Adaptive Texture Alignment for Japanese Kimono Design

A yukata is a type of traditional Japanese kimono. An alignment of its texture pattern is an important factor of the yukata design. There are rules in the texture alignment of the yukata. The rules are comparratively simple. However, the texture alignment is difficult for the designer because the texture alignment should be performed with consideration to the rules and the wearer's taste, Additioanlly, it is necessary for the designer to create the cutting pattern from a limited length of the kimono cloth. Consequently, a design support system for the yukata is required. We have developed the image processing algorithm to simulate the condition of the texture alignment. It becomes possible to perform the texture alignment based on the traditioonal rule automatically. However, some cutting pattern becomes over length of a standard kimono cloth. In this paper, we describe a multi-agent system for supporting the texture alignment of the yukata. We developed texture alignment agents, and a management agent. The management agent acts management of the wearer's body sizes, the condition of the texture alignment and the cutting pattern and orders to the texture alignment agents to carry out the texture alignment according to the wearer's taste. By repeating trial and error, the realistic texture alignment became possible.