High-Precision Assessment of Stereoacuity in Virtual Reality with STRIVE (STereo Random-dot Interactive Virtual Environment)
TL;DR
STRIVE is presented, a real-time, depth-map-driven random-dot stereogram framework that can turn arbitrary VR scenes into disparity-isolated environments and establishes a foundation for studying stereovision in dynamic, interactive, and ecologically relevant environments.
Abstract
Accurate measurement of stereoacuity is essential for understanding binocular depth perception and its contribution to visuomotor behavior. Common clinical tests (e.g., Randot, Frisby) do not completely remove potential monocular cues to depth, are constrained by resolution limits, and are difficult to extend as printed media. Here, we present STRIVE, a real-time, depth-map-driven random-dot stereogram framework that can turn arbitrary VR scenes into disparity-isolated environments. This approach provides fine control over disparity, supports arbitrarily complex and interactive stimuli, and enables systematic manipulation of competing depth cues. In an initial test, we evaluated its utility for measuring static stereoacuity thresholds. Thirteen adults with self-reported normal binocular vision completed a three-alternative forced-choice depth discrimination task in which one of three random-dot-defined discs appeared closer than the others. Stereo thresholds were estimated in arcseconds using a 2-up-1-down adaptive staircase followed by a method of constant stimuli. The VR-based assessment yielded well-fitted psychometric functions and reliable threshold estimates across repeated blocks. Thresholds ranged from 6.9 to 28.5 arcseconds, with low within-participant variability (mean within-participant SD = 2.1 arcseconds) and good test-retest reliability (Intraclass Correlation Coefficient, ICC(2,1)) = 0.862. In contrast, Randot thresholds suffered from floor effects and clustered at only three discrete levels, limiting sensitivity to fine individual differences. Together, these findings demonstrate that the proposed method provides a valid and reliable measure of stereoacuity while offering substantially greater precision and experimental flexibility than standard clinical tests. More broadly, this approach establishes a foundation for studying stereovision in dynamic, interactive, and ecologically relevant environments.