Abstract
Pearls are graded by experts according to their shape, glossiness, and structural color. Although we have attempted to reveal how novices grade pearls (Kato et al., ECVP 2011), it is still unknown. The aim of this study is to investigate the reproducibility of experts' grading and to compare the experts' grading and novices' grading. Eight experts and six novices participated in the experiment. The task was to grade ten size-controlled pearls, which were labeled by other experts, who did not participate in the experiment. We used two sets of pearls for experts and two or three sets for novices. Experts graded five times for each set, and novices graded ten times for each set alternately. Novices were not given any instruction about how pearls should be graded. We found that the reproducibility of experts' grading was higher than that of novices' grading, but novices' reproducibility was enough higher than the chance. The correlation between novices' grading and labeled rank was as high as the correlation between experts' grading and labeled rank. But we found the difference between them in the results of multiple regression analysis in which the measured values of glossiness and structural color were used as independent variables and the grade was set as dependent variable. Although all novices had significant trials, which could be explained by the linear regression function, there were only three experts who had such trials.z88;This could be interpreted that novices without knowledge about how to grade pearls could make use of these two factors simply and that experts used them in a complex manner, because experts also ought to rely on these two factors to grade pearls. These results would suggest that people, at least Japanese people, inherently have a nature to be fascinated by the glossiness and structural color of pearl.
Meeting abstract presented at VSS 2013