Abstract
Purpose :
To estimate the luminance contrast thresholds using a stimulus based on the use of chromatic and spatial noise to mask other cues than luminance differences.
Methods :
The stimuli were developed in 2 systems: ViSaGe (CRS, UK), MATLAB R2010a, and CRS Toolbox for MATLAB; MacBook PRO (Apple, USA) and MATLAB R2010a. Stimuli were mosaic of circles with 10 sizes (spatial noise), colored by 1 out of 16 chromaticities (color noise) around a chromaticity in the CIE 1976 color space. The 5° stimulus had a target (Landolt C, 4.4° outer and 2.2° inner diameters, and 1° gap) differing from background by luminance. Initially, the target had 4 cd/m2 and background had 40 cd/m2. Subjects were tested to compare luminance contrast threshold estimates using ViSaGe (n = 10) or MacBook (n = 7). Contrast thresholds also were estimated using 3 chromaticity noises centered at the CIE1976 C2: u’=0.219, v’= 0.48. Ten subjects were tested using the ViSaGe system to compare the contrast threshold estimated using colors from different regions of the CIE1976 color space; C1: u’= 0.215, v’= 0.531; C2; and C3: u’= 0.225, v’= 0.415. Each noise was modulated along four chromatic vector sizes (0.06, 0.03, 0.015, 0.0075). The subject identified the Landolt C gap orientation (up, down, left, or right). A forced-choice staircases controlled the contrast between target and background. The staircase rule was 2 hits:1 error, and 21 reversals ended the test. Thresholds were estimated using the last 15 reversals. ANOVA was used to compare results estimated at each condition.
Results :
For all conditions, the larger the chromatic vector, the larger the contrast threshold (p < 0.05). There was no significant difference between contrast threshold as a function of the chromatic noise using ViSaGe and MacBook Pro systems (p > 0.05), as well as the same functions estimated from 3 regions of the CIE1976 color space. Since no difference between different systems and locations in the color space were observed, the luminance contrast thresholds were averaged for each vector size: 0.06 = 35.5 ± 7.7%; 0.03 = 30.8 ± 6.4%; 0.015 = 27.1 ± 7.8%; and 0.0075 = 21.7 ± 6.7%.
Conclusions :
Luminance contrast threshold was dependent on the chromaticity noise, but independent of color space location where the noise was modulated, and also independent from the system used to test. We introduced a new method to investigate the luminance vision.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.