Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.593417
Title: Detection problems with the displays of radioisotope imaging devices
Author: Sharp, Peter F.
Awarding Body: University of Aberdeen
Current Institution: University of Aberdeen
Date of Award: 1973
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
The final interpretation of the image produced by a radioisotope imaging device depends upon a human observer. The object of this investigation is to see what factors influence the observer's response, The interpretation of a display record is defined as the detection of a perturbation in an otherwise normal pattern, distribution, of radioisotope. For the purpose of this investigation the simplest normal pattern is chosen, that of a uniform distribution of radioisotope. The perturbation, or target, consists of a circular area of radioisotope, 2 cm. in diameter, superimposed on the normal pattern, or background. The observer's ability to detect the target is measured by his visual response i.e. the percentage of the number of times he manages to detect the target. Three types of displays are investigated: a gamma-camera oscilloscope display, a multi-element display and a rate-meter controlled display with colour-coded statistical bonds. The test-pattern is looped on each display and the resulting display records are shown to a group of observers. The observer's results are plotted in the form of visual response curves, each of which shows visual response as a function of target Intensity for a given value of background intensity. These curves are measured for a series of values of the background intensity. The visual response curves are found to be ogival in shape. This is due both to the day-to-day variation in the responses given by an observer and to the fact that the visual appearance of a target is not completely defined by its Intensity measured in units of counting-density or counting-rate. The target intensity that produces a visual response of 50% is taken as representative of the detectability of the target at a particular background intensity. The ratio of this target intensity to the background intensity is called the 50% contrast. With the oscilloscope display the 30% contrast was found to decrease with increasing background intensity. The two other displays differ from the oscilloscope display in that they are bend-limited. There are only a limited number of different 'intensity' unite with which to display the information. The multi-element display has 8 different grey shades end the counting-rate display 6 colours. Therefore a range, or band, of values of the counting-density or counting-rate is assigned to each of the display 'intensity units. In the multi-element display the position of the limits defining each band is fixed. This results in e saw-tooth shaped variation of the 50% contrast with background intensity. However on the rate-meter controlled display the position of the bands can be altered. With the bends always positioned symmetrically about the mean background intensity the 50% contrast decreases monotonically with increasing background intensity. The performances of the displays are intercompared. The performance of a display is expressed by the surface defined by the axes showing visual response, the ratio of the target-to-background intensity, and the background intensity. A good display is one which has low values of the target-to-background ratio, for a given visual response, end which shows a rapid rise in visual response with increasing target intensity. The multi-element and counting-rate displays are both shown to be superior to the oscilloscope display end comparable to each other in performance. A theory is also proposed to explain the manner in which such perturbations might be detected by the visual system. It is suggested that the observer identifies as the perturbation any one area whose intensity is much greater than that of any other area of comparable size found elsewhere in the background. This contrasts with the statistical theories which compare the target with the average of the signals from the background. The limitations of this way of measuring display performance are indicated and, in conclusion, further areas of work are suggested.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.593417  DOI: Not available
Share: