Novel Stereo Vision Test for Far Distances Measuring Perception Time as a Function of Disparity in a Virtual Environment

Abstract

 

Purpose:

 

To evaluate performance of stereo vision for far distances by measuring perception time as a function of disparity in a virtual environment.

 

Methods:

 

Our already presented stereo vision test was extended to evaluate far stereo vision in human subjects. The test setup consists of a polarized dual projection system for 3D simulation producing a virtual environment. To obtain lower disparities we use full HD projectors (1920 x 1080p). We included a depth sensor and developed an automated gesture detection to provide an intuitive and fast test control. We display four disks of the same size at the same disparity in arc seconds (’’). In each iteration one disk is randomly shown to be in front of the others by a certain disparity difference. The subject has to identify the front disk by pointing on it using our gesture control (see figure). The time to detect the correct disk was measured. In five iterations we showed the disparity differences 300’’, 225’’, 150’’, 75’’, and 15’’. This enables the evaluation of perception time as a function of decreasing disparity. We tested 10 healthy subjects (17-36 years) at a distance of six meters. We classified a given disparity to be perceptible by the subject, if the subject was able to identify the front disk at least four out of five times.

 

Results:

 

The average perception times for correct decisions varied from 1234 ms to 2951 ms in the range of 75’’-300’’ disparity and from 2358 ms to 3877 ms with the disparity of 15’’ (see diagram). In average the perception times decreased with increasing disparity and seemed to saturate in the range of 75’’-300’’. A clear gain of perception time is observable at 15’’. Two subjects were not able to detect the correct disk and were removed from the statistics.

 

Conclusions:

 

The presented test is a novel tool to analyze stereo vision in far distances by measuring perception time as a function of disparity using a virtual environment.  

 

Keywords: binocular vision/stereopsis • depth • clinical (human) or epidemiologic studies: systems/equipment/techniques