Showing papers on "Line segment published in 1997"

Automatic line matching across views

Abstract: The paper presents a new method for matching individual line segments between images. The method uses both grey-level information and the multiple view geometric relations between the images. For image pairs epipolar geometry facilitates the computation of a cross-correlation based matching score for putative line correspondences. For image triplets cross-correlation matching scores are used in conjunction with line transfer based on the trifocal geometry. Algorithms are developed for both short and long range motion. In the case of long range motion the algorithm involves evaluating a one parameter family of plane induced homographies. The algorithms are robust to deficiencies in the line segment extraction and partial occlusion. Experimental results are given for image pairs and triplets, for varying motions between views, and for different scene types. The three view algorithm eliminates all mismatches.

Geometric pattern matching under Euclidean motion

Abstract: Given two planar sets A and B, we examine the problem of determining the smallest ϵ such that there is a Euclidean motion (rotation and translation) of A that brings each member of A within distance ϵ of some member of B. We establish upper bounds on the combinatorial complexity of this subproblem in model-based computer vision, when the sets A and B contain points, line segments, or (filled-in) polygons. We also show how to use our methods to substantially improve on existing algorithms for finding the minimum Hausdorff distance under Euclidean motion.

Microwave imaging-Location and shape reconstruction from multifrequency scattering data

Abstract: The problem of determining the shape and location of an object embedded in a homogeneous dissipative medium from measurements of the field scattered by the object is considered in this paper. The object is assumed to be an infinite cylinder of known cross section illuminated by a TM plane wave and the scattered field is measured on a line segment perpendicular to the direction of incidence. Measurement data are carried out at three different frequencies for a homogeneous cylinder of known dielectric constant. The location and contour shape are determined using two different reconstruction algorithms, a Newton-Kantorovich (NK) method and the modified gradient (MG) method whose effectiveness and robustness are compared. Both methods are based on domain integral representations of the field in the body. They involve an iterative minimization of the defect between an integral representation of the field measured on the line and the actual measured data. The NK method involves a linearization of the nonlinear relation between the field and the contrast, as well as the solution of a direct scattering problem at each iteration. The MG method seeks the simultaneous reconstruction of the field and the characteristic function of the support of the scatterer without solving a direct problem at each step. Both methods employed the same initial guess and the a priori information that the characteristic function is nonnegative.

Deformed map automatic generation system including automatic extraction of road area from a block map and shape deformation of at least one road area drawn in the map

Abstract: A vector block information indicating a plurality of line segment rows is read out from a block map. Each line segment row divides an area of a block from an area of a road in an image drawing region. Each point at which one end of a line segment row contacts with a boundary line of the image drawing region is set as a boundary point. In this case, a road crossing the boundary line of the image drawing region is indicated by a pair of line segment rows having a pair of boundary points close to each other. To determine an area of the road, the pair of boundary points are connected with each other through a connecting line segment. Therefore, a road area in the image drawing region can be automatically extracted from the block map. In addition, an end area of the road placed at the boundary line of the image drawing region is reshaped to have two right-angle corners at the end area of the road, a length of a road area at the boundary line of the image drawing region is shortened, an area of a narrow road is deleted from the road area, an area of a narrow road is widened, or a plurality of blocks placed in an administrative district is unified by deleting areas of a plurality of roads dividing the blocks.

How easy is matching 2D line models using local search

Abstract: Local search is a well established and highly effective method for solving complex combinatorial optimization problems. Here, local search is adapted to solve difficult geometric matching problems. Matching is posed as the problem of finding the optimal many-to-many correspondence mapping between a line segment model and image line segments. Image data is assumed to be fragmented, noisy, and cluttered. The algorithms presented have been used for robot navigation, photo interpretation, and scene understanding. This paper explores how local search performs as model complexity increases, image clutter increases, and additional model instances are added to the image data. Expected run-times to find optimal matches with 95 percent confidence are determined for 48 distinct problems involving six models. Nonlinear regression is used to estimate run-time growth as a function of problem size. Both polynomial and exponential growth models are fit to the run-time data. For problems with random clutter, the polynomial model fits better and growth is comparable to that for tree search. For problems involving symmetric models and multiple model instances, where tree search is exponential, the polynomial growth model is superior to the exponential growth model for one search algorithm and comparable for another.

Angle-restricted tours in the plane

Abstract: For a given set A ⊆ (− π ; + π ] of angles, the problem “Angle-Restricted Tour” (ART) is to decide whether a set P of n points in the Euclidean plane allows a closed directed tour consisting of straight line segments, such that all angles between consecutive line segments are from the set A . We present a variety of algorithmic and combinatorial results on this problem. In particular, we show that any finite set of at least five points allows a “pseudoconvex” tour (i.e., a tour where all angles are nonnegative), and we derive a fast algorithm for constructing such a tour. Moreover, we give a complete classification (from the computational complexity point of view) for the special cases where the tour has to be part of the orthogonal grid.

Method and apparatus for detecting a lane on a road

Abstract: The invention is directed to a method and apparatus for detecting a lane on a road, the latter of which includes an edge sensor (ED) for detecting edges indicative of boundaries of opposite sides of each lane, an edge memory (EM) for storing the edges to form a couple of continuous lines of edge groups indicative of the opposite boundaries of the lane, and an image processor (DP) for defining the opposite boundaries of the lane on a plane in a three-dimensional geometry. A center of curvature defining device (RC) defines a center of curvature to the opposite boundaries of the lane. An edge line extracting device (ES) divides the edge groups stored in the edge memory into a plurality of continuous regions each having a predetermined length along a concentric circle arc having the same center as the center of curvature, and extracts a pair of edge lines from each region. A tangent defining device (TL) defines a couple of tangents to the edge lines, and a normal defining device (NL) defines a normal to the tangents. Then, a segment setting device (SD) defines a couple of intersections of the normal with the tangents, and sets a line segment between the intersections as a segment of the lane. The image processor (DP) renews the opposite boundaries of the lane on the basis of the segment of the lane.

Motion compensated digital video decoding with buffered picture storage memory map

Abstract: A digital video presentation system (30) is provided with hardware and software logic for mapping the picture data into buffer memory (48) in a way that permits both the reading of motion vector compensated macroblocks (S) of data and the reading of horizontal picture wide scan lines with a low number of memory page crossings. Preferably, the memory (48) is a plurality of rows (74), for example 16 rows, wide. Preferably, 16 lines of 8-pixel (two 32 pixel wide column) line segments of 8x8 pixel blocks (A-P) are stored in consecutive storage locations (76) followed by the consecutive storage vertically adjacent line segments until one line segment is stored in each logical row (74) of the memory (48). Then the next horizontally adjacent set of line segments of similarly stored until the right boundary of the picture is reached, then the each additional row of 16 lines of the picture similarly are stored until the bottom of the picture is reached. Each 16x16 pixel macroblock (S) of data is stored on a single page (75); preferably, two horizontally adjacent macroblocks (S) are stored on one page of memory (48). Each line of the picture is stored in contiguous locations on the same row of the memory. The motion compensation logic (77) interprets motion vectors from the incoming data (32) and calculates adresses for a macroblock (S) of picture data by separating read commands into separate commands where a page boundary divides the macroblock into vertically adjacent rectangles. Memory controller logic (50) further divides such rectangles where they cross boundaries between horizontally adjacent pages of the memory. One fixed address 8-hex increment steps from line segment to vertically adjacent line segment while another fixed address 80-hex increment steps horizontally from one 8-pixel line segment to the next, such as across a scan line of the picture.

Flash memory array and decoding architecture

Abstract: A flash memory circuit having a word line decoder with even and odd word line latches and a source line decoder with a source line latch is disclosed. The word line decoders and source line decoder provide the capability of erasing the memory cells of two adjacent word lines in a flash memory simultaneously and verifying the memory cells word line by word line. By erasing two adjacent rows simultaneously, the embodiments of this invention eliminates over-erasure and source disturbance problems associated with conventional flash memory circuits. The decoding architecture provides flexible erase size that can be from a pair to a large number of multiple pairs of word lines. By dividing the memory cells of a word line into a number of segments, the decoding circuit further provides the capability of selecting the memory cells of a word line segment for erasing.

Matching sets of 3D line segments with application to polygonal arc matching

Abstract: In this paper, we consider two sets of corresponding 3D line segments of equal length. We derive a closed-form solution for the coordinate transform (rotation and translation) that gives the best match between the two sets; best in the sense of a least-squares distance measure between the sets. We use these results as the basis to construct efficient algorithms for solving other problems in computer vision. Specifically, we address the problem of matching polygonal arcs, that is, the problem of finding a match between a short arc and a piece of long arc.

New results on binary space partitions in the plane

Abstract: We prove the existence of linear size binary space partitions for sets of objects in the plane under certain conditions. In particular, we construct linear size binary space partitions for sets of fat objects, for sets of line segments where the ratio between the lengths of the longest and shortest segment is bounded by a constant, and for homothetic objects. For all cases we also show how to turn the existence proofs into efficient algorithms.

Road area extracting apparatus for extracting a road area from a block map, deformed map automatic generation system for generating a deformed map from road area data obtained by the road area extracting apparatus, map information providing system, geographical information providing system and geographical information describing method

Abstract: A vector block information indicating a plurality of line segment rows is read out from a block map. Each line segment row divides an area of a block from an area of a road in an image drawing region. Each point at which one end of a line segment row contacts with a boundary line of the image drawing region is set as a boundary point. In this case, a road crossing the boundary line of the image drawing region is indicated by a pair of line segment rows having a pair of boundary points close to each other. To determine an area of the road, the pair of boundary points are connected with each other through a connecting line segment. Therefore, a road area in the image drawing region can be automatically extracted from the block map. In addition, an end area of the road placed at the boundary line of the image drawing region is reshaped to have two right-angle corners at the end area of the road, a length of a road area at the boundary line of the image drawing region is shortened, an area of a narrow road is deleted from the road area, an area of a narrow road is widened, or a plurality of blocks placed in an administrative district is unified by deleting areas of a plurality of roads dividing the blocks.

Homotopic shape deformation

Abstract: An image warping method is presented that deforms an image continuously without foldover while observing a given set of trajectories of feature elements. Any intermediate image during the morph is homeomorphic to the initial image and the morphing process is a homotopy. The method permits points, line segments, and polygons to be included as features in the image. Techniques are described to make the deformation process smooth in spatial as well as temporal dimensions. An experimental result is included to demonstrate the feasibility of our approach and complexity of the algorithm is analyzed. Morphing between images that are topologically unequivalent is also briefly discussed.

Feature space trajectory for distorted-object classification and pose estimation in synthetic aperture radar

Abstract: Classification and pose estimation of distorted input objects are considered The feature space trajectory representation of distorted views of an object is used with a new eigenfeature space For a distorted input object, the closest trajectory denotes the class of the input and the closest line segment on it denotes its pose If an input point is too far from a trajectory, it is rejected as clutter New methods for selecting Fukunaga-Koontz discriminant vectors, the number of dominant eigenvectors per class and for determining training, and test set compatibility are presented

A polynomial time solution for labeling a rectilinear map

Abstract: Given a rectilinear map consisting of n disjoint line segments, the corresponding map labeling problem is to place a maximum width rectangle at each segment using one of the three natural ways. In a recent paper, it is shown that if all segments are horizontal then the problem can be solved in optimal Θ(n log n) time. For the general problem a factor-2 approximate solution and a Polynomial Time Approximation Scheme are also proposed. In this paper, we show that the general problem is polynomially solvable with a nontrivial use of 2SAT and the solution can be even generalized to the case of allowing k natural placements for each segment, where k is any fixed constant. We believe this technique can be also used to solve other geometric packing problems.

Motion compensated digital video decoding and buffer memory addressing therefor

Abstract: A digital video presentation system is provided with hardware and software logic for mapping the picture data into buffer memory in a way that permits both the reading of motion vector compensated macroblocks of data and the reading of horizontal picture wide scan lines with a low number of memory page crossings. Preferably, the memory is a plurality of rows, for example 16 rows, wide. Preferably, 16 lines of 8-pixel (two 32 pixel wide column) line segments of 8×8 pixel blocks are stored in consecutive storage locations followed by the consecutive storage vertically adjacent line segments until one line segment is stored in each logical row of the memory. Then the next horizontally adjacent set of line segments of similarly stored until the right boundary of the picture is reached, then the each additional row of 16 lines of the picture similarly are stored until the bottom of the picture is reached. Each 16×16 pixel macroblock of data is stored on a single page; preferably, two horizontally adjacent macroblocks are stored on one page of memory. Each line of the picture is stored in contiguous locations on the same row of the memory. The motion compensation logic interprets motion vectors from the incoming data and calculates addresses for a macroblock of picture data by separating read commands into separate commands where a page boundary divides the macroblock into vertically adjacent rectangles. Memory controller logic further divides such rectangles where they cross boundaries between horizontally adjacent pages of the memory. One fixed address increment of 8 hex steps from line segment to vertically adjacent line segment while another fixed address increment of 80 hex steps horizontally from one 8 pixel line segment to the next, such as across a scan line of the picture.

Edge cycle collision detection in graphics environment

Abstract: A collision of a moving object with another object in a 3D virtual world is efficiently determined in real time so that the technique is usable over a network. An object in the 3D virtual world is projected onto a 2D plane. The projection of the object is bounded by a two-sided edge cycle that is formed from a sequence of line segments having endpoints that are consecutively joined at an ordered series of vertices. One side, which is formed by a counterclockwise ordering of the line segments and their respective vertices, is characterized as "front facing." The other side is characterized as "back facing." Additionally, the front and back facing characterization of a side of the edge cycle is determined by the disposition of the moving object, i.e., the closest side of the edge cycle to the moving object is its back facing side. A path line segment is developed for the projection of a path of a moving object onto the 2D plane, and a path line equation is determined for the path line segment. The path line equation is employed to identify any positive to negative sign transition between consecutive vertices of a front facing line segment of the edge cycle. An edge cycle line equation is then evaluated with the endpoints of the path line segment for the moving object to determine if the path line segment has intersected the line segment of the edge cycle during a particular time period.

An Efficient Code‐Based Voxel‐Traversing Algorithm

Abstract: The paper considers an efficient approach to traversing a uniformly-subdivided space pierced by a line segment. A voxel, as the basic constituent element of the uniformly subdivided space, is restricted to having the form of a cube. The algorithm works in two steps. In the first step, the so-called Bresenham voxels are identified and, by comparing their position codes, their type of connectivity is determined. To achieve the required connectivity between neighbouring voxels, the second step of the algorithm is applied to find the missing voxels. In this way, the algorithm efficiently switches between face-, edge- and vertex-connectivity. Although the algorithm works with oating-point precision, it is extremely computationally efficient, and tests of speed compared with the Muller, Cleary & Wyvill, Amanatides & Woo, and Zemcik algorithms are described.

Steiner Trees for Terminals Constrained to Curves

Abstract: We give a polynomial time algorithm for solving the Euclidean Steiner tree problem when the terminals are constrained to lie on a fixed finite set of disjoint finite-length compact simple smooth curves. The problem is known to be NP-hard in general. We also show it to be NP-hard if the terminals lie on two parallel infinite lines or on a bent line segment provided the bend has an angle of less than $120^\circ$.

Bulge Elimination in Implicit Surface Blends

Abstract: Implicit surface blends can be obtained by combining distance surfaces. These are surfaces defined by distance to ‘skeletal’ elements such as points, line segments, polygons, or any free-form curve, surface, or volume. To illustrate, we consider distance to the planar curve shown below. Computing this distance is demanding [Bloomenthal 1989], [Schneider 1990], and often a piecewise linear approximation is substituted for the curve.

Fast Line and Rectangle Detection by Clustering and Grouping

Abstract: Fast algorithms to detect lines and rectangles in grey-scale images are proposed. At first, a contour image is obtained by the modified edge detection scheme. The linear primitives are extracted in the contour image and joined into line segments by cluster analysis method. The line merging algorithm is developed to get straight lines from segments. Algorithm to detect rectangles from the extracted straight lines is suggested. The developed algorithms are fast and permit to get the qualitative result.

Detector for original

Abstract: PROBLEM TO BE SOLVED: To allow a detector to detect the position and inclination of an original with high accuracy and few midsections, even in the case of a non- rectangular original by deciding an original area, based on line segments extracted by a line segment extract means and the inclination of the original obtained by an original tilt detection means. SOLUTION: An original tilt detection means defines the inclination of the longest line segment among line segments extracted by a line segment extract means as the inclination of an original, Then an original area is decided based on the line segment among line segments extracted by a line segment extract means and the inclination of the original obtained by the original inclination detection means. A CPU 314 of the detector detects the change of the inclination between coordinate data of each original edge, based on coordinate data stored in a stack memory 308 to obtain points (edge change points), where the inclination is changed and to extract line segments tying the edge change points. Thus, the longest line segment is decided among the detracted line segments, straight lines surrounding the extected lie segments are obtained, and then four straight lines surrounding the original area are decided. COPYRIGHT: (C)1999,JPO

The recognition of form documents based on three types of line segments

Abstract: Almost all form documents contain line segments. We propose an efficient method to recognize the form document that contains at least one line segment. Our method is based on an efficient representation model of the form. The representation model uses three types of line segments to represent a form. All line segments are normalized and sorted after they are extracted. The normalization and sorting not only solve the problems of magnification and contraction but also provide a unified and efficient way of matching between forms. To make the recognition method more robust, fuzzy matching is used. Experimental results show the effectiveness and efficiency of the method.

Apparatus and method for generating antialiased polygons

Abstract: A first forming means forms a first line segment along an edge of a polygon which has been delineated on a picture memory. A second forming means forms a plurality of second line segments along the length of the first line segment. The second line segments are comprised of a plurality of pixels and are generally perpendicular to the direction of the first line segment. The values of the second-line-segment pixels may be found using the values of pixels lying on the first line segment. Alternatively, the values of the second-line-segment pixels may be found by interpolation using the values of pixels adjacent the pixels lying on the first line segment.

Advances in computer vision

Abstract: A semidiscrete nonlinear scale-space theory and its relation to the Perona-Malik paradox.- Topological approach to mathematical morphology.- Segmentation by watersheds: definition and parallel implementation.- A graph network for image segmentation.- Associative memory for images by recurrent neural subnetworks.- Optimal models for visual recognition.- Order of points on a line segment.- Subjective contours detection.- Texture feature based interaction maps: potential and limits.- Non-Markov Gibbs image model with almost local pairwise pixel interactions.- Equivalent contraction kernels to build dual irregular pyramids.- Towards a generalized primal sketch.- Categorization through temporal analysis of patterns.- Detection of regions of interest via the Pyramid Discrete Symmetry Transform.- Dense depth maps by active color illumination and image pyramids.- Local and global integration of discrete vector fields.- A new approach to shape from shading.- Recent uniqueness results in shape from shading.- Computation of time-varying motion and structure parameters from real image sequences.- A theory of occlusion in the context of optical flow.- Algebraic method for solution of some best matching problems.- Determining the attitude of planar objects with general curved contours from a single perspective view.- CAD based 3D object recognition on range images.- Dual quaternions for absolute orientation and hand-eye calibration.- Segmentation of behavioral spaces for navigation tasks.- Geometric algebra as a framework for the perception-action cycle.- List of contributors.

Camera-blur parameter detection method and information recording medium

Abstract: PROBLEM TO BE SOLVED: To automatically detect parameters for camera-blur such as the direction and magnitude for the correction of an image which is deteriorated due to a camera-blur and to easily confirm/correct the parameters. SOLUTION: Prameters that denote camera-blurs, such as the direction and magnitude are detected automatically from an image that is obtained by transformed an object image into a 2-dimensional frequency space. A partial image (A) of the object image is displayed on display screen, and an image (B) resulting from overlapping a line segment 10 corresponding to a camera-blur direction parameter, and line segments 12, 14 corresponding to a camera-blur magnitude parameter in the converted image is displayed on another display screen. The line segment 10 is rotated, and the line segments 12, 14 are moved through the operation of a mouse by the operator to correct the parameters. Corrected partial images D1, D2, D3 according to an automatic detection of 1st parameter candidate, an automatic detection of 2nd parameter candidate and corrected parameters respectively are displayed at each occasion and the user selects one of the corrected partial images D1, D2, D3 so as to confirm the parameters.