Our objective of this research series is to find the image feature that represents Harajuku Kawaii style by means of computer vision. With the hypothesis that Harajuku Kawaii has a mixture of colorful items, we hypothesized that it is distinct not only from Japanese street snaps but also from Thailand street snaps. We applied a computer vision technique to measure colorfulness proposed by Hasler and Süsstrunk to test the hypothesis in this work.
We investigated the differences of colorfulness among three fashion styles, which were Thai street snaps, Japanese street snaps and Harajuku Kawaii fashion. The finding showed that the colorfulness of Harajuku Kawaii fashion has the highest average, followed by Thai street snaps and Japanese street snaps, and they were significantly different. It can also conclude that Harajuku kawaii fashion is distinct and unique to each person in colorfulness.
(Peeraya SRIPIAN, Kejkaew THANASUAN, Keiko MIYATAKE, Michiko OHKURA)