Cognitive Control of Depth/Transparency

Two experiments examined the ability of observers to control the relative depth of multistable transparency displays and the relationship of that control to default biases for attended surfaces. These biases were assessed by a task that induced particular distributions of attention over different surfaces of the display. The first experiment concerned asymmetrical displays in which one percept was fully transparent and the other percept was partially transparent; the second experiment concerned completely ambiguous displays in which both percepts were fully transparent. Three different control strategies were evaluated using observers who had previously demonstrated good or poor control of structure from motion displays.