2015 年 53 巻 5 号 p. 255-263
A support system for clothes selection by oneself is necessary as a psychological support to help the visually impaired to live independently and be able to socialize. We previously developed an environmental light utilization system in which a smartphone used the light of an indoor environment for illumination. However, for correct color recognition, it was necessary to photograph the clothes with color-rendering indexes (references:color standards) by a smartphone camera, and then correct the colors using appropriate numerical formula for adjusting colors suitable for the illumination conditions. In this study, we developed a novel technique that allows correct color recognition using only a smartphone without the references. In the new system, we used a flash light as the dominant source of illumination, and we developed a numerical formula for adjusting colors which allows correct color recognition even when the indoor light environment changes. We confirmed the precision of color recognition by experiments. Good color difference between 5 and 7 was obtained in the CIE1976 L*a*b*color space. In addition, colors and patterns on clothes were output almost precisely by sounds in experiments on subjects. The results were evaluated on a four-grade scale (between 0 and 3), and average scores of color recognition and pattern recognition were 2.34 and 2.58, respectively. The present results demonstrate that this system is very effective to support visually impaired persons in clothes selection. The use of this system is expected to be widespread in the future because it does not require references.