A simple microcomputer-based nuclear medicine data processing system design and performance testing
This thesis investigates the feasibility of designing a simple nuclear medicine data processing system based on an inexpensive microcomputer system, which is affordable to small hospitals and to developing countries where resources are limited. Since the main need for a computer is to allow dynamic studies to be carried out, the relevant criteria for choosing the computer are its speed and memory capacity. The benchmark chosen for these criteria is renography, one of the commonest nuclear medicine procedures. The Acorn Archimedes model 310 microcomputer was found to meet these requirements, and a suitable camera-computer interface has been designed. Because of the need for ensuring that the gain and offset controls of the interface are set optimally before connecting to the camera, it was necessary to design a circuit which produces a test pattern on the screen for use during this operation. Having also developed and tested the data acquisition and image display software successfully, atttention was concentrated on finding ways of characterising and measuring the performance of the computer interface and the display device, two important areas which have been largely neglected in the quality control of camera-computer systems. One of the characteristics of the interface is its deadtime. A procedure has been outlined for measuring this by means of a variable frequency pulse generator and also for interpreting the data correctly. A theoretical analysis of the way in which the interface deadtime affects the overall count rate performance of the system has also been provided. The spatial linearity, resolution and uniformity characteristics of the interface are measured using a special dual staircase generator circuit designed to simulate the camera position and energy signals. The test pattern set up on the screen consists of an orthogonal grid of points which can be used for a visual assessment of linearity, while analysis of the data in memory enables performance indices for resolution, linearity and uniformity to be computed. The thesis investigates the performance characteristics of display devices by means of radiometric measurements of screen luminance. These reveal that the relationship between screen luminance and display grey level value can be taken as quadratic. Characterisation of the display device in this way enables software techniques to be employed to ensure that screen luminance is a linear function of display grey level value; screen luminance measurements, coupled with film density measurements, are also used to optimise the settings of the display controls for using the film in the linear range of its optical densities. This in turn ensures that film density is a linear function of grey level value. An alternative approach for correcting for display nonlinearity is by means of an electronic circuit described in this thesis. Intensity coding schemes for improving the quality of grey scale images can be effective only if distortion due to the display device is corrected for. The thesis also draws attention to significant variations in film density which may have their origins in nonuniformities in the display screen, the recording film, or in the performance of the film processor. The work on display devices has been published in two papers.