-dimensional image approximations, is introduced. In the image analysis application, the relation between the Hermite transform and existing line/edge detection schemes is described. It is shown that, by concentrating on more specific patterns, the coding efficiency can be increased since fewer coefficients have to be coded. Meanwhile, sufficient descriptive power can be maintained for approximating the most interesting features in natural images View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Maximum-likelihood narrow-band direction finding and the EM algorithm

    Page(s): 1560 - 1577
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1332 KB)  

    Generalized EM (expectation-maximization) algorithms have been derived for the maximum-likelihood estimation of the direction-of-arrival of multiple narrowband signals in noise. Both deterministic and stochastic signal models are considered. The algorithm for the deterministic model yields estimates of the signal amplitudes, while that for the stochastic model yields estimates of the powers of the signal. Both algorithms have the properties that their limit points are stable and satisfy the necessary maximizer conditions for maximum-likelihood estimators. It is shown via simulation that the maximum-likelihood method allows for the resolution of the directions-of-arrival of signals at angular separation and noise levels for which other high-resolution methods will not work. Algorithm convergence does depend on initial conditions; however, convergence to a global maximum has been observed in simulation when the initial estimates are within a significant fraction if one beamwidth (componentwise) of this maximum. Simulations also show that the deterministic model has a significant impact on the angle estimator performance View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Weight adjustment rule of neural networks for computing discrete 2-D Gabor transforms [image processing]

    Page(s): 1654 - 1656
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (192 KB)  

    It is demonstrated that the weight adjustment rule used in the neural network for computing the 2-D Gabor transform proposed by J. Daugman (1988) can be shown to be equivalent to the Jacobi iteration scheme for solving simultaneous linear equations. It is shown that faster convergence, of the algorithm can be achieved by using Gauss-Seidel iteration, successive overrelaxation, conjugate gradient algorithms, and multigrid methods View full abstract»

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

Aims & Scope

This Transactions ceased production in 1990. The current retitled publication is IEEE Transactions on Signal Processing.

Full Aims & Scope