|
described for fixated vision in earlier chapters. In an empty field situation - that
is, where the background scene contains little structure - the jumps, or
voluntary saccades, are usually considered to be random (e.g. Koopman and
Krendel & Wodinsky). Conversely, for structured backgrounds, they are often
found to be concentrated in areas of high interest, as might be expected (e.g.
Enoch and Williams). Each dwell period is conventionally referred to in search
modelling as a glimpse. The glimpse frequency is usually of the same order as
that for involuntary saccades - that is, about three per second on average -although
there is some evidence that, for tasks which are felt to be difficult, this
frequency drops to around one per second (e.g. Enoch). At each and every
glimpse there exists a probability either that the gaze will be centred on the
object - yielding a fovea1 single glimpse viewing situation - or that the object
may be seen anywhere within a larger or smaller region of peripheral vision as
determined by the viewing conditions. As shown in Section 4.8, the retinal
position on which the object falls on a particular glimpse will modify greatly the
chances of seeing the object on that glimpse. Furthermore, it seems in practice
that, if there is a reasonably high confidence of having seen an object in the
peripheral field at a given glimpse, the next glimpse will be drawn towards that
object, thus strengthening the effect of the stimulus and allowing confirmation
or rejection as being the object of interest.
In this chapter we shall be concerned with the rudimentary search
situation - the detection of an object in an unknown position in a plain field.
Other forms of search - and in particular that of searching a structured field for
a particular object amongst similar (or confusible) objects - will be dealt with in
Section 13.3.
|
