Digitized images: what type of grey scale should one use?

One of the issues faced by engineers when designing a system which records an external event and represents it in the form of a digitized image on VDU screen is which type of grey scale to use. An experiment is described which compares, in a simulated digitized image, the effect of a linear and a logarithmic grey scale on the detectability of a straight-line signal embedded in visual noise. It was found that both bright and dark signals were detected more easily with the linear scale. A signal detection theory analysis was carried out to compare human performance with that of an 'ideal' observer who performed the detection task with a filter spatially matched to the signal. It was found that the model of performance for this ideal observer accounted well for the results provided the assumption of a linear transformation of luminance was made. The analysis showed that the superiority of the linear over the logarithmic grey scale was simply due to the higher signal-to-noise ratio of the signals in the former.     

Modelling visual detection: Luminance response non-linearity and internal noise

Two experiments that investigate the effect of various display factors on the detectability of a thin line signal in random visual noise are described. Three statistical decision models are described, together with their ability to account for the results. The first is an “ideal detector” model, the second an “energy integrator” model, and the third a model based upon the operation of retinal ganglion cells which incorporates a gain control mechanism. The ideal detector model fails to give a good account of human performance, whereas the other two models provide a good fit to the data. The digital Laplacian with gain control model has the slight advantage over the energy integrating model in being able to account for a small superiority in the detection of dark as opposed to bright signals. Finally, both models require the inclusion of an estimate of the internal noise of the human visual system to account for the pattern of performance observed under changing conditions of display contrast.     

On the interference of task-irrelevant hue variation on texture segmentation.

Although natural images often include discordant information about object boundaries, the majority of research on texture segmentation has involved variation along a single dimension, e.g. colour, orientation, size. In this study, we examined orientation-based texture segmentation in the presence and absence of task-irrelevant colour variation. Previously, it had been shown that orientation-based texture segmentation was impaired if the elements, normally of one colour, were randomly allocated one of two colours (Morgan et al, 1992 Proceedings of the Royal Society of London, Series B 248 291-295). We found that this interference disappeared, however, when the spatial pattern of the colour variation was regular, as opposed to random, and when the elements were randomly positioned. We consider four models of how relevant and irrelevant texture information might combine to produce the interference effect, with special regard to these new findings. None of the models could account for the dependency of the interference effect on the spatial arrangement of colour and orientation in the texture. We suggest that inter-element separation and spatial-frequency selectivity are critical variables in the interference effect.