Showing papers by "Azriel Rosenfeld published in 1991"
TL;DR: The least-median-of-squares (LMedS) method, which yields the correct result even when half of the data is severely corrupted, is described and compared with the class of robust M-estimators.
Abstract: Regression analysis (fitting a model to noisy data) is a basic technique in computer vision, Robust regression methods that remain reliable in the presence of various types of noise are therefore of considerable importance. We review several robust estimation techniques and describe in detail the least-median-of-squares (LMedS) method. The method yields the correct result even when half of the data is severely corrupted. Its efficiency in the presence of Gaussian noise can be improved by complementing it with a weighted least-squares-based procedure. The high time-complexity of the LMedS algorithm can be reduced by a Monte Carlo type speed-up technique. We discuss the relationship of LMedS with the RANSAC paradigm and its limitations in the presence of noise corrupting all the data, and we compare its performance with the class of robust M-estimators. References to published applications of robust techniques in computer vision are also given.
653 citations
TL;DR: The authors develop a scheme for resolving conflicts for the case where range images share a common reference surface, i.e. when the transformation matrix between two overlapping images involves only three components: two translations and one rotation.
Abstract: A scheme is developed to register range images in an environment where distinctive features are scarce. Another issue addressed here is conflicting situations that may arise from pair-wise registration of multiple overlapping range images whether or not they contain distinctive features. There can be two causes for this: (a) error in individual registrations or (b) compression or bending in range images. The authors develop a scheme for resolving such conflicts for the case where range images share a common reference surface, i.e. when the transformation matrix between two overlapping images involves only three components: two translations and one rotation. The authors implemented this scheme to map the floor of the ocean, where the range data is obtained by a multibeam echo-sounder system installed aboard a sailing ship producing multiple overlapping range images. The system developed is the first automated system for correctly registered mapping of the ocean floor; it is efficient and robust. >
80 citations
[...]
65 citations
TL;DR: New algorithms for computing the Euler number of a 3D digital image S are given, based on smoothing the image to a differentiable object and applying theorems of differential geometry and algebraic topology.
44 citations
Patent•
09 Sep 1991
TL;DR: In this paper, a method is described for automatic detection of a large class of industrial parts, which is faster and more reliable than current methods, by detecting changes in the light intensity at these edges.
Abstract: A method is described for automatic detection of a large class of industrial parts, whch is faster and more reliable than current methods. Many industrial parts have visible outlines which are either circular or elliptical or contain segments of these. Circles are seen as ellipses unless viewed head-on. Thus the ability to detect ellipses in an image of an industrial part, quickly and reliably, can be of great help in identification of this part by a robotic assembly process. In our method one first scans the image electronically by a video camera. When a scan line goes through an image of some object one can detect the edges of this image by detecting changes in the light intensity at these edges. For an object of the shape of a conic section (e.g. an ellipse), generally two edges will appear on each scan line and the midpoints between them will all lie on one straight line. This straight line can be detected with Hough's method.
43 citations
TL;DR: A system for identifying small clouds and their shadows on aerial photographs segments an image into homogeneous regions; selects bright and dark regions as cloud and shadow candidates, respectively; and finds acceptable (cloud, shadow) pairs based on consistent relative position.
38 citations
TL;DR: A fuzzy disk with center P is a fuzzy set in which membership depends only on distance from P, and there is a maximal fuzzy disk gpƒ ⩽ ƒ centered at every point P, which is called the fuzzy medial axis transformation (FMAT) of Ɠ.
37 citations
TL;DR: An algorithm for matching two polygonal arcs by decomposition into pairs of equal length line segments that finds the best match by calculating analytically the relative position and orientation of the arcs that minimize the distance measure.
Abstract: In this paper we present an algorithm for matching two polygonal arcs. We first show how to match two arcs of the same length by decomposition into pairs of equal length line segments. The distance measure used is the sum of the squared distances between corresponding points on the two arcs. We find the best match by calculating analytically the relative position and orientation of the arcs that minimize the distance measure. After analyzing the case of arcs having the same length, we show how to use this result in the algorithm for arc matching, i.e., finding the piece of a long arc that best matches with a short arc. We illustrate its application with an example and suggest how to implement it efficiently. We also discuss simplifications of the algorithm in the case of digital images.
32 citations
02 Jun 1991
TL;DR: The authors present an approach to the recovery and recognition of 3-D objects from a single 2-D image based on grouping the regions into aspects and using the aspect hierarchy to infer a set of volumetric primitives and their connectivity.
Abstract: The authors present an approach to the recovery and recognition of 3-D objects from a single 2-D image. Given a recognition domain consisting of a database of objects, they select a set of object-centered 3-D volumetric modeling primitives that can be used to construct the objects. They take the set of primitives and generate a hierarchical aspect representation based on their projected surfaces; conditional probabilities capture the ambiguity of mappings between levels of the hierarchy. From a region segmentation of the input image, they present a novel formulation of the recovery problem based on grouping the regions into aspects. Once the aspects are recovered, they use the aspect hierarchy to infer a set of volumetric primitives and their connectivity. The recovered primitives are then used as indices into the object database for recognition. >
16 citations
TL;DR: A bibliography of nearly 1200 references related to computer vision and image analysis, arranged by subject matter is presented, covering topics including architectures; computational techniques; feature detection and segmentation; image analysis; and motion.
Abstract: This paper presents a bibliography of nearly 1200 references related to computer vision and image analysis, arranged by subject matter. The topics covered include architectures; computational techniques; feature detection and segmentation; image analysis; two-dimensional shape; pattern; color and texture; matching and stereo; three-dimensional recovery and analysis; three-dimensional shape; and motion. A few references are also given on related topics, such as geometry, graphics, image input/output and coding, image processing, optical processing, visual perception, neural nets, pattern recognition and artificial intelligence, as well as on applications.
16 citations
TL;DR: A bibliography of over 1600 references related to computer vision and image analysis, arranged by subject matter is presented, covering topics including architectures; computational techniques; feature detection, segmentation, and imageAnalysis.
Abstract: This paper presents a bibliography of over 1600 references related to computer vision and image analysis, arranged by subject matter. The topics covered include architectures; computational techniques; feature detection, segmentation, and image analysis; matching, stereo, and time-varying imagery; shape and pattern; color and texture; and three-dimensional scene analysis. A few references are also given on related topics, such as computational geometry, computer graphics, image input/output and coding, image processing, optical processing, visual perception, neural nets, pattern recognition, and artificial intelligence.
TL;DR: The structure of the response is to define the central problems of computer vision and the methodological paradigms that have been developed over the past 35 years, and to address the question of why vision is hard.
Abstract: In their paper, Jain and Binford claim that the ignorance, myopia, and naivete of today’s computer vision systems are due to the inadequacy of our methodological paradigms. They attempt to define the central problems in the field, and they make recommendations as to how these problems can be successfully addressed. The structure of our response is as follows: We start by defining the central problems of computer vision and the methodological paradigms that have been developed over the past 35 years. Then we address the question of why vision is hard, and we build around the authors’ statements our suggestions as to how progress in the field can be accelerated. We do not find computer vision to be a field in shambles. On the contrary, we find that it is becoming a well defined science and is well on its way to becoming a postparadigmatic field in the sense that research projects evolve naturally from a central paradigm.’
TL;DR: A general 'consensus' approach to combining multiple techniques, or multiple versions of the same technique, to obtain more reliable results is illustrated.
TL;DR: This work presents simple algorithms for recovering both X and ∂X from the set of corner vertices together with specifications of their types, in serial and parallel versions.
TL;DR: An algorithm of this type may account for the ability of humans to perceive at a glance whether the dots in a pattern are randomly or regularly spaced, i.e., their neighbor distances have high or low variance.
TL;DR: This paper presents a theoretical framework for designing solutions to computer vision problems and describes three methods of handling three-dimensional scenes and time-varying scenes.
Book•
01 Sep 1991
TL;DR: A survey of digital metrics by R. A. Rosenfeld and T. Y. Wu can be found in this article, where the authors present a set, its complement, and their common boundary.
Abstract: Self-similarity properties of digitized straight lines by A. M. Bruckstein Galleries and light matchings: Fat cooperative guards by J. Czyzowicz, I. Rival, and J. Urrutia Chord centers for convex polygons by M. Diaz and J. O'Rourke Discrete straight line segments: Parameters, primitives, and properties by L. Dorst and A. W. M. Smeulders Vision, geometry, and Minkowski operators by P. K. Ghosh Discrete multidimensional Jordan surfaces by G. T. Herman A survey of digital metrics by R. A. Melter Combinatorial and computational aspects of Minkowski decompositions by D. Mount and R. Silverman Connectedness of a set, its complement, and their common boundary by A. Rosenfeld and T. Y. Kong "Digital geometry" on graphs by A. Rosenfeld and A. Y. Wu Star-shapedness of digitized planar shapes by D. Shaked, J. Koplowitz, and A. M. Bruckstein Algorithms for the decomposition of convex polygons by R. Silverman and A. H. Stein Decomposition of discrete curves into piecewise straight segments in linear time by A. W. M. Smeulders and L. Dorst Digitization schemes and the recognition of digital straight lines, hyperplanes, and flats in arbitrary dimensions by I. Stojmenovic and R. Tosic Computational geometry and computer vision by G. T. Toussaint Convexity, visibility, and orthogonal polygons by D. Wood, G. J. E. Rawlins, and S. Schuierer.