Visualization in Scanning

Scanning is essentially a method of making the unknown known, and for the vast majority of scanning work this means making the invisible visible. Visualization is developed as a process involving a model of eye and brain function, display performance and counter performance combined with the spatial frequency components in the object scanned. The performance of a scanner or gamma camera depends in part on its display system and experiments have been performed to determine the fractional change in count-rate which can be perceived by the eye from colour, photoscan and cathode-ray tube displays, for both circular and line sources of radioactivity. For circular sources, colour and photoscan displays are capable of making 10% increases in count-rate perceptible in backgrounds ranging from 5 to 100 counts/s for sources of 1 to 4 cm diameter, whilst cathode-ray tube and black-mark displays show only 20% at best. The results for line sources are also given. In testing for the statistical significance of count-rate differences on a scan, the counts are usually compared for equal areas. On this basis the cathode-ray tube display is only able to show five standard deviations of difference whereas colour and photoscan displays show from 2.5 to 3 standard deviations, levels not usually accepted statistically as significant. It is suggested that if the eye examines the areas surrounding the area of interest then the significance of a count-rate difference may be established more accurately and in general will be increased, thereby accounting for the visualization of apparently 'insignificant' regions. This theory is used to predict the areas of count density which should be visually perceived in the presence of noise, and extended to consider practical display conditions, which have been described in this paper. The performance of a scanner can be considered in terms of its modulation transfer function (derived from a line spread function). Fourier analysis has been performed of simple objects, (e.g. discs, lines, etc.) so that the shape is defined in terms of fundamental and harmonic spatial frequencies. The performance of a counter and collimator for a selected object is then calculated from the MTF values and compared directly with phantom measurements. (author)