Research on the Generation of Artistic Patterns in Fashion Design Based on Style Transfer

In the field of fashion design, consumers are increasingly demanding personalized and customized clothing, and the traditional design process relies on the experience and inspiration of designers, which is time-consuming and costly. Therefore, a study on the generation of artistic patterns in fashion design based on style transfer is put forward. Hand-held camera was used instead of traditional industrial camera or digital camera for image acquisition. On the basis of bilateral filtering, weight variable values were set by adaptive method, and Inceptionv3 model was used to classify clothing design art patterns. On the basis of the style transfer algorithm, the pre-trained VGG-19 model was used for feature extraction, and the bidirectional mapping performance of Cycle GAN was utilized to construct the art pattern generation model while retaining the input image features. The experimental results show that the loss function value of the design method is kept below 0.5, the search time is 0.28[Formula: see text]s, the running time is always lower than 0.2[Formula: see text]s, and the fitness value is higher than 80. Therefore, this method effectively improves the design efficiency, reduces the cost, and meets the needs of consumers for personalized design, providing strong support for automation and intelligence in the field of clothing design.