微型电脑应用2025，Vol.41Issue(3)：117-121,5.
基于特征映射优化的无监督绘画图像风格迁移研究
Research on Unsupervised Painting Image Style Migration Based on Feature Mapping Optimization
摘要
Abstract
Existing migration methods usually rely on supervised learning and require a large amount of paired training data,which limits their application in real-world scenarios.In order to solve this problem,based on a derivative in generative adver-sarial networks,i.e.,deep convolutional generative adversarial networks,this research introduces the self-attention mechanism and domain-specific painting for feature mapping optimization,and proposes a novel painting image style migration algorithm.Experimental results show that the migration style reproduction accuracy of this algorithmic is up to 98.8％,which is 3.6％higher compared to 95.2％of the neural style migration model.Its peak signal-to-noise ratio value is up to 9.684,and its structural similarity value is up to 0.683.Inter-application tests show that the migrated images of this new model are signifi-cantly better than other models of the same type in terms of color,line and texture.The above results illustrate that the pro-posed algorithm has better generalization performance and practical application value,and can provide new ideas and methods for the development of the field of unsupervised painting image style migration.关键词
特征映射/风格迁移/生成对抗网络/自注意力机制/无监督学习Key words
feature mapping/style migration/generative adversarial network/self-attention mechanism/unsupervised learning分类
信息技术与安全科学引用本文复制引用
曹璐..基于特征映射优化的无监督绘画图像风格迁移研究[J].微型电脑应用,2025,41(3):117-121,5.基金项目
基于2021年院级教学改革研究项目(YJJG2021017) （YJJG2021017）
