Showing papers by "Ioannis Pitas published in 1993"

Digital Image Processing Algorithms

Abstract: Digital image processing fundamentals digital image transfor algorithms digital image filtering digital image compression edge detection algorithms image segmentation algorithms shape description.

Parallel Algorithms for Digital Image Processing, Computer Vision and Neural Networks

Abstract: Low Level Parallel Image Processing Parallel FFT-like Transform Algorithms on Transputers Parallel Edge Detection and Related Algorithms Parallel Segmentation Algorithms MIMD and SIMD Parallel Range Data Segmentation Parallel Stereo and Motion Estimation Parallel Implementations of the Backpropagation Learning Algorithm Based on Network Topology Parallel Neural Computation Based on Algebraic Partitioning Parallel Neural Computing Based on Network Duplicating PARALLEL EIKONA: A Parallel Digital Image Processing Package Parallel Architectures and Algorithms for Real Time Computer Vision Index.

Knowledge-based image analysis for geophysical interpretation

Abstract: Geophysical seismic interpretation is part of geophysical oil prospecting. It evaluates and analyses seismic reflection data, aiming at the detection of the position of hydrocarbon reservoirs. This paper provides a review of current efforts to automate, at least partially, seismic interpretation. As will be shown, this research area is very active and is a melting pot of various different approaches and techniques: artificial intelligence, pattern recognition, image processing, graphics, fuzzy set theory and, of course, geophysics and geology. Some methods of seismic pattern recognition (e.g. remote correlation, fuzzy seismic modeling, recognition of reservoir boundaries) and of seismic image processing (horizon following, texture analysis) are presented and some applications are shown. Expert systems used in geophysical interpretation (mainly in well log interpretation) are also briefly described. Finally, an automated system for knowledge-based image analysis for geophysical interpretation is dicussed. Its low-level vision techniques, its knowledge representation, and the control strategy for seismic pattern search are described.

Application of adaptive order statistic filters in digital image/image sequence filtering

Abstract: An application of adaptive order statistic filters in digital image filtering and in image sequence filtering is presented. The use of optimal minimum mean-square error (MMSE) and adaptive L-filters and L1-filters is examined in detail. It is proven that these filters adapt fairly well and remove effectively noise having various probability distributions (e.g. uniform, Gaussian, Laplacian, impulsive). >

Voronoi tessellation and Delauney triangulation using Euclidean disk growing in Z/sup 2/

Abstract: A novel approximation of Euclidean distance in Z/sup 2/ is proposed, and a novel algorithm for the computation of Voronoi tessellation and Delauney triangulation is presented based on this approximation. The proposed method has low computational complexity (of order O(1/N)) and allows parallel implementation. Mathematical morphology is used to implement the Voronoi tessellation and the Delauney triangulation. >

Multichannel L-filters based on M-ordering

Abstract: The extension of single-channel nonlinear filters whose output is a linear combination of the order statistics of the input samples to the multichannel case is presented in this paper. The subordering principle of marginal ordering (M-ordering) is used for multivariate data or dering. A unified framework for a discrete calculation of the moments of the bivariate order statistics required for the design of the multichannel marginal L-filters is outlined. The derivation of a bivariate distribution, namely the Laplacian (bi-exponential) distribution which belongs to Morgenstern's family is discussed. It is shown by simulations that the proposed multichannel L-filters perform better than other multi channel nonlinear filters such as the marginal median and the vector median proposed elsewhere as well as their single-channel counterparts.

A variant of learning vector quantizer based on the L/sub 2/ mean for segmentation of ultrasonic images

Abstract: The segmentation of ultrasonic images using self-organizing neural networks (NN) is investigated. A modification of learning vector quantizer (called L/sub 2/ LVQ) is proposed so that the weight vectors of the output neurons correspond to the L/sub 2/ mean instead of the sample arithmetic mean of the input observations. The convergence in the mean and in the mean square of the proposed variant of LVQ is studied. Experimental results show that L/sub 2/ LVQ outperforms other segmentation techniques that employ thresholding a filtered ultrasonic image with respect to the probability of detection for the same probability of false alarm in all cases. >

State of the art of morphological and nonlinear digital image processing

Abstract: A multitude of nonlinear digital image processing techniques has appeared in the literature. The following classes of nonlinear digital image/signal processing techniques can be identified at present: (1) order statistic filters; (2) homomorphic filters; (3) polynomial filters; (4) mathematical morphology (5) neural networks; (6) nonlinear image restoration. One of the main current limitations of nonlinear techniques is the lack of unifying theory, that can encompass all existing nonlinear filter classes. Each class of nonlinear processing techniques possesses its own mathematical tools that can provide reasonably good analysis of its performance. Cross-fertilization of these classes has been proven to be promising. For example, mathematical morphology and order statistic filters have been efficiently integrated in one class, although they come from completely different origins. The authors focus on the morphological and order statistic filters their properties and especially on trends in these two areas.< >

Nonlinear digital signal/image processing: overview and perspectives

Abstract: This paper presents an overview of the various nonlinear digital filter classes that exist in the literature. The properties of these classes are presented as well as their applications. The interrelation of the the filter classes and the efforts towards unification are exposed as well. Finally, the current trends in this area are described.

Application of directional statistics to nonlinear vector field filtering and analysis

Abstract: Vector fields play an important role in the area of color image processing and image sequence processing. Certain classes of vector fields can be handled more efficiently in a polar coordinate system. Optimal L-filters for filtering of vector magnitude are derived. The principles of directional data are presented and the notion of ordering is extended to angular data in order to filter and analyze the vector direction.

Combined Evaluation of Motion and Disparity Vector Fields for Stereoscopic Sequence Coding

Abstract: Efficient, disparity and motion compensated compression techniques are necessary for the transmission of 3DTV signals. Stereo vision geometry imposes certain coherence constraints between vector fields. The method proposed in this paper uses such a constraint in an effort to jointly optimize the four vectors involved in it. Both a candidate testing algorithm and a gradient-based method are tested for the optimization.

Error analysis of recursive algorithms for computing the correlation matrix of the order statistics

Abstract: The computation of the expected values and moments of the order statistics has received much attention since the mid-1950's, because of the prominent position order statistics in the field of robust estimation [1]-[6],[9]. The computation of the second-order and product moments appearing in the correlation matrix of the order statistics is of fundamental importance, because the correlation matrix of the order statistics is involved in the de sign of L-filters both single-channel [7,8,10] as well as multichannel ones [13]. Furthermore, a need for computing the correlation matrix of the order statistics is also arisen when the assessment of the performance of adaptive nonlinear filters based on Least Mean Squares (LMS) or Recursive Least Squares algorithms is to be pursued [11,12]. For example, the stability as well as the rate of convergence of LMS adaptive L-filters depend on the extreme eigenvalues of the correlation matrix of the order statistics.