3D Garment Design Using Interactive Genetic Algorithm and k-Means Clustering

Today, with increasing development and utilization of digital technologies to promote and accelerate artistic production trends, the use of computers has found a special place in fashion design. In this study, a 3D garment design system using interactive genetic algorithms and k-means clustering is proposed. Components of a swimsuit (top style, waist and bottom style) were individually designed using the 3D garment design software. The fabric designs were reviewed and selected by users to make a database of favorites. The designs were generated using genetic algorithms on a dummy. By rotation of the dummy, the users could see all parts of the swimsuit intuitively. In order to reduce the eye strain/fatigue, the total population was limited into 8 clusters using k-means clustering. Then, the fitness of the designs was determined by the user and the next generations were produced by this fitness and evolution principles. The results of evolutions indicate the system efficiency in fashion set design using a fabric pattern set at minimum cost and shortest time according to user’s satisfaction.