Eye Movement Measurement in the Assessment and Training of Visual Performance

By analysing visual search behavior knowledge may be guided to improve the performance of search tasks through human engineering design and training. Two ideas have suggested the need for further analysis of visual search behavior. The first is the suggestion of Krendel and Wodinsky (1960) who supposed that visual search strategies could be inferred from the underlying distributions of search times. They postulated that an exponential distribution is indicative of a random strategy where successive fixations are independent of each other and therefore will overlap each other. A linear distribution, on the other hand, indicates a systematic search strategy without overlapping of fixations. This strategy is called an exhaustive search, which may reduce the mean search time in a trial considerably (Howarth and Bloomfield, 1968). The other idea is related to the size of the useful field of view or the visual lobe area. It is defined as the peripheral area around the central fixation point from which specific information can be extracted. The size of the useful visual field depends on the context of a target; e.g. it is reduced when the target is embedded in a complex background or surrounded by irregularly positioned nontarget items (Brown and Monk, 1975).