Evaluating the Mechanism by Which the TNO Stereo Test Overestimates Stereo Thresholds

Abstract

Several studies have revealed that results of the TNO stereo test may overestimate the stereoacuity value (the less the better) compared with other testing measurements. The manner in which vision is divided among two eyes of a person wearing anaglyph glasses may play an important role. This study aimed to examine the effect of anaglyph glasses on stereopsis measurements. A stereopsis measurement system using a phoropter and two Sony smartphones was established. Four types of test patterns, including the original TNO stereo test pictures, isoluminant red-green pictures, grayscale pictures, and black and white dots pictures, were designed. A total of 32 participants were recruited for this study. A significant difference was found among the four groups (Friedman test, chi-square = 50.985, $P<0.001$ ). The Wilcoxon signed-rank test was used to detect differences between the groups. The stereoacuity of the original TNO group was significantly worse than those of the isoluminant, grayscale, and black-white groups. However, no significant difference was found between the isoluminant and grayscale groups. The correlation coefficient between the original TNO and isoluminant groups was 0.952 (Spearman’s rho, $P<0.001$ ; 95% confidence interval (CI), 0.901–0.988), while that between the original and grayscale groups was 0.771 (Spearman’s rho, $P<0.001$ ; 95% CI, 0.550–0.916). Anaglyph glasses played an important role in determining the stereoacuity values with the TNO stereo test, and the results were overestimated when compared with that of the other testing methods. The imbalance of chroma and luminance between the two eyes caused by the anaglyph glasses was indicated as one of the reasons for the overestimation of stereo thresholds.