Empirically defined regions of influence for clustering analyses

Osbourn, G. C.; Martínez, Rosa

doi:10.1016/0031-3203(95)00032-u

Cited by 33 publications

(22 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this approach, the VERI shape was regarded as an empirical entity to be fitted to data on human cluster judgments. The VERI ROI shape was first discovered by Osboum and Martinez through a set of psychophysical studies of human cluster perception [1,2]. That work has confirmed that VERI provides a reasonable model of visual cluster judgments obtained from a consensus of human subjects with normal visual perception.…”

Section: Introductionsupporting

confidence: 50%

“…This follows from the apparent need to examine every pair of points for potentiaI clustering, and the apparent need to consider every remaining point in the data set as a potential inhibitor of each pair. We have previously described [1,6] an 0(N2) cluster implementation and an O(Nti. *N@,t) VERI pattern recognition algorithm that exist if the total number of pairwise VERI clusterings scales as O(N) when only the first nearest neighbors of each vector are considered as inhibitors.…”

Section: Introductionmentioning

confidence: 99%

“…The visual empirical region of influence (VERI) is a specific version [1,2] of the more general region of influence (ROI) concept from graph theory [3]. ROIS are twodimensional (2-D) shapes that are defined with respect to pairs of vectors.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Some Provable Properties of VERI Clustering

Osbourn¹

2000

View full text Add to dashboard Cite

Section: Introductionsupporting

confidence: 50%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Some Provable Properties of VERI Clustering

Osbourn¹

2000

View full text Add to dashboard Cite

“…In Figure 2, there are fourteen data points, each point has three dimensions. There are four classes of data (1,2,3,5). All the data in a class must be stored contiguously.…”

Section: Data Formatmentioning

confidence: 99%

A Visual Empirical Region of Influence Pattern Recognition Tool for Leave-One-Out Data Analysis

Martinez¹

2002

View full text Add to dashboard Cite

In previous research at Sandia National Laboratories new pattern recognition (PR) algorithms based on a human visual perception model were developed. We named these algorithms Visual Empirical Region of Influence (VERI) algorithms. This document describes a graphical user interface tool developed to control the VERI algorithm and a visualization technique that allows users to graphically animate and visually inspect multi-dimensional data after it has been classified by the VERI algorithms. The VERI Interface Tool was written using the TcVTk Graphical User Interface (GUI) programming language, version 8.1. Although the TcVTk packages are designed to run on multiplatforms, we have concentrated our efforts to develop a user interface for the ubiquitous DOS environment. The VERI algorithms are compiled, executable programs. The commands to execute VERI algorithms are quite difficult to master when executed from a DOS command line. The algorithm q u i r e s several parameters to operate correctly. From our own experiences we realized that if we wanted to provide a new data analysis tool to the PR community we would have to make the tool powerful, yet easy and intuitive to use. That was our motivation for developing the VERI Interface tools. This document illustrates step-by-step examples of how to use the interface. This interface runs VERI in Leave-One-Out mode using the Euclidean metric. For a description of this type of data analysis, and for a general pattern Recognition tutorial, refer to our website at: htto://www.sandia.eov/imrVXVisionScience/Xusers.htm.

show abstract

“…The initial clustering algorithm we developed in previous work used a straight foirward approach of examining all (Om2) pairs) point pair combinations to determine if they cluster (group) together or are inhibited by any third point which falls within the VERI template (1). The algorithm is an OV3) algorithm, which would severely limit the size of data sets which could be operated on in a practical amount of time.…”

Section: Motivation To Improve Algorithm Runtimesmentioning

confidence: 99%

Empirically defined regions of influence for clustering analyses

Cited by 33 publications

References 14 publications

Some Provable Properties of VERI Clustering

Some Provable Properties of VERI Clustering

A Visual Empirical Region of Influence Pattern Recognition Tool for Leave-One-Out Data Analysis

Extending applicability of cluster based pattern recognition with efficient approximation techniques

Contact Info

Product

Resources

About