关键词: 服装建模, 虚拟展示, 部件设计, 交互式遗传算法, 服装款式推荐

Abstract:

Because of the high threshold of professional knowledge in the virtual design and customization of clothing, customers are not able to obtain their favorite customized styles efficiently. Taking the Polo shirt as an example, the research reported in this paper designed a componentized 3-D model of the Polo shirt and the appearance of its striped knitted fabric, and established a database of Polo shirt design components. The research worked out an optimized interactive genetic algorithm, and an adaptive weighted allocation strategy was used based on similarity. Through algorithm coding, initial population generation, using genetic similarity to obtain individual fitness, selection, subchromosome preference fixation or elimination, crossover and mutation, a style recommendation system, which gradually generate user-satisfied Polo shirt style model design schemes, was established. It was verified that with the increase of evolutionary generation, the maximum fitness value and average fitness value of the population increase. While ensuring a certain search space, it is gradually inclined to the psychological preferences of users, and users can gradually search for more preferred design solutions to achieve user-oriented personalized quick style generation and recommendation of Polo shirts.

Key words: clothing modeling, virtual display, component design, interactive genetic algorithm, garment style recommendation