Computer-aided generation of performance measures for man-machine systems

Author(s):  
Edward M. Connelly ◽  
Patricia A. Knoop ◽  
Francis J. Bourne ◽  
Diane G. Loental
2005 ◽  
Vol 27 (3) ◽  
pp. 181-198 ◽  
Author(s):  
Ulrich Scheipers ◽  
Christian Perrey ◽  
Stefan Siebers ◽  
Christian Hansen ◽  
Helmut Ermert

The application of the receiver operating characteristic (ROC) curve for computer-aided diagnostic systems is reviewed. A statistical framework is presented and different methods of evaluating the classification performance of computer-aided diagnostic systems, and, in particular, systems for ultrasonic tissue characterization, are derived. Most classifiers that are used today are dependent on a separation threshold, which can be chosen freely in many cases. The separation threshold separates the range of output values of the classification system into different target groups, thus conducting the actual classification process. In the first part of this paper, threshold specific performance measures, e.g., sensitivity and specificity; are presented. In the second part, a threshold-independent performance measure, the area under the ROC curve, is reviewed. Only the use of separation threshold-independent performance measures provides classification results that are overall representative for computer-aided diagnostic systems. The following text was motivated by the lack of a complete and definite discussion of the underlying subject in available textbooks, references and publications. Most manuscripts published so far address the theme of performance evaluation using ROC analysis in a manner too general to be practical for everyday use in the development of computer-aided diagnostic systems. Nowadays, the user of computer-aided diagnostic systems typically handles huge amounts of numerical data, not always distributed normally. Many assumptions made in more or less theoretical works on ROC analysis are no longer valid for real-life data. The paper aims at closing the gap between theoretical works and real-life data. The review provides the interested scientist with information needed to conduct ROC analysis and to integrate algorithms performing ROC analysis into classification systems while understanding the basic principles of classification.


1974 ◽  
Vol 18 (3) ◽  
pp. 359-367 ◽  
Author(s):  
Edward M. Connelly ◽  
Patricia A. Knoop ◽  
Francis J. Bourne ◽  
Diane G. Loental

The design, application, and evaluation of man-machine systems are limited by our ability to measure system performance in a reliable and sensitive manner. Without adequate performance measures, there is no way to test system designs, to plan and execute training systems, or to effectively evaluate operational systems. Typically, measures are manually produced by selecting a set of candidate performance measures which are subsequently tested for reliability and validity. Since the measurement value of a given candidate measure is not known until these tests are complete, this process, which may be an iterative process, can be both time consuming and costly. Also, since only a few candidate measures can be investigated manually, there is a high probability that superior measures are not even considered. Automating at least some of the manual operations required can result in improved performance measures in less time and at lower cost. The performance measurement generating processor described in this paper accepts demonstration data representing various levels of performance, and under user control, analyzes the data to provide candidate performance measures. The processor also conducts validation tests and orders candidate measures according to their measurement value. Output from the FORTRAN IV processor includes results from validation tests and specifications for objective performance measures.


1987 ◽  
Vol 109 (1) ◽  
pp. 87-94 ◽  
Author(s):  
S. Dubowsky ◽  
J. F. Deck ◽  
H. Costello

An effective method is presented for modeling the full three-dimensional dynamics of high performance spatial machine systems, including the vibrations of their links, supporting structure and enclosures, and impacts in their connection clearances. The method combines four-by-four matrix modeling techniques with finite-element techniques. The system’s elements can be described in rich detail, including detailed models of the internal structure of the connections, yet the dynamic calculations are performed efficiently. The method is formulated for implementation on computer-aided design systems. Examples are presented.


Author(s):  
Edward M. Connelly ◽  
Francis J. Bourne ◽  
Diane G. Loental ◽  
Patricia A. Knoop

Author(s):  
Mark Ellisman ◽  
Maryann Martone ◽  
Gabriel Soto ◽  
Eleizer Masliah ◽  
David Hessler ◽  
...  

Structurally-oriented biologists examine cells, tissues, organelles and macromolecules in order to gain insight into cellular and molecular physiology by relating structure to function. The understanding of these structures can be greatly enhanced by the use of techniques for the visualization and quantitative analysis of three-dimensional structure. Three projects from current research activities will be presented in order to illustrate both the present capabilities of computer aided techniques as well as their limitations and future possibilities.The first project concerns the three-dimensional reconstruction of the neuritic plaques found in the brains of patients with Alzheimer's disease. We have developed a software package “Synu” for investigation of 3D data sets which has been used in conjunction with laser confocal light microscopy to study the structure of the neuritic plaque. Tissue sections of autopsy samples from patients with Alzheimer's disease were double-labeled for tau, a cytoskeletal marker for abnormal neurites, and synaptophysin, a marker of presynaptic terminals.


Author(s):  
Greg V. Martin ◽  
Ann L. Hubbard

The microtubule (MT) cytoskeleton is necessary for many of the polarized functions of hepatocytes. Among the functions dependent on the MT-based cytoskeleton are polarized secretion of proteins, delivery of endocytosed material to lysosomes, and transcytosis of integral plasma membrane (PM) proteins. Although microtubules have been shown to be crucial to the establishment and maintenance of functional and structural polarization in the hepatocyte, little is known about the architecture of the hepatocyte MT cytoskeleton in vivo, particularly with regard to its relationship to PM domains and membranous organelles. Using an in situ extraction technique that preserves both microtubules and cellular membranes, we have developed a protocol for immunofluorescent co-localization of cytoskeletal elements and integral membrane proteins within 20 µm cryosections of fixed rat liver. Computer-aided 3D reconstruction of multi-spectral confocal microscope images was used to visualize the spatial relationships among the MT cytoskeleton, PM domains and intracellular organelles.


Sign in / Sign up

Export Citation Format

Share Document