개인화 디자인을 위한 디자인 생성 도구 적용 연구 : 가구 디자인 사례를 중심으로

Personalization is recognized as an important factor that enhances market competitiveness. To develop personalized product successfully, the concept of co-design is suggested, which involves users in the design phase during product development. Whereas co-design requires UX experts to elicit user needs and foster users to visualize their thoughts, personalized products are usually designed by small group of designers who are not trained in user research methods. Therefore, we developed a method applying a generative design tool using GAN(Generative Adversarial Network) to help users produce design themselves from their own needs. The method is composed of three steps: GAN generation, need analysis, and concept design. First, we crawl images and train a network using StyleGAN2 algorithm to generate fake images. Second, we define PA(Perceived Attribute) from online review data and interview, and map related images by conducting card sorting analysis. Third, a generative tool is developed to explore latent space of GAN based on the personal result of card sorting. Users generate personal images using this tool, and designers evaluate concept design by interviewing users referring their personal images. For case study, we developed concept designs for dining chair, pendant light, and living room table. The effectiveness of the suggested method is verified by survey result.