A poster presented at the third meeting of the Vision Sciences Society, Sarasota, Fl., U.S.A., May 9-14.
In most visual search experiments, only one target is specified.Many real-world searches involve looking for multiple targets simultaneously. ("Find me either a phone booth or a policeman.")
Some models of top-down search suggest that specifying multiple targets from the same feature dimension simultaneously should be difficult or impossible.
How efficient is search for multiple targets?
Previous experiments with multiple disjunctive targets
Quinlan and Humphreys (1987), Treisman (1988): efficient search for two distinct targets, defined by separate unique features.
D'Zmura (1991): two distinct coloured targets can be sought simultaneously in parallel if they are together linearly separable from the distractors.
Other experiments in colour search
Nagy and Sanchez (1990): if the difference between target colour and distractors colours reaches a critical point, then search remains efficient.
Bauer et al. (1998): linear separability of the target from the distractors determines search performance.
Poisson and Wilkinson (1992): grouping has an important role in subset selection.
Wolfe et al. (1990): more difficult to search for a conjunction of same-dimension features than for conjunctions of features from different dimensions; perhaps because it requires union rather than intersection of feature subsets.
Search for multiple differently coloured targets is inefficient.This has implications for real-life situations such as airport security X-ray scanning, where scanners search for multiple items simultaneously.
Preliminary experiments with search for complex shapes shows a very different pattern.We are also conducting experiments with orientation, and with combinations of colour and orientation.
This work was supported by a grant from U. S. Transportation Security Administration.