—With the development of the Internet, people purchase various products online, with items for purchase being proposed on the basis of recommending algorithms. Fashion items are one of them. However, difficulties in recommending fashion items arise from problems in detecting the item's characteristics such as color, detail, and silhouette, in many still images on online shopping sites. In particular, color is a simple characteristic, but different color combinations give different impressions. Some people obtain a lovely and cute impression from the color combination of white and pink. Thus, this type of information is needed to focus on recommendation of items to users via the Internet. Our main aim in this study is to extract the rules related to the information described above, i.e., to obtain the associated rules from color combinations to derive impressions. We accumulated fashion styling data from online shopping sites related to fashion items. Then, we attempted to acquire the associated rules. After this procedure, we extracted approximately 6,000 rules. Subsequently, we tested the validity of the rules using questionnaires. The results demonstrated that our extracted rules were highly significant when compared with a controlled condition. Thus, we have successfully acquired association rules for fashion item recommendation.
[1]
Ling-Ling Wang,et al.
Color texture segmentation for clothing in a computer-aided fashion design system
,
1996,
Image Vis. Comput..
[2]
Henry Lieberman,et al.
What am I gonna wear?: scenario-oriented recommendation
,
2007,
IUI '07.
[3]
Jun Rekimoto,et al.
suGATALOG: Fashion Coordination System That Supports Users to Choose Everyday Fashion with Clothed Pictures
,
2013,
HCI.
[4]
Shigenobu Kobayashi,et al.
The aim and method of the color image scale
,
2009
.
[5]
S. TAKAHASHI,et al.
MIRROR APPLIANCE: RECOMMENDATION OF CLOTHES COORDINATION IN DAILY LIFE
,
2008
.
[6]
Tomoharu Iwata,et al.
Fashion Coordinates Recommender System Using Photographs from Fashion Magazines
,
2011,
IJCAI.