The quantification of imaging data gathered by scientific instruments depends on the capability and precision of the analysis techniques and on visual feedback characterized by physiological visual properties. This extension of data analysis to visual tasks requires data display technologies that can present to visual perception the undistorted and complete data information content. Many image enhancing technologies are available but none provides a simple tool for general data presentation that can offset the insensitivity, variability and subjectivity of the physiological process of perception. This limitation is most prominently evident for high precision halftone data that are generated by medical diagnostic imaging or technical imaging modalities. A false reading rate of 30% curtails mammography and, similarly, will trim technical precision imaging tasks. In general, most sensors' high resolution information is perceptually inaccessible and this will hinder progress in all areas of medical, technical and scientific research and their applications. Modem imaging sensors need a physiological display modality that can translate the data into perceptual and quantifiable information. For this translation, the investigators propose a challenging paradigm of information encoding in visual perception that is based on their discovery of an unique new coherence property of 2D data point arrays. Coherence describes the intensity relation between the data points as local areas of similar intensity values. They postulate that coherence patterns represent the perceivable data as characterized by the physiological contrast sensitivity function of the visual system. They propose to generate digital test patterns for a characterization of these perceptual contrast patterns, to develop a display modality for measurement of the patterns and to characterize quantitatively the perceptual task of pattern recognition. The results will verify the existence and determine the meshing of perceptual coherence patterns and will lead to a new way of accessing objectively and quantitatively image information in science, technology, and medical diagnostic imaging as well as digital image communication through networking.

Agency
National Institute of Health (NIH)
Institute
National Center for Research Resources (NCRR)
Type
Exploratory/Developmental Grants (R21)
Project #
5R21RR014258-02
Application #
6188537
Study Section
Special Emphasis Panel (ZRR1-BT-4 (01))
Program Officer
Marron, Michael T
Project Start
1999-09-01
Project End
2001-08-31
Budget Start
2000-09-01
Budget End
2001-08-31
Support Year
2
Fiscal Year
2000
Total Cost
$104,475
Indirect Cost
Name
University of Connecticut
Department
Radiation-Diagnostic/Oncology
Type
Schools of Medicine
DUNS #
022254226
City
Farmington
State
CT
Country
United States
Zip Code
06030
Peters, Klaus-Ruediger; Ramsby, Gale R (2002) Natural display mode for digital DICOM-conformant diagnostic imaging. Acad Radiol 9:1025-35