Abstract
J. Gould mentioned the following general factors which affect the eye movement during visual search tasks. 1) The nature of the search task. 2) The information the observer has about the target's characteristics. 3) The observer's expectanceies. 4) The observer's past experience. Based on these four factors, the first target is roughly fixed to the position within the peripheral field, and then it is introduced to the fovea which has sharp discrimination. Repeating this procesess, cognitive fields are gradually enlarged. This process is a kind of choice through eye movement, and so targets with higher chances of the search are imagined to be processed and discriminated succesively. As N. H. Mackworth pointed out in his thesis "Visual noise causes tunnel vision", noise factors of visual search tasks during this process are especially the noise around the target, although it has many kinds of noise factors. In this paper, in order to make the effect of many kinds of noise influences to the eye movement clear, models were prepared which has only the factors with no varieties which the targets normally have. With these models, the following items were analysed by using eye movement parameters about disperse ratio of a fixated point, location and mean duration of fixations. 1) In the experiment about factor density and visual search field within S. V. F. (static visual field) and C. V. F. (Cinetic Visual Field), the following matters are concluded. The distribution density is higher in the S. V. F., the cognitive field is narrower, and in the C. V. F., fixation ratio toward the target is lower because visual search tasks are down without fixating the target because of the noise caused by the factor density. 2) In the experiment about visual search task and eye movement, the following matters are concluded. There are large differences of eye movement before and after targets are searched. That is, before targets are searched, scan paths are arranged according to the position of the elements. On the other hand, once targets are searched, scan paths are getting random, because eye movement for visual search for new targets and the eye movement for cognitive targets are mixed up. 3) In the experiment of eye movement in the connection with elements for the objects (distribution, regularity, variety), the following matters are concluded. The distribution of elements in the model is less, the time for cognition is longer. The arrangement of targets are more irregular, and the numbers of contents of the elements are more, quantities of visual search tasks per hour are less, and duration of fixation for cognition between elements are longer. Probability measures according to the quantity of noise can be adopted because the visual search tasks are the process of cognition continuously changing the target by the successive eye movement within duration of fixation per one target.
