Design Criteria for Reducing 'Popping' in Area-of-Interest Displays: Preliminary Experiments

摘要

A helmet-mounted display being procured for the Navy's Visual Technology Research Simulator will display two computer-generated images of the scene to be viewed by the trainee pilot, the first inset inside the second, following eye direction and showing a higher level of detail version of the scene than the second, low resolution, low detail, wide angle image. Experiments in which spatially and temporally varying sinusoidal bar patterns are displayed on an Optronix Vision Tester suggest that a contrast blend region around the inset will be effective in suppressing the 'popping' of objects and image detail in and out of view as they move across the inset boundary. These experiments also suggest a restriction upon scene modeling which may be helpful in mitigating popping--namely that the largest size of detail associated with a form (the external border between form plus shadow, and background) be changed least. (Internal contours are added or subtracted first.) In further experimentation, locus of attention is shown to have little effect upon perceived popping, but adaptation may have large effects.