文献类型：会议论文

英文题名：Generation and Design of Helan Mountain Petroglyphs Based on Style Transfer Algorithm

作者：Li, Yating[1];Chen, Yumiao[1]

机构：[1]East China Univ Sci & Technol, Sch Art Design & Media, 130 Meilong Rd, Shanghai, Peoples R China

会议论文集：11th International Conference on Culture and Computing (C&C) part of the 25th International Conference on Human-Computer Interaction (HCI)

会议日期：JUL 23-28, 2023

会议地点：Copenhagen, DENMARK

语种：英文

外文关键词：Helan Mountain Petroglyphs; Art Style Transfer; Deep Neural Networks

摘要：Primitive people used stoneware as tools to paint and record people's life contents with Petroglyphs, which reflected people's cognition of nature and life, and petroglyphs have become classic works in civilization. In recent years, as style transfer algorithm has been widely used in art, style transfer technology has been taken seriously in the field of art creation. With high ornamental value, Helan Mountain petroglyphs are perfect style samples. Aiming at the limited universality of the current simulation methods of Helan Mountain primitive petroglyphs, this paper proposes a method of art style transfer and simulation based on deep neural network. Using the VGG19 feature extraction model and taking Helan Mountain primitive petroglyphs as the prototype of artistic style, this paper generates petroglyphs of the corresponding style from pictures of modern family life, and tries to apply the style transfer algorithm technology to the presentation of Helan Mountain petroglyphs techniques, which is helpful to realize the inheritance and regeneration of primitive petroglyphs. At the same time, this paper also explores the style transfer algorithm and the new design mode of cultural and creative products, which can provide new ideas for the development of related cultural and creative products. As experiments show, the proposed method can convert ordinary photos into petroglyphs with a good effect, complete the art style transfer and simulation of petroglyphs. In addition, by combining with peripheral product design, the style transfer algorithm technology provides new innovative ideas for cultural and creative product design.