Browse Prior Art Database

METHOD AND APPARATUS FOR EVALUATING THE QUALITY OF MEDICAL IMAGING SYSTEMS AND COMPONENTS USING MODEL OBSERVERS AND COMPUTER AIDED DETECTION ALGORITHMS

IP.com Disclosure Number: IPCOM000130693D
Publication Date: 2005-Nov-01

Publishing Venue

The IP.com Prior Art Database

Related People

Jian Yang: AUTHOR [+4]

Abstract

This disclosure describes methods and apparatus for evaluating the image quality of medical imaging systems using a mathematical model observer to estimate human visual talk performance, or by measuring the efficacy of a computer aided detection (CAD) algorithm. These automated image quality monitoring methods enable assessment of full-system image quality and emulation of equipment component behaviors to identify sources of image quality degradation.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 5% of the total text.

Method and Apparatus for Evaluating the Quality of Medical Imaging Systems and Components Using Model Observers and Computer Aided Detection Algorithms

Authors:

Jian Yang, Richard B. Wheeler, Brian W. Keelan, Karin Töpfer

This disclosure describes methods and apparatus for evaluating the image quality of medical imaging systems using a mathematical model observer to estimate human visual task performance, or by measuring the efficacy of a computer aided detection (CAD) algorithm. These automated image quality monitoring methods enable assessment of full-system image quality and emulation of equipment component behaviors to identify sources of image quality degradation.

Image quality (IQ) tools have been used in medical imaging areas to evaluate the quality of image capturing and displaying devices.  Existing image quality tools evaluate physical properties of the image systems and components, such as resolution, noise, artifacts, tonescale, geometrical distortion, etc.

Medical images are used by medical specialists to identify whether clinical targets, such as tumors and lesions, exist in a patient’s body.  The above-mentioned physical properties often do not correlate well with diagnostic performance.  It has been suggested to use target detection performance to evaluate medical image quality and to use mathematical observers to evaluate target detection performance (Meyers et al., 1990; ICRU report 54, 1996; Abbey et al., 1997; Eckstein, 2001).

Using target detection performance to evaluate the image quality of medical imaging systems is now a well-accepted approach in medical imaging community.  For example, the visibility of three types of disease-like features (fibers, microcalcification groups, and masses) in ACR phantom images are used to evaluate the image quality of mammography systems in the United States, and the dot visibilities of CDMAM phantom images are used to evaluate the image quality of mammography systems in European countries. The normal procedure requires 3 or more human observers to identify the concerned targets in each case, which is costly and time consuming. Furthermore, the obtained results contain substantial observer variability. To avoid these disadvantages, several investigators have tried using automated computer analysis to replace human observers in the evaluation loops.

Chakraborty (1996; 1997) developed a computer analysis method, known as CAMPI, to evaluate the quality of ACR phantom images.  The method includes estimating the perfect images of ACR phantom targets and the actual phantom target images.  The image quality is then evaluated using several metrics to indicate the difference between the test and perfect target images. The most useful metric is signal-to-noise ratio (SNR), which is the estimate of the ideal observer predication for white noise or nonprewhitening model observer for other noise statistics.  Therefore, the CAMPI tool simulates some aspects of ideal observer performance.

A...