Showing papers by "Azriel Rosenfeld published in 1980"
TL;DR: This paper presents an algorithm for converting from quadtrees to a simple class of boundary codes and is shown to have an execution time proportional to the perimeter of the region.
Abstract: There has been recent interest in the use of quadtrees to represent regions in an image. It thus becomes desirable to develop efficient methods of conversion between quadtrees and other types of region representations. This paper presents an algorithm for converting from quadtrees to a simple class of boundary codes. The algorithm is shown to have an execution time proportional to the perimeter of the region.
173 citations
TL;DR: Cooperating local parallel processes can be used as aids in assigning numerical or symbolic labels to image or scene parts in low-level vision.
163 citations
TL;DR: In experiments using a color image of a house, the relaxation approach gave markedly superior performance; relaxation eliminated 4-8 times as many errors as the other methods did.
Abstract: Three approaches to reducing errors in multispectral pixel classification were compared: 1) postprocessing (iterated reclassification based on comparison with the neighbors' classes); 2) preprocessing (iterated smoothing, by averaging with selected neighbors, prior to classification); and 3) relaxation (probabilistic classification followed by iterative probability adjustment). In experiments using a color image of a house, the relaxation approach gave markedly superior performance; relaxation eliminated 4-8 times as many errors as the other methods did.
86 citations
TL;DR: In segmenting an image by pixel classification, the sequence of gray levels of the pixel's neighbors can be used as a feature vector that yields classifications at least as good as those obtained using other local properties as features.
34 citations
01 Jan 1980
TL;DR: A method of representing arrays by trees of degree 4, which are constructed by recursively subdividing the array into blocks of constant value, can be applied to binary arrays representing regions.
Abstract: This paper describes a method of representing arrays by trees of degree 4, which are constructed by recursively subdividing the array into blocks of constant value. In particular, this method can be applied to binary arrays representing regions (1's are region points). Algorithms for conversion between this and other region representations, and for measuring geometric properties of regions represented in this way, are informally described.
13 citations
TL;DR: If the automaton is allowed to shrink (i.e., cells are deleted, and their predecessors become directly connected to their successors), language acceptance in less than linear time becomes possible.
12 citations
01 Feb 1980
TL;DR: An approach to image segmentation based on using a quadtree approximation image to identify regions in the image is examined, finding that by traversal of the tree it is possible to find the approximate location of the regions within an image.
Abstract: : This paper examines an approach to image segmentation based on using a quadtree approximation image to identify regions in the image. The advantages of this approach are firstly that noise points are not present in the quadtree approximation of an image and secondly that by traversal of the tree it is possible to find the approximate location of the regions within an image. It is thus possible to extract a region by applying thresholds in the vicinity of the region and to identify the best threshold by correlating the extracted region with an edge map of the original image. Since thresholds to extract regions are applied locally, noise points which may be present elsewhere in the edge map have no effect on the correlation of extracted regions to the edge map of the original image.
11 citations
01 Jan 1980
TL;DR: Experimental results are presented on some properties of random mosaic models for textures and their theoretically predicted values are compared with observations on a real Image.
Abstract: Experimental results are presented on some properties of random mosaic models for textures. These observations are compared with the theoretically predicted values. The predictions are also compared with observations on a real Image.
10 citations
TL;DR: A class of parallel computers is described, and general examples are given illustrating how such a computer could initially configure itself to represent a given decomposition of an image into regions, and dynamically reconfigure itself, in parallel, as regions merge or split.
10 citations
01 Nov 1980
TL;DR: In this article, a number of variations on the basic block linking approach are investigated, and some tentative conclusions are drawn regarding preferred methods of initializing the process and of defining the links, yielding improvements over the originally proposed method.
Abstract: When an image is smoothed using small blocks or neighborhoods, the results may be somewhat unreliable due to the effects of noise on small samples. When larger blocks are used, the samples become more reliable, but they are more likely to be mixed, since a large block will often not be contained in a single region of the image. A compromise approach is to use several block sizes, representing versions of the image at several resolutions, and to carry out the smoothing by means of a cooperative process based on links between blocks of adjacent sizes. These links define "block trees" which segment the image into regions, not necessarily connected, over which smoothing takes place. In this paper, a number of variations on the basic block linking approach are investigated, and some tentative conclusions are drawn regarding preferred methods of initializing the process and of defining the links, yielding improvements over the originally proposed method.© (1981) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.
9 citations
01 Dec 1980
TL;DR: In this paper, a tutorial paper on the subjects of digital straightness and convexity is presented, where the central questions treated are: When can a digital arc be the digitization of a real straight line segment? When can an object be the digital object in a real convex set?
Abstract: : This tutorial paper reviews the subjects of digital straightness and convexity. The central questions treated are: When can a digital arc be the digitization of a real straight line segment? When can a digital object be the digitization of a real convex set?
01 Sep 1980
TL;DR: In this paper, it was shown that characterizations of simple closed curves in two-dimensional arrays (digital pictures) are insufficient to characterize simple closed surfaces in 3D arrays; but the 2D analog of a recently given 3D definition of "surface" does characterize curves.
Abstract: : Several characterizations of simple closed curves in two-dimensional arrays (digital pictures) are shown to be inadequate to characterize simple closed surfaces in three-dimensional arrays; but the 2d analog of a recently given 3d definition of 'surface' does characterize curves. (Author)
TL;DR: A bibliography of 700 references related to the computer processing of pictorial information, arranged by subject matter is presented, restricted to a selected set of U.S. journals and proceedings of specialized meetings.
TL;DR: Sequential and cellular d-graph automata are defined, and it is shown that they can simulate each other and are also equivalent to web-bounded automata.
23 Dec 1980
TL;DR: A general review of texture analysis can be found in this paper, which summarizes some past work on texture analysis, including comparisons between different classes of features based on spatial statistics and Fourier analysis; recently developed refinements of the spatial statistics approach; methods based on the ex-traction and description of texture primitives; texture models based on random geometric processes; methods of improving texture classification by feature value smoothing; and methods of segmenting images into textured regions.
Abstract: This paper briefly summarizes some past work on texture analysis, including comparisonsbetween different classes of features based on spatial statistics and Fourier analysis;recently developed refinements of the spatial statistics approach; methods based on the ex-traction and description of texture primitives; texture models based on random geometricprocesses; methods of improving texture classification by feature value smoothing; andmethods of segmenting images into textured regions. References are given to reports andpapers in which further details can be found.IntroductionTexture analysis has been used extensively in many image Processing and recognition ap-plications over the past 25 years. A general review of the subject can be found in '.This paper briefly summarizes work on texture analysis, primarily at the University ofMaryland, during the past five years, and gives references to technical reports and papersin which further details can be found.Feature comparisons
TL;DR: In this article, the use of color gradient magnitude information as an aid in classifying pixels based on their colors or spectral signatures is discussed, where the gradient magnitude can be used in various ways to aid in thresholds selection.
Abstract: : When pixels in a black-and-white image are classified by thresholding their gray levels, gradient magnitude information can be used in various ways as an aid in thresholds selection. This note deals with the use of color gradient magnitude as an aid in classifying pixels based on their colors or spectral signatures.
TL;DR: Given a binary image containing a connected component, parallel propagation algorithms are presented for constructing upright and 45°-tilted framing rectangles around the component in time proportional to the sum of the dimensions of these rectangles.
01 Sep 1980
TL;DR: This algorithm should be especially useful in a relaxation type scheme for classifying linear feature segments when the degree of anti-parallelism of the segments is needed.
Abstract: : A method for finding anti-parallel straight edges is presented and discussed. This method is based on information about the object sides of edges, the similarity and homogeneity of gray level between the edges, the angle difference, the amount of overlap, and also information about the estimated object gray level. Two figures of merit are defined to calculate the mutual support of two antiparallel linear features. Examples are shown of applying the method to high resolution aerial photographs. Results indicate that cultural features such as roads and buildings can be extracted and that a significant reduction in the complexity of the image description can be obtained. This algorithm should be especially useful in a relaxation type scheme for classifying linear feature segments when the degree of anti-parallelism of the segments is needed.
01 Aug 1980
TL;DR: In this article, a method of linking compatible straight edges is presented and discussed, which is based on information about the geometrical configurations of the edges, the similarity of the gray levels on their object sides, and the similarity with the line joining their endpoints.
Abstract: : A method of linking compatible straight edges is presented and discussed. The linking is based on information about the geometrical configurations of the edges, the similarity of the gray levels on their object sides, and the similarity of their object sides with the line joining their endpoints. Three figures of merit are defined for evaluating pairs of segments for possible linking. Examples are shown of applying the method to high resolution aerial photographs. Results indicate that cultural features such as roads and buildings can be extracted and that a significant reduction in the complexity of the image description can be obtained. This approach should be especially useful for defining the degrees of compatibility of pairs of edges in a relaxation scheme for classifying linear feature segments.
01 Jul 1980
TL;DR: In this article, straight line segments are fitted to sets of edge pixels; compatibilities between pairs of these segments, based on gray level and geometric information, are computed; and the segments are then grouped into building-like and road-like groupings based on these compatibilities.
Abstract: : This paper describes steps toward the recognition of cultural features such as buildings and roads on aerial photographs. The approach involves several successive stages of grouping of edge segments. Straight line segments are fitted to sets of edge pixels; compatibilities between pairs of these segments, based on gray level and geometric information, are computed; and the segments are then grouped into building-like and road-like groupings based on these compatibilities. Examples of the results obtained using this approach are given, and some variations on the initial stages of the process are also investigated.