Dynamic virtual convergence for video see-through head-mounted displays: maintaining maximum stereo overlap throughout a close-range work space

We present a technique that allows users of video see-through head-mounted displays to work at close range without the typical loss of stereo perception due to reduced nasal side stereo overlap in most of today's commercial HMDs. Our technique dynamically selects parts of the imaging frustums acquired by wide-angle head-mounted cameras and re-projects them for the narrower field-of-view displays. In addition to dynamically maintaining maximum stereo overlap for objects at a heuristically estimated working distance, it also reduces the accommodation-vergence conflict, at the expense of a newly introduced disparity-vergence conflict. We describe the hardware (assembled from commercial components) and software implementation of our system and report on our experience while using this technique within two different AR applications. (Plus color plates).