Showing papers on "Line segment published in 1995"

Visual feature integration and the temporal correlation hypothesis

Abstract: The mammalian visual system is endowed with a nearly infinite capacity for the recognition of patterns and objects. To have acquired this capability the visual system must have solved what is a fundamentally combinatorial prob­ lem. Any given image consists of a collection of features, consisting of local contrast borders of luminance and wavelength, distributed across the visual field. For one to detect and recognize an object within a scene, the features comprising the object must be identified and segregated from those comprising other objects. This problem is inherently difficult to solve because of the combinatorial nature of visual images. To appreciate this point, consider a simple local feature such as a small vertically oriented line segment placed within a fixed location of the visual field. When combined with other line segments, this feature can form a nearly infinite number of geometrical objects. Any one of these objects may coexist with an equally large number of other

Structure and motion from line segments in multiple images

Abstract: This paper presents a new method for recovering the three dimensional structure of a scene composed of straight line segments using the image data obtained from a moving camera. The recovery algorithm is formulated in terms of an objective function which measures the total squared distance in the image plane between the observed edge segments and the projections (perspective) of the reconstructed lines. This objective function is minimized with respect to the line parameters and the camera positions to obtain an estimate for the structure of the scene. The effectiveness of this approach is demonstrated quantitatively through extensive simulations and qualitatively with actual image sequences. The implementation is being made publicly available. >

Facial feature extraction method and apparatus for a neural network acoustic and visual speech recognition system

Abstract: A facial feature extraction method and apparatus uses the variation in light intensity (gray-scale) of a frontal view of a speaker's face. The sequence of video images are sampled and quantized into a regular array of 150×150 pixels that naturally form a coordinate system of scan lines and pixel position along a scan line. Left and right eye areas and a mouth are located by thresholding the pixel gray-scale and finding the centroids of the three areas. The line segment joining the eye area centroids is bisected at right angle to form an axis of symmetry. A straight line through the centroid of the mouth area that is at right angle to the axis of symmetry constitutes the mouth line. Pixels along the mouth line and the axis of symmetry in the vicinity of the mouth area form a horizontal and vertical gray-scale profile, respectively. The profiles could be used as feature vectors but it is more efficient to select peaks and valleys (maximas and minimas) of the profile that correspond to the important physiological speech features such as lower and upper lip, mouth corner, and mouth area positions and pixel values and their time derivatives as visual vector components. Time derivatives are estimated by pixel position and value changes between video image frames. A speech recognition system uses the visual feature vector in combination with a concomitant acoustic vector as inputs to a time-delay neural network.

Reasoning about ordering

Abstract: Ordering information is a special type of spatial information that derives from the linear, planar or spatial ordering of points. A definition of ordering information in terms of the orientation of simplexes is used in this paper to introduce a system of line segment relations which generalizes Allen's system of interval relations to two dimensions. It shows that this generalization differs in interesting properties from the generalizations based on topological relations which have been proposed so far. The conceptual neighborhood structure of the line segment relations provides the foundation of ordering information reasoning. This is illustrated with an example from motion planning. Finally, the problem of representing ordering information is addressed. In that context the cell complex representation of Frank and Kuhn is compared with the approach presented here.

Estimating motion and structure from correspondences of line segments between two perspective images

Abstract: Presents an algorithm for determining 3D motion and structure from correspondences of line segments between two perspective images. To the author's knowledge, this paper is the first investigation of use of line segments in motion and structure from motion. Classical methods use their geometric abstraction, namely straight lines, but then three images are necessary for the motion and structure determination process. In this paper the author shows that it is possible to recover motion from two views when using line segments. The assumption used is that two matched line segments contain the projection of a common part of the corresponding line segment in space, i.e., they overlap. Indeed, this is what the author uses to match line segments between different views. This assumption constrains the possible motion between two views to an open set in motion parameter space. A heuristic, consisting of maximizing the overlap, leads to a unique solution. Both synthetic and real data have been used to test the proposed algorithm, and excellent results have been obtained with real data containing a relatively large set of line segments.

Source data interpolation method and apparatus

Abstract: A method of generating an upsampled target pixel positioned between two lines of input source data includes the step of comparing pixels of different lines of the source data in a region surrounding the upsampled target pixel to be generated in at least two different directions. An interpolation direction based on the comparison is selected and interpolations between selected pixels of the source data in the determined interpolation direction are carried out to compute intermediate pixels on a line segment passing through the upsampled target pixel. An interpolation between the intermediate pixels is carried out to generate the upsampled target pixel. An apparatus for performing the method is also disclosed.

Workflow modelling system

Abstract: A system generates a model of a workflow. The system begins by displaying on a computer screen a first flowgraph between two horizontal time lines representing two respective participants. The first flowgraph comprises a plurality of vertical directional line segments extending between the two horizontal time lines and representing respective communications or other actions between the participants. Next, the computer receives a selection of a second flowgraph for display between one of the two time lines and a third, horizontal time line representing a third participant. The second flowgraph comprises a plurality of vertical directional line segments extending between the one time line and the third horizontal time line and representing respective communications or other actions between the participants represented by the one and third time lines. Then, the computer displays the second flowgraph between the one and third time lines. The computer also generates a verbal description of communications or other actions represented by vertical line segments that lead to or from each time line.

Rounding arrangements dynamically

Abstract: We describe a robust, dynamic algorithm to compute the arrangement of a set of line segments in the plane, and its implementation. The algorithm is robust because, following Greene7 and Hobby,11 it rounds the endpoints and intersections of all line segments to representable points, but in a way that is globally topologically consistent. The algorithm is dynamic because, following Mulmuley,16 it uses a randomized hierarchy of vertical cell decompositions to make locating points, and inserting and deleting line segments, efficient. Our algorithm is novel because it marries the robustness of the Greene and Hobby algorithms with Mulmuley's dynamic algorithm in a way that preserves the desirable properties of each.

Lane guidance system

Abstract: A method and apparatus for continuously prompting a pilot to manually control and guide a vehicle along a selected line segment of a pattern of coverage of a target area. An on-board receiver and computer receives navigation reference signals, computes navigation parameters and generates navigation signals to an indicator. The indicator is placed within the normal forward field of outside vision of the vehicle's pilot. In a virtual flagman mode, the indicator has lower and upper horizontal lightbars on which discreet lights are illuminated to the left or right of the centerpoint to simulate the relative lateral position of actual flagmen at the beginning and end points of the line segment being tracked. In a flight director mode, the indicator shows a composite sum of factored navigation parameters including angle of intercept between the line segment and vehicle track, perpendicular distance of the vehicle from the line segment, and rate of turn of the vehicle, as a control instruction to the pilot. The system may include audible tone or voice prompts as well.

On the determinant of elliptic boundary value problems on a line segment

Abstract: In this paper we present a formula for the determinant of a matrix-valued elliptic differential operator of even order on a line segment [0, T] with boundary

A Robust Hough Transform Technique for Complete Line Segment Description

Abstract: When the Hough transform is applied to the detection of straight lines in images, it provides the parameters but not the length or the end points of the line. We propose a new algorithm for the determination of the length and the end points of a line in an image. It is very efficient in terms of computing time and does not depend on the sharpness of the peak in the accumulator array.

A Benchmark: Performance Evaluation of Dashed-Line Detection Algorithms

Abstract: This paper describes a protocol for systematically evaluating the performance of dashed-line detection algorithms. It includes a test image generator which creates random line patterns subject to prespecified constraints. The generator also outputs ground truth data for each line in the image. The output of the dashed line detection algorithm is then compared to these ground truths and evaluated using a set of criteria.

External-Memory Algorithms for Processing Line Segments in Geographic Information Systems (Extended Abstract)

Abstract: In the design of algorithms for large-scale applications it is essential to consider the problem of minimizing I/O communication. Geographical information systems (GIS) are good examples of such largescale applications as they frequently handle huge amounts of spatial data. In this paper we develop efficient new external-memory algorithms for a number of important problems involving line segments in the plane, including trapezoid decomposition, batched planar point location, triangulation, red-blue line segment intersection reporting, and general line segment intersection reporting. In GIS systems, the first three problems are useful for rendering and modeling, and the latter two are frequently used for overlaying maps and extracting information from them.

Method and apparatus for parametric signature verification using global features and stroke-direction codes

Abstract: A signature verification method operates by comparing the numerical values of parameters evaluated on a trial signature with stored reference data derived from previously entered reference signatures. The parameters include global features of the signature, and also include a stroke-direction code (SDC) of the signature. An SDC is derived by subdividing the signature into a sequence of time-ordered, spatially oriented line segments, each segment extending between a pair of discrete points along the signature. Each line segment has a stroke-direction value. The SDC is the ordered sequence of these stroke-direction values.

Ferroelectric nonvolatile random access memory utilizing self-bootstrapping plate line segment drivers

Abstract: A ferroelectric nonvolatile random access memory array includes multiple ferroelectric memory cells arranged in rows and columns, a word line coupled to a word line input of each of the ferroelectric memory cells in a row, and a bit line coupled to a bit line input of each of the ferroelectric memory cells in a column. The array also includes multiple plate lines, each plate line being arranged into a plurality of plate line segments each coupled to a plate line input of a predetermined number of the ferroelectric memory cells in a row, and multiple NMOS plate line segment drivers coupled to each of the plate line segments for selectively driving the corresponding plate line segment to a full rail voltage. The rows of ferroelectric memory cells and the NMOS plate line segment drivers have substantially the same layout pitch. The plate line segment drivers are each coupled to a center portion of the corresponding plate line segment. Each NMOS plate line segment drivers includes a first NMOS transistor having a first current node coupled to the word line associated with the ferroelectric memory cells coupled to the plate line segment, a gate coupled to a source of supply voltage, and a second current node, and a second NMOS transistor having a first current node coupled to the plate line segment, a second current node coupled to a plate clock line, and a gate coupled to the second current node of the first NMOS transistor.

Apparatus and method for converting line segment data to three-dimensional data

Abstract: An apparatus and method in which an operator can input a three-dimensional (3D) graphics to a graphic system by performing an operation that is as if he were sketching on a paper surface. A line segment of two-dimensional (2D) data is input through an input device such as a mouse, and is then temporarily stored in a storage device as well as being displayed on a 2D display screen. The line segment of 2D data is converted to 3D data when it obtains information in the depth direction based on its relationship to another line segment that constitutes a 3D graphics and has 3D data already. 3D data of line segments that have that data in addition to 2D data are stored in a storage device, and are projected by a projection conversion circuit onto a 2D projection plane that has been set by a visual field input device, so as to be converted to 2D data, which are then displayed on the display screen. Therefore, the line segments that have already been converted to 3D data are displayed on the display screen together with the line segments that have been input through the input device and still have only 2D data. An operator can construct a 3D graphics constituted of 3D line segments simply by inputting 2D line segments as if he were sketching.

Estimating motion and structure from correspondences of line segments between two perspective images

Abstract: Presents an algorithm for determining 3D motion and structure from correspondences of line segments between two perspective images. To our knowledge, the paper is the first investigation of use of line segments in motion and structure from motion. Classical methods use their geometric abstraction, namely straight lines, but then three images are necessary for the motion and structure determination process. We show that two views are in general sufficient when we use line segments. The assumption we use is that two matched line segments contain the projection of a common part of the corresponding line segment in space. Indeed this is what we use to match line segments between different views. Both synthetic and real data have been used to test the proposed algorithm, and excellent results have been obtained with real data containing a relatively large set of line segments. The results are comparable with those obtained using stereo calibration. >

A new approach for merging edge line segments

Abstract: A novel method for the linking or merging of edge line segments belonging to a single line is proposed; the segments can be non-overlapping or overlapping, either fully or partially. Explicit criteria for merging, based on proximity of segments and on directional difference are included in the process. I INTRODUCTION In many areas of computer vision it is often required to link or to merge line segments belonging to the same line. This may be required to reduce the number of entities having common characteristics in an image, thus building new entities with stronger characteristics (e.g., length, orientation). The method herein proposed is quite versatile, being appropriate either for linking nonoverlapping segments, or for merging line segments partially or fully overlapping. The main advantage of the method derives from the consideration of the lengths of each segment as explicit weights for the definition of the orientation and placement of the resulting line. II OVERVIEW Several methods have been proposed to merge two segments which are close of each other and have similar directions. We shall briefly describe two methods, both analysed in [1]. The first method is due to Liu and Huang. The principle of the method consists in defining the resulting segment as the one that connects the ends of the given segments which are farther apart, as shown in Fig. 1. _ The authors are with Departamento de Engenharia Electrotecnica e de Computadores, Faculdade de Engenharia da Universidade do Porto, Rua dos Bragas, 4099 PORTO Codex. Contact E-mail: tavares@obelix.fe.up.pt lr lj

Subpixel edge location in binary images using dithering

Abstract: This paper concerns the problem of obtaining subpixel estimates of the locations of straight edges in binary digital images using dithering. By adding uniformly distributed independent random noise it is shown that estimation bias may be removed and that the estimation variance is inversely proportional to the length of the line segment. The sensitivity to incorrect dither amplitude is calculated, and implementation is discussed. >

Time-optimal visibility-related algorithms on meshes with multiple broadcasting

Abstract: Given a collection of objects in the plane along with a viewpoint /spl omega/, the visibility problem involves determining the portion of each object that is visible to an observer positioned at /spl omega/. The visibility problem is central to various application areas including computer graphics, image processing, VLSI design, and robot navigation, among many others. The main contribution of this work is to provide time-optimal solutions to this problem for several classes of objects, namely ordered line segments, disks, and iso-oriented rectangles in the plane. In addition, our visibility algorithm for line segments is at the heart of time-optimal solutions for determining, for each element in a given sequence of real numbers, the position of the nearest larger element within that sequence, triangulating a set of points in the plane, determining the visibility pairs among a set of vertical line segments, and constructing the dominance and visibility graphs of a set of iso-oriented rectangles in the plane. All the algorithms in this paper involve an input of size n and run in O(log n) time on a mesh with multiple broadcasting of size n/spl times/n. This is the first instance of time-optimal solutions for these problems on this architecture. >

Snake beam: a paraxial arbitrary focal line.

Abstract: The creation of paraxial arbitrary focal lines by a Fourier computer-generated hologram is demonstrated. The desired focal line is represented by a series of connected straight line segments, each of which is implemented by a radial harmonic function located on a different radial portion of the entire hologram. Each subhologram is multiplied by appropriate linear and quadratic phase functions and is shifted by some distance from the center. The two phase factors determine the location of each line segment, while the in-plane shift determines the tilt angle of the segment.

Efficient method and apparatus for determining texture coordinates for lines and polygons

Abstract: A method of generating a pixel in a computer system. The computer system includes a processor and a memory. The memory includes a first polygon and a texture map. The method comprises the steps of: accessing a first point on a first polygon. A step of determining a first line segment having less than a predetermined error and determining whether a next line segment causes less than the predetermined error. The next line segment is longer than the first line segment in one direction. If the next line segment causes less than the predetermined error, then determine whether a new next line segment causes less than the predetermined error. Otherwise, access a texel in the texture map at a position corresponding to the first point and the first line segment, and generate the pixel from the texel.

Topologically sweeping the visibility complex of polygonal scenes

Abstract: In order to do efficient visibility computations in the plane, one must deal with maximal free line segments, that is line segments of maximal length in free space. To that end, [PV93b] introduce, for scene of convex curved objects, a new data structure, the visibility comp~ex, whose elements represent classes of maximal free segments having the same visibilities. In particular the visibility complex contains the visibility graph. We have studied the visibility complex for general polygonal scenes, and have devised an algorithm for topologically sweeping the visibility complex (and so for computing the visibility graph with the same complexity) in optimal time O(n log n + m) and optimal space O(n), where n is the total number of polygon vertices, and m the size of the visibility graph. Since it is a sweep algorithm, it improves the space storage of [GM87] for computing the visibility graph of polygonal scenes. Moreover, we have implemented our algorithm and we have made experimental comparisons with two other sweep algorithms for computing visibility graphs : the algorithm of [OW88] and the algorithm of [PV93a], which both run in time O(m log n). They were tested on scenes composed of line segments. We see that the gain due to the suppression of the logarithmic cost is significant.

Consistent Non-parametric Estimation for a One-dimensional Line Segment Process Observed in an Interval

Abstract: A line segment process on the line is observed through a finite window (0, x). This means that the data consists of line segments that are non-empty intersections of the process with the window. The EM-algorithm is used to construct the NPMLE on (0, t) of the distribution function of the length of the line segments. Furthermore the EM-algorithm provides the self-consistency equations. The NPMLE is unique and consistent. Consistency is proved by generalizing a method of Jewell (1982), which can be used in other models. The approach is based on convexity arguments. Some applications, another estimator and an extension to a two-dimensional line segment process will be discussed.

A contour approximation apparatus for representing a contour of an object

Abstract: A contour approximation apparatus for representing a contour image of an object comprises a polygonal approximation section for determining a number of vertices on the contour image and fitting the contour image with a plurality of line segments to provide a polygonal approximation of the contour image, a sampling circuit for providing N sample points for each of the line segments, an error detector for calculating an error for each of the N sample points on each of the line segments to produce a set of errors for each of the line segments, a discrete sine transform and quantization block for transforming each set of errors into a set of discrete sine transform coefficients, and for converting the set of discrete sine transform coefficients into a set of quantized transform coefficients.

Method and apparatus for dividing shape expressed on three-dimensional space into elements

Abstract: A shape model as an object to be analyzed and division information used to produce a finite element mesh are entered, and a recognition model is produced by allocating each edge line segmented by structural points of the shape model to any one coordinate axial direction in an orthogonal coordinate system. A mapping model is produced in which an orthogonal lattice is produced on a surface and the inside of an approximated shape model, and is displayed on a screen. The mapping model is modified by a system user in an interactive operation. A finite element mesh is generated from lattice points of the modified mapping model. The data input by the system user is processed in an input/output data processing unit, and shape model data of the object to be analyzed is produced in a shape model producing unit. Based on the division information input by the system user, a finite element mesh is generated on the shape model by a finite element mesh producing unit. The line segment allocation directions of the recognition model are displayed in a recognition model producing unit, so that an allocation direction can be modified in a recognition model modifying unit by the system user in a interactive mode. The finite element model data produced from the optimized recognition model is registered into a database.