文献类型：会议论文

中文题名：Shanghai-Style Realistic Style Watercolor Painting Style Transfer by Using RSIM Evaluation

作者：Rongrong Fu[1];Yuanyuan Wang[1];Jiajun Lin[2];Ruiyang Fan[2];Han Zhao[1]

机构：[1]School of Art Design and Media,East China University of Science and Technology;[2]School of Information Science and Engineering,East China University of Science and Technology

会议论文集：2021中国智能自动化大会(CIAC 2021)论文集

会议日期：20211106

会议地点：线上会议

语种：英文

中文关键词：Style transfer;Deep learning;Shanghai-style watercolor painting;RSIM

摘要：As an important branch of Shanghai-style painting and the source of the development of Chinese watercolor painting, Shanghai-style watercolor painting has distinctive style characteristics, but it is undergoing a dilemma in the inheritance due to the limited application scenes. This paper exploits the VGG19 feature extraction model and Gatys algorithm to carry out image transfer, where reduced structural similarity index RSIM(Reduced Structural Similarity) is proposed to measure the transfer image by sub-index for selecting the optimal transfer image.The style samples of Shanghai-style realistic watercolor paintings and the content samples of Hudecs buildings are transferred to generate materials with distinctive features of Shanghai-style watercolor paintings. On this basis of generated materials, a series of cultural and creative product designs can be created.