KANSEI evaluation of color images presented in color gamuts with different blue primaries

Abstract

-The effect of blue primary color on image quality was studied. We used the KANSEI evaluation method, and for experiments, we used Semantic Differential method. First, four different blue primaries of 430, 450, 470, and 480 nm with full brightness of L max equal 60 cd/m 2 were tested.The same experiment was performed for 450, 470, and 480 nm at L max of 170 cd/m 2 . The results for the four primaries show that 450-and 470-nm blue primaries are the best. Then 430 nm was next, and 480-nm blue primary shows the worst performance. In the tests of three primaries, the 470 nm shows the highest scores for bluish test images. Results are compared with color naming data in the entire chromaticity diagram.

53-1:Invited Paper: The Winter Green and Summer Blue

“…Then allowable variability range of primaries along the spectral locus is also important. However, no study has been reported to examine the effect of gamut change using different color primaries along the spectral locus but ours [6]. In the previous study, we examined the effect of different monochromatic blue primaries, 430nm, 450nm, 470nm, and 480nm, at low luminance level (around 60 cd/m 2 ), and 450nm, 450nm, and 570nm at high luminance level (around 170 cd/m 2 ).…”

“…Test images, apparatus, evaluation words, and procedure are common to Exp.1 and Exp.2. Details of Exp.1 are described in Ayama et al [6].…”

P-33: Color Classification of Images using Categorical Color Database

“…Example of "blue" image and the percentage of pixel distribution are shown in Figure 2. This image was used as a "blue" test image in the KANSEI evaluation study [9]. Ten images that were not classified into any of the 15 groups were added to test images.…”

“…First, we used the criteria indicated in Table 1 in the color classification. This was done originally for selecting proper color images that represent typical colors in the KANSEI evaluation experiment [9]. Fifteen color categories, red, green, blue,yellow, orange, red-green, green-blue, yellow-blue, yellowgreen,red-blue, pink, purple, multi-colors, monochromatic, and skin were employed.…”

“…Color distributions of more than 2600 images were analyzed and they were classified into one of the groups in Table 1. The most appropriate image was selected from each group and employed as test image in the KANSEI evaluation study [9]. Classification of 2600 images did not seem improper so far, but we thought some kind of confirmation should be done.…”