It is vitally important to estimate visibilities of all visual targets in real environments to maintain visual safety. As luminous color images in real lit environments are relatively easily obtained, it is reasonable to attempt to establish a method to estimate visibilities of objects using these images.
Authors have proposed C (luminance contrast) -A (adaptation luminance) graph which can estimate visibilities from luminance images using contrast profile method. In other previous studies, authors have proposed a method to analyze effects of color on visibility using XYZ stimulus values and L*a*b* color space. In addition, authors have conducted visibility evaluation experiments using circular objects, varying luminance and chromaticity for common, protan, and deutan color vision type people. Experiment results were analyzed using CIE 1931 xy chromaticity diagram and aC (red-green contrast) - bC (yellow-blue contrast) graph. Experiment results of common, protan, and deutan types were compared respectively and characteristics of each type were acquired. These characteristics are caused by differences in functions of LMS cone cells on the retina among color vision types. In this study, these experiment results were expressed in contrasts of LMS stimulus values which is based on responses of LMS cone cells and the visibility estimation method was considered expecting implementations in color universal design.
First, XYZ stimulus values of the background and circular targets were converted to LMS stimulus values and LMS stimulus value contrasts were calculated. When the distribution of “invisible” stimuli for common type was expressed on the plane where LCC (L contrast) was represented in horizontal axis, MCC (M contrast) in vertical axis, and SCC (S contrast) was from -0.10 to 0.10, the slope was almost 1. However, the distribution of “invisible” stimuli for the protan type was elongated along LCC axis, and the distribution of “invisible” stimuli for the deutan type was elongated along MCC axis. Using these results, four vertices on the LCC-MCC plane (SCC≒0) were set for each distributions and the visibility threshold estimation method was established. It expresses of “invisible” stimuli as rhombus-like quadrangles. Sizes and angles of quadrangle diagonals differ depending on color vision types. As described above, visibility estimation method expecting to achieve color universal design was established.
Furthermore, visibilities of red visual objects were estimated from the luminous color images of real lit environments using proposed the visibility estimation method. As it was suggested that it is the possible to estimate visibilities from ratios of visibility threshold luminance contrasts in the previous study of authors, it may also be possible to assume that it is possible to estimate visibilities from ratios of visibility thresholds of contrasts of LMS stimulus values. Based on contrast profile method, and by adjusting detection size and point in order to achieve the maximum aC value, contrast values were obtained for each images. The visibility was estimated from LCC, MCC, and SCC values using the proposed method. As a result, it was seen that the ratios of protan and deutan types were lower than the ratio of common types. For red visual objects, it was estimated that visibility of object is low and unsafe for protan or deutan type people.
In this study, experiment results of previous studies were shown as contrast of LMS stimulus values, and visibility estimation method expecting to achieve color universal design was established. Furthermore, visibility estimations using the proposed method and luminous color image were conducted. It was confirmed that the method is useful for achieving color universal design.