Showing papers on "Centroid published in 1998"

Method and apparatus for cluster exploration and visualization

Abstract: A method and apparatus for visualizing a multi-dimensional data set in which the multi-dimensional data set is clustered into k clusters, with each cluster having a centroid. Then, either two distinct current centroids or three distinct non-collinear current centroids are selected. A current 2-dimensional cluster projection is generated based on the selected current centroids. In the case when two distinct current centroids are selected, two distinct target centroids are selected, with at least one of the two target centroids being different from the two current centroids. In the case when three distinct current centroids are selected, three distinct non-collinear target centroids are selected, with at least one of the three target centroids being different from the three current centroids. An intermediate 2-dimensional cluster projection is generated based on a set of interpolated centroids, with each interpolated centroid corresponding to a current centroid and to a target centroid associated with the current centroid. Each interpolated centroid is interpolated between the corresponding current centroid and the target centroid associated with the current centroid. Alternatively, the intermediate 2-dimensional cluster projection is generated based on an interpolated 2-dimensional nonlinear cluster projection that is based on the selected current centroids and the selected target centroids.

Method and system for clustering data in parallel in a distributed-memory multiprocessor system

Abstract: A method, apparatus, article of manufacture, and a memory structure for clustering data points in parallel using a distributed-memory multi-processor system is disclosed. The disclosed system has particularly advantageous application to a rapid and flexible k-means computation for data mining. The method comprises the steps of dividing a set of data points into a plurality of data blocks, initializing a set of k global centroid values in each of the data blocks k initial global centroid values, performing a plurality of asynchronous processes on the data blocks, each asynchronous process assigning each data point in each data block to the closest global centroid value within each data block, computing a set of k block accumulation values from the data points assigned to the k global centroid values, and recomputing the k global centroid values from the k block accumulation values.

A method for using three points to define a 2d mr imaging section

Abstract: A method and apparatus for producing an imaging plane on an image of a structure of interest, such as an anatomical structure, positioned in an MRI system. An operator interactively pages through real-time, planar sections of the structure of interest. Using an input device, the operator selects three separate points in a planar section of the structure under study. Within approximately one second of selection of the third point, the method of the present invention determines the imaging plane containing the three selected points, determines the centroid of the imaging plane centered on a triangle defined by the three selected points, sends such imaging geometry and in-plane offsets of the imaging plane directly to the MRI system to generate a new imaging plane optimally positioned with respect to the selected points on the structure of interest and displaying such new imaging plane. The operator can also selectively maneuver the imaging plane on the image of the structure of interest. The operator uses a graphical user interface in conjunction with the input device and a display screen for producing the imaging plane on the structure of interest. Such graphical user interface is referred to as a three point tool.

VLSI Implementation of Motion Centroid Localization for Autonomous Navigation

Abstract: A circuit for fast, compact and low-power focal-plane motion centroid localization is presented. This chip, which uses mixed signal CMOS components to implement photodetection, edge detection, ON-set detection and centroid localization, models the retina and superior colliculus. The centroid localization circuit uses time-windowed asynchronously triggered row and column address events and two linear resistive grids to provide the analog coordinates of the motion centroid. This VLSI chip is used to realize fast lightweight autonavigating vehicles. The obstacle avoiding line-following algorithm is discussed.

Robust method for finding registration marker positions

Abstract: A method and system for detecting registration markers having the shape of an `X` on a scanned test target used to calibrate an image processing system. The procedure is impervious to scanning and printing distortion. A captured image is converted to a binary image and then segmented into a list of eight-connected components. A set of features is extracted from each component in the list. An accurate decision rule is executed on each component. If a component is found to be a marker, its centroid is returned as its position.

Unsupervised texture segmentation using multiresolution Markov random fields

Abstract: In this thesis, a multiresolution Markov Random Field (MMRF) model for segmenting textured images without supervision is proposed. Stochastic relaxation labelling is adopted to assign the class label with highest probability to the block (site) being visited. Class information is propagated from low spatial resolution to high spatial resolution, via appropriate modifications to the interaction energies defining the field, to minimise class-position uncertainty. The thesis contains novel ideas presented in Chapter 4 and 5, respectively. In Chapter 4, the Multiresolution Fourier Transform (MFT) is used to provide a set of spatially localised texture descriptors, which are based on a two-component model of texture, in which one component is a deformation, representing the structural or deterministic elements and the other is a stochastic one. Experiments show that the algorithm is efficient in alleviating class-position uncertainty via data propagation across resolutions. However, the blocking artifacts of the segmentation results show that it is preferable to combine both class and position information so as to achieve smoother and more accurate boundary estimation. In Chapter 5, based on the same MFT-MMRF framework, a boundary process is proposed to refine the segmentation result of the region process proposed in Chapter 4. At each resolution, all the image blocks on either sides of the preliminary boundary detected in the region process are treated as potential boundary-containing blocks (PBCB's). The orientation and the centroid of the boundary-segment contained in each PBCB are calculated. The sequence of PBCB's are then modelled as a MRF and the interaction energy between each pair of neighbouring blocks is defined as a function of the 'distance' D between the centroids of the two boundary segments. During the stochastic relaxation process boundary/non-boundary labels are assigned to the PBCB's. Once the algorithm converges, the centroids of the identified true boundary blocks are connected to form the refined boundary which is propagated down to the next resolution for further refinement.

Centroid based Multilayer Perceptron Networks

Abstract: In this study we investigate a hybrid neural network architecture for modelling purposes. The proposed network is based on the multilayer perceptron (MLP) network. However, in addition to the usual hidden layers the first hidden layer is selected to be a centroid layer. Each unit in this new layer incorporates a centroid that is located somewhere in the input space. The output of these units is the Euclidean distance between the centroid and the input. The centroid layer clearly resembles the hidden layer of the radial basis function (RBF) networks. Therefore the centroid based multilayer perceptron (CMLP) networks can be regarded as a hybrid of MLP and RBF networks. The presented benchmark experiments show that the proposed hybrid architecture is able to combine the good properties of MLP and RBF networks resulting fast and efficient learning, and compact network structure.

Automated detection of BB pixel clusters in digital fluoroscopic images

Abstract: Small ball bearings (BBs) are often used to characterize and correct for geometric distortion of x-ray image intensifiers. For quantitative applications the number of BBs required for accurate distortion correction is prohibitively large for manual detection. A method to automatically determine the BB coordinates is described. The technique consists of image segmentation, pixel coalescing and centroid calculation. The dependence of calculated BB coordinates on segmentation threshold was also evaluated and found to be within the uncertainty of measurement.

Extended source effects in astrometric gravitational microlensing

Abstract: Extended source size effects have been detected in photometric monitoring of gravitational microlensing events. We study similar effects in the centroid motion of an extended source lensed by a point mass. We show that the centroid motion of a source with uniform surface brightness can be obtained analytically. For a source with a circularly symmetric limb-darkening profile, the centroid motion can be expressed as a one-dimensional integral, which can be evaluated numerically. We find that when the impact parameter is comparable to the source radius, the centroid motion is significantly modified by the finite source size. In particular, when the impact parameter is smaller than the source radius, the trajectories become clover-leaf like. Such astrometric motions can be detected using space interferometers such as the Space Interferometry Mission. Such measurements offer exciting possibilities for determining stellar parameters, such as stellar radius, to excellent accuracy.

Adaptive suboptimum detection of an optical pulse-position-modulation signal with a detection matrix and centroid tracking

Abstract: In some applications of optical communication systems, such as satellite optical communication and atmospheric optical communication, the optical beam wanders on the detector surface as a result of vibration and turbulence effects, respectively. The wandering of the beam degrades the communication system performance. In this research, we derive a mathematical model of an optical communication system with a detection matrix to improve the system performance for direct-detection pulse-position modulation. We include a centroid tracker in the communication system model. The centroid tracker tracks the center of the beam. Using the position of the beam center and an a priori model of the beam spreading, we estimate the optical power on each pixel (element) in the detection matrix. Using knowledge of the amplitudes of signal and noise in each pixel, we tune adaptively and separately the gain of each individual pixel in the detection matrix for communication signals. Tuning the gain is based on the mathematical model derived in this research. This model is defined as suboptimal, owing to some approximations in the development and is a suboptimum solution to the optimization problem of n multiplied by m free variables, where n, m are the dimensions of the detection matrix. Comparison is made between the adaptive suboptimum model and the standard model. From the mathematical analysis and the results of the comparison it is clear that this model significantly improves communication system performance.

Extended Source Effects in Astrometric Gravitational Microlensing

Abstract: Extended source size effects have been detected in photometric monitoring of gravitational microlensing events. We study similar effects in the centroid motion of an extended source lensed by a point mass. We show that the centroid motion of a source with uniform surface brightness can be obtained analytically. For a source with circularly symmetric limb-darkening profile, the centroid motion can be expressed as a one-dimensional integral, which can be evaluated numerically. We found that when the impact parameter is comparable to the source radius, the centroid motion is significantly modified by the finite source size. In particular, when the impact parameter is smaller than the source radius, the trajectories become clover-leaf like. Such astrometric motions can be detected using space interferometers such as the Space Interferometry Mission. Such measurements offer exciting possibilities of determining stellar parameters such as stellar radius to excellent accuracy.

Method for improving the accuracy in the determination of a waveform center of a waveform signal

Abstract: Prior to performing a centroid calculation on a waveform signal that is discretely sampled at a limited number of sample points, the last sample point (V N , A N ) is eliminated if the magnitude of the amplitude at the first sample point (A 1 ) is greater than the last sample point (A N ), and the difference in magnitude between the first and last sample points (A 1 −A N ) is greater than the difference in magnitude between the second to last sample point and the first sample point (A N−1 −A 1 ). The first sample point (V 1 , A 1 ) is eliminated prior to the centroid calculation if the magnitude of the amplitude at the last sample point (A N ) is greater than the first sample point (A 1 ), and the difference in magnitude between the last and first sample points (A N −A 1 ) is greater than the difference in magnitude between the second sample point and the last sample point (A 2 −A N ). In a second embodiment of the invention, a first centroid calculation is performed using a set of samples in which one side of the waveform signal has the lowest amplitude value sample. Sample values on the side of the waveform initially having the lowest amplitude value are then eliminated until the opposing side of the waveform has the lowest amplitude value sample. A second centroid calculation is then performed and the two centroid calculations are averaged together to arrive at an average centroid calculation. In a third embodiment of the invention, the amplitude components of the waveform sample values are normalized to the lowest amplitude value sample point and a first centroid calculation is performed on the normalized waveform signal. Next, the waveform is normalized to the lowest amplitude value sample point on the other side of the waveform signal and a second centroid calculation is performed. The two centroid calculations are then averaged to provide an averaged normalized centroid calculation.

CLUSTER: Stata module to perform nonhierarchical k-means (or k-medoids) cluster analysis

Abstract: Cluster performs nonhierarchical k-means (or k-medoids) cluster analysis of your data. Centroid cluster analysis is a simple method that groups cases based on their proximity to a multidimensional centroid or medoid. Cluster.do provides an example of usage.

Edge detection method of parts picture, and parts form recognition device

Abstract: PROBLEM TO BE SOLVED: To precisely detect the edge points of a parts image by setting the nozzle center position of a picked up image to be a radial scan original point, radically scanning the picked up image with a scan start radius as a value larger than a nozzle radius, and detecting the edge points on the outline of the parts image. SOLUTION: The center position O of a nozzle in the image pickup image is set as the radial scan original point. The radius (scan start radius) for starting scanning is set to be a value larger than the radius of the nozzle in the image pickup image. A differential processing is executed while the image pickup image is radically scanned. Thus, a point 104 (edge point) on the outline of the partial image 101 and the gradient direction (edge vector) α of the point are detected. A detected xy-coordinate value is transformed into the coordinate value (r, θ) of a polar coordinate system (r, θ) with a centroid position G as the original point. Thus, the coordinate values (r, θ) of the respective edge of the parts image and edge vector α are extracted.

Using centroid covariance in target recognition

Abstract: An automatic target recognition algorithm for low quality imagery is reported. Compact shaped targets are represented by their 2D silhouettes. Associated with each point on the silhouette, there is a direction roughly perpendicular to the local segment of the silhouette. The location of each silhouette point is assumed to be perturbed along that direction. A statistical technique is used to estimate the variance of that perturbation for the silhouette points of the hypothesized target. This variance is then used to estimate the location covariance of the target centroid. Target detection and recognition is based on this covariance. Target scaling, aspect, and rotation are not considered. Experiments on 31 FLIR images give a correct recognition of target identity and target location for 29 of the 31 images.

Analyzing the probabilities of natural resting for a component with a virtual resting face

Abstract: This paper introduces a new approach to predicting the natural resting behavior of part of irregular prismatic cross-sectional shape. The methods proposed by Boothroyd, Redford, Poli, and Murch (1972) and Boothroyd and Ho (1977), successfully analyzed parts whose centers of gravity were located at the geometric center. The hypothesis presented here employs the centroid solid angle of parts of irregular cross-sectional shape with virtual resting faces, such as the L-section. It postulates that the probability of the respective natural resting surfaces of a part coming to rest is directly proportional to the centroid solid angle and inversely proportional to the height of the center of gravity from that surface. The proposed hypothesis is compared with the Energy Barrier Method and experimental results obtained from drop-tests of the L-shaped part. The theoretical results agree well with the experimental results, with the maximum deviation not exceeding 6 percent.

A Stochastic Technique for the Removal of Artifacts in Compressed Images and Video

Abstract: The perceived quality of images and video sequences reconstructed from low bit rate compressed bit streams is severely degraded by the appearance of coding artifacts. This chapter introduces a technique for the post-processing of compressed images based on a stochastic model for the image data. Quantization partitions the transform coefficient space and maps all points in a partition cell to a representative reconstruction point, usually taken as the centroid of the cell. The proposed technique selects the reconstruction point within the quantization partition cell which results in a reconstructed image that best fits a non-Gaussian Markov Random Field image model. This approach results in a convex constrained optimization problem that can be solved iteratively. Efficient computational algorithms can be used in the minimization. This technique is extended to the post-processing of video sequences. The proposed approach provides a reconstructed image with reduced visibility of transform coding artifacts and superior perceived quality.

Dense shape and motion from region correspondences by factorization

Abstract: In this paper, we propose an algorithm for estimating dense shape and motion of dynamic piecewise planar scenes from region correspondences using factorization. Region correspondences are used since they are easier to establish and more reliable than either line or point correspondences. The image measurements required are the centroid and area for each region. Singular value decomposition is employed to find the basis of range space of the motion, shape, and surface normal matrices. By imposing model constraints, motion, shape, and surface normal can be recovered only from region correspondences.

Method and system for determining component dimensional information

Abstract: A method and system for determining component dimensional information in a component placement system. The component information is derived from the pick and place equipment which is used to place the components on the substrate. The component is grasped at substantially the centroid position and moved over the aperture of a camera. The edges of the component are located, and from the edges an accurate location of the centroid may be obtained. The centroid, as well as the length, width and thickness of the component is stored in a database for use during placing of the component. Other features derived include orientation indicia obtained from component features which are viewed by the camera. During placement, all features associated with the component necessary to accurately locate the component on a substrate are obtained from the database.

Method, device for determining number of moving objects to be arranged and programming system for the same

Abstract: PROBLEM TO BE SOLVED: To objectively determine the number of moving objects to be arranged at the destination of arrangement so as to minimize the time for visiting respective spots in an object area from the arrangement destination of moving object. SOLUTION: The object area is divided into plural small blocks by a data preparation part 100 for determination, and the point of centroid representative among the respective small blocks is calculated. Next, the shortest road distance between each centroid point and a fire station and the fire station closest to the respective centroid points are extracted, an arrangement estimate pattern estimating the arrangement of vehicles is lined up based on the number of vehicles enabled as mobile object machines, the shortest road distance to the estimated fire station from the centroid point in each pattern is divided with the number of vehicles estimated for that fire station by an arithmetic processing part 116, this operation is performed for all the centroid points, its total sum is calculated, further, the total sum is calculated concerning the respective estimated arrangement patterns, the estimated arrangement pattern to minimize the total sum is preserved and outputted in the manner of update and according to the preservation pattern to be outputted, the fire station to arrange vehicles and the number of vehicles to be arranged are determined.

Individual identification method/device

Abstract: PROBLEM TO BE SOLVED: To improve the precision of individual identification and to remarkably reduce storage capacity by operating the plural feature amounts of a centroid and an inclination as the features of silhouette pictures by means of a feature operation means and storing a reference function itself calculated by adding and averaging a feature function with the number of samples by means of a difference degree operation means in a storage device. SOLUTION: The silhouette picture of a pedestrian is extracted (S112) and a feature operation processing is executed from the silhouette picture (S114). Namely, a centroid operation processing calculating a centroid and an inclination operation processing calculating an inclination are executed at every prescribed number of silhouette pictures. A function of time on the centroid and the inclination, which is obtained in the operation processing, is Fourier-transformed at high speed into the function of a frequency (S116). A feature function generation processing extracting the feature function of the prescribed number of samples is executed (S118). The addition average of the respective feature functions is obtained and it is stored in a file as the reference function of a specified individual (S128).

Centroid, centroid from edge vectors, and shape descriptor using only boundary information

Abstract: A simple algorithm is presented to obtain the area and the centroid of a plane homogeneous feature, by using only the coordinates of the boundary points. The algorithm also allow higher order moments to be calculated. From such centroid, and the centroid from edge vectors, we defined a shape descriptor invariant to Rotation, Scale, and Translation transformations. This parameter is able to characterize features with asymmetric distribution of roughness or reentrants.

Curve curvature estimating device

Abstract: PROBLEM TO BE SOLVED: To provide a curve curvature estimating device which can estimate the curvature of a curve with accuracy by taking the centroid slip angle of a vehicle into consideration SOLUTION: A curve curvature estimating device is provided with an image pickup means 2 which is mounted on a vehicle and picks up the image of the running road of the vehicle in front of the vehicle, an image converting means 2A which converts the image into a top-view image, and a curvature calculating means 30 which finds the angle θ between the axial directions of the running road and the main body of the vehicle at a point which is in front of the reference point of the vehicle by a prescribed distance L based on the image information from the converting means 2A and calculates the curvature of the curved section of the running road from the angle θ and distance L The estimating device is also provided with a correcting means 31 for centroid slipping angle which corrects the value of the angle θ or distance L in accordance with the centroid slip angle of the vehicle which occurs when the vehicle makes a turn at the curved section at the time of calculating the curvature is provided

Method for partially selecting three-dimensional shape dividing model and information processor

Abstract: PROBLEM TO BE SOLVED: To easily set border conditions by easily extracting the area to set the border conditions even in the case of a border condition setting area having a complicated shape SOLUTION: A selection reference point P is set (step S31) The surface turned toward the direction observing the relevant surface from this selection reference point P is selected out of the surfaces of a divided model in three- dimensional shape (steps S32-S35) Namely, first of all, all the surfaces of the divided model are extracted as model surface data F0 (step S32) Continuously, concerning all the respective model surface data F0 (step S35), it is discriminated whether the inner product of a direction vector (a) having the direction of connecting the selection reference point P from a centroid point on the surface and a normal vector (b) on this surface becomes positive or not (step S33) and when it is positive, that surface is registered as model surface data F1 (step S34)

Method and device for detecting weld line

Abstract: PROBLEM TO BE SOLVED: To provide a method capable of stably detecting a weld line and greatly improving detecting precision, even if the shape or material changes of the weld line to be detected exist. SOLUTION: In calculating the centroid for variable-density projection values in the horizontal and vertical directions, the values are segmented at a set value determined by the maximum and the minimum for the inflection point of the variable-density projection values that are obtained from their processing part 8; with the centroid calculation performed for the segmented variable- density projection values in the horizontal and vertical directions; the position of a picture element is determined corresponding to the horizontal and vertical centroid positions obtained from the calculation result; and the position of a weld line is determined from that position of the picture element. An image processor 2 is provided with horizontal and vertical variable-density projection processors 8, 10, 9, 11 and with an arithmetic processor 13 for calculating the centroid based on the projection processing.