Showing papers in "Signal Processing in 1989"

TL;DR: Some classical results about error bounds in classification and feature selection for pattern recognition are recalled, which are obtained with the aid of properties of distance measures.

TL;DR: A simple and reliable histogram-based method to estimate mutual information and the accuracies of this mutual information estimator and of a similar entropy estimator are discussed.

TL;DR: Various graphs that can be defined on a given set S of objects, depending on the intensity of the desirable neighborhood relationships are presented, and the main operators of Mathematical Morphology are defined.

TL;DR: In this article, the authors investigated the problem of sub-band coding of images and showed that extending the signal so that continuity is maintained may be a sufficient remedy, as shown in an image coding example.

TL;DR: An “extended invariance principle” is derived which is valid under the conditions describing that situation and which reduces to the classical IP in the special conditions under which the latter is known to hold.

TL;DR: In this paper, a differential motion estimation technique for television sequences is presented which measures the parameters of global motion in the image plane due to zoom and pan of the camera using a two-dimensional signal model and a three-parameter model for motion description.

TL;DR: This paper proposes the use of a representation for a stochastic linear system from the point of view of a Stochastic SVD and shows its relationship to the classical Karhunen-Loeve expansion.

TL;DR: Using the notion of homotopy, several very fast and efficient algorithms for constructing the skeleton are derived and the resulting skeleton is a good candidate for being the correct discretization of the skeletons in the continuous plane.

TL;DR: In this paper, it was proved that A is a complete lattice with respect to the activity ordering, according to which a mapping is more active than another when it modifies more elements of E. In this lattice, the infinum ⋏ is nothing but the already known morphological center.

TL;DR: In this article, the envelope characteristics of an unknown signal in noise have been used to identify the modulation type of the unknown signal by its envelope characteristics, such as the ratio (R ) of the variance of the envelope to the square of the mean of the signal.

TL;DR: An adaptive noise cancelling structure suitable for situations where the noise reference transducer is closely spaced relative to the primary transducers, based on two LMS delay line cancellers with cross-coupled feedback is described.

TL;DR: With three specified construction procedures and based on few starting arrays, many new perfect binary arrays with increasing dimensions and sizes can be synthesized.

TL;DR: The algorithm is based on the kernels used in the non-parametric estimation of probability density and regression functions and possesses tracking properties as the sample size grows large and the conditions for the mean square error convergence and the almost sure convergence are given.

TL;DR: In this paper, a sampling theorem for entire functions of exponential type, bounded on the real line, is established, and the sampling points have to fulfill certain periodicity conditions, but they may even lie arbitrarily close together.

TL;DR: The main properties and the main distribution functions of the models (e.g. statistical behavior of the RF under the non-linear operations provided by dilation or erosion) are presented and illustrate the MM approach for random structure modeling.

TL;DR: A geometrical approach based on mathematical morphology is proposed to remove speckle noise in coherent imagery, and two simple sequences based on the general theory of Alternating Sequential Filters are presented.

TL;DR: In this article, the effect of increasing the number of Yule-Walker equations on estimation accuracy is analyzed and an explicit expression for the asymptotic covariance matrix of the estimation errors is derived.

TL;DR: It was shown that very deep compression of one of the images of a stereo pair does not interfere with the perception of depth in the stereo image pair.

TL;DR: Approximate Gaussian filtering of equidistant data can be obtained by regularizing the data with Tikhonov's second order stabilizing functionals with good performance compared with that of single- and multi-stage FIR Gaussian filter approximations.

TL;DR: This class of nonlinear digital smoothing filters is shown to exhibit similar type root signal properties as the standard median filter, and the noise attenuation is optimal for certain types of exponential noise distributions.

TL;DR: After structuring the set of the morphological transformations, it is shown how to automatically choose the components of the high-level primitives and their chaining by means of geometric reasoning.

TL;DR: A version of this type of algorithm based on the orthogonal triangular factorization technique known as the QR decomposition is presented and its computational complexity and numerical properties are briefly discussed.

TL;DR: This paper considers the reconstruction of support-limited signals from noisy samples of their Fourier Transform and Iterative algorithms devised to solve these problems by using a convex programming method.

TL;DR: The steepest descent and conjugate gradient methods have been used for finding the first few large/small eigenvalues and eigenvectors of a Hermitian operator.

TL;DR: In this article, the authors considered spectral estimation via an L 1 solution to a set of overdetermined linear prediction equations (1-D) for Gaussian and non-Gaussian noise.

TL;DR: In this paper, an alternative technique for 2-D variable digital filters is proposed, where the filter coefficients are expressed as analytic functions with their independent variables being the frequency specifications of interest.

TL;DR: The in-place growing FMH filter based trend detector was found to have significant advantages over the FMH based methods and can be successfully applied to the filtering of the beginning and end of a finite length data sequence.

TL;DR: This paper revises the basic properties of the MLMS algorithm for stationary inputs, and it is shown that, under slow convergence conditions, this new algorithm is equivalent to the usual LMS algorithm, but it outperforms the L MS algorithm for fast convergence cases and for inputs containing inpulsive noise components.

TL;DR: A fresh approach to the problem of robust object recognition, where the edges are not the fundamental picture primitives, but rather the local axes of mirror symmetry (or local orientation) in polar coordinates, which can perform reliable object recognition in complex scenes.