In general, there is a stereo anomaly for people who do not perceive the stereo visual depth from images with various disparities. People with ophthalmological problems such as strabismus are categorized as stereo blind because they cannot perceive stereo visual depth from 3D images .
To exclude such participants from the experiment, a screening test was performed to find stereo anomalies by measuring near-distance and far-distance stereovision. The “stereo fly test” was used to test the near-distance stereovision. The test equipment was composed of nine circles with different stereo visual resolutions similar to the stereo visual assessment tools recommended in Recommendation ITU-R BT.1438. The visual angles of each stereovision had a magnitude of 40 ~ 800 seconds, as shown in the following Figure. For the far-distance stereovision test, participants wearing 3D glasses were asked to watch 3D videos from 3DTV from a distance of three meters from the screen.
The screen displayed four grey rectangles or randomized dot rectangles, and the participants were asked to choose one rectangle with a different depth from the other rectangles, and to fill in the questionnaire with the answer. A total of 48 (2 stimulus × 6 disparity × 4 repetition) 3D images were presented in randomized order.
Far-distance stereovision test (a) and (b) are the images presented to the left and right eyes respectively. Image (c) is the image showing the level of disparity when (a) and (b) were presented to both eyes at the same time. In (c), the white rectangle implies the location of the grey picture for the right eye overlapping the picture for the left eye. In all the stereoscopic images, one randomized rectangle appears to have different depth than other rectangles. For example, the upper right rectangle from (c) has a different disparity level than those of the other rectangles, and hence it is perceived to have different depth.
Subjects were presented with 45 randomized video clips. They filled in 5-point scaled questionnaires on dizziness and/or visual discomfort while watching those video clips. Comfort/discomfort assessment while 3D content watching was the viewing safety evaluation. Questions were used to evaluate test images designed with five levels for each question item in the ACR-HR (Absolute Category Rating-Hidden Reference) method of ITU-T P.910. The specific answers were on a 5‑point scale (high discomfort, mild discomfort, nil, mild comfort, and high comfort) for each test video clip.