Showing papers in "Signal Processing in 1995"

TL;DR: A new separation scheme whose convergence is proved analytically is proposed, based on the observation that it is possible to extract one of the source signals by a simple algorithm obtained by extending to the source separation context some of the ideas developed by Shalvi-Weinstein in the framework of blind deconvolution.

TL;DR: It is found that at high and medium signal-to-noise ratio, the OPM performs quite like MUSIC with a complexity reduced by the ratio of the number of sources to theNumber of sensors.

TL;DR: This implementation yields an infinite impulse response filter that has six MADDs per dimension independent of the value of σ in the Gaussian kernel.

TL;DR: It is proved in this paper that the estimation of filter coefficients may be based on the cancellation of 4th-order output cross-cumulants, and that the FIR model of mixtures is not realistic enough and must be improved.

TL;DR: This paper introduces the general topic of optical character recognition (OCR), and introduces a five stage model for AOTR systems and classify research work according to this model, and presents an historical review of the Arabic text recognition systems.

TL;DR: A set of decision criteria for identifying different types of digital modulation is developed and it is found that all modulation types of interest have been classified with success rate ≥90% at SNR = 10 dB.

TL;DR: From the simulation results, it is observed that the performance of the proposed FLANN based equalizer outperforms the other two in terms of bit-error rate (BER) and attainable MSB level over a wide range of eigenvalue spread, signal to noise ratio and channel nonlinearities.

TL;DR: The focus of the research is on the combination, by means of the sup- and inf-operations, of alternating filters by reconstruction when their component filters belong to a granulometry and an antigranulometry (by reconstruction).

TL;DR: A global procedure for recognition of analogue modulation types is developed and it is found that all types of analog modulation have been classified with success rate ⩾ 90% at SNR = 10 dB.

TL;DR: A continuum of “fractional” domains making arbitrary angles with the time and frequency domains is considered, derived by the fractional Fourier transform, to derive transformation, commutation, and uncertainty relations among coordinate multiplication, differentiation, translation, and phase shift operators between domains making arbitrarily angles with each other.

TL;DR: The processing algorithm developed in this paper does not need a potentially costly polarization filter bank to cover the unknown target polarization parameter and is compared with other algorithms that do not utilize the polarization information or assume that the target polarization is known.

TL;DR: The results show the gain in performance if an MMD filter is used instead of a Volterra filter, due to the fact that with a given computational power longer memory lengths can be achieved by the MMD model.

TL;DR: The asymptotic covariance matrix and bias of the estimators of dsω(t)/dts, s = 0,1,2,…, m − 1, are obtained for the frequency with bounded m-derivative and the estimator is shown to be strongly consistent and Gaussian for a polynomial frequency.

TL;DR: Two image transforms are presented: one performs an optimal enhancement and stabilization of the overall dynamic range, and the other does of the mean dynamic range.

TL;DR: A new method is proposed for modelling state duration in hidden Markov model (HMM) speech recognition systems that improves the recognition accuracy by 5.6%.

TL;DR: A general approach for deriving 3D invariants that can be used as input to a statistical classifier, such as a k-nearest-neighbours algorithm or a neural network.

TL;DR: An analysis of ambiguity-domain weighting functions for comparing the interference attenuation and time-frequency concentration properties of the three TFRs shows important effects and performance limitations whose understanding is essential for a practical application of the TFR's.

TL;DR: It is demonstrated that the various neural networks based on radial basis functions can be very useful for interference cancellation problems in which traditional linear cancelers may fail badly.

TL;DR: This paper explores the nonstationarity of Markov chains and proposes a nonstationary HMM that is defined with a set of dynamic transition probability parameters, which is more capable of capturing the dynamic nature of signals with higher discrimination power in on-line character recognition.

TL;DR: A multiscale vision model based on the wavelet transform is introduced which furnishes an isotropic vision, with a unique wavelet function, which gives correct restored images without or with adding a Gaussian noise.

TL;DR: Input/output relations in terms of cyclic higher-order statistics for multi-input multi-output linear almost-periodically time-variant systems that are excited by cyclostationary inputs are derived.

TL;DR: A new method is proposed for the separation of mixed digital sources, based on geometrical considerations, which is applied to the separated of binary and n-valued sources.

TL;DR: The methods of estimating fractal and multifractal dimensions of gray-tone digital images of texture images from the Brodatz album are dealt with.

TL;DR: A symmetric wideband ambiguity function is proposed, by recognizing the relationship between ambiguity functions and the theory of group representations, and the affine Wigner distribution is recovered, a Mellin-frequency/time-scale phase space function with properties that parallel those of the Wigners.

TL;DR: Some extensions of the RBF models, which are derived from the general expression for a nonlinear stochastic model, suggest new mechanisms of production of complex dynamical features such as intermittency and chaos.

TL;DR: It is shown that the DOAs of two uncorrelated signals can be uniquely determined with one electromagnetic vector sensor, regardless of the states of polarization, and the relationship between uniqueness in measurements and uniqueness in MUSIC estimates is established.

TL;DR: New decompositional methods based on Fibonacci p-codes for computing the output for different stack filters are presented and the computational complexities of these methods are studied.

TL;DR: Asymptotic analysis of constrained and unconstrained MUSIC demonstrates that (asymptotically) improved subspace estimates always result from the use of constraints, and the variance of constrained MUSIC is less than that of unconstrains under either high coherence, large numbers of sensors, or high SNR conditions.

TL;DR: Frequency response analysis reveals that the LMS algorithm proposed for a single-frequency case possesses the characteristics of an IIR notch filter.

TL;DR: It is shown that the texture energy method can be understood as the equivalent of an envelope detector receiver commonly used in AM communication techniques and a further linear relationship between the local variances of the filter outputs and the autocorrelation function of the texture process is derived.