Innovative Development Pathways for Preserving Chinese Liangshan Yi costume culture in the Digital Age

Abstract

Taking Liangshan Yi traditional costumes as the carrier, we extracted the cultural factors from them, and achieved the purpose of inheriting Yi culture through the innovative application in product design. The study extracts five major factors based on hierarchical analysis: form, color, pattern, material and emotional semantics. It also introduces the image style migration technology based on deep learning, using VGG network and ReLU activation function, by calculating the content loss and style loss, so that the white noise image gradually approximates the characteristics of the content image and the style image, thus realizing the stylized expression of the cultural factors. The morphology factor is the core of the design with an absolute high weight of 0.563, and the most important morphology factor is the “erwar cloak” with a weight of 0.578. In the deepest Conv5-3 layer, the average SSIM value of the structural similarity of neighboring feature maps is 0.8471, which proves that the model can stably extract the abstract style features. In the side-by-side comparison experiment of color migration, the model of this paper has PSNR=15.81, which is better than the 11.81 of comparative HSV and 12.56 of RGB, and its SSIM=0.6959, while the FID value is even lower than 168.05, which indicates that the color style of the generated image is the closest to the distribution of the real Yi dress. By analyzing the design and applying the style migration to the Yi dress culture, the study provides a model that can be used as a reference for the digital inheritance and innovation of the Liangshan Yi dress and even other ethnic cultures.