基于生成对抗网络的神兽形象设计  

Image Design of Mythical Creatures Based on Generative Antagonistic Network

作　　者：胡玥茗 

机构地区：[1]中国美术学院

出　　处：《上海工艺美术》2024年第4期80-82,共3页Shanghai Arts & Crafts

摘　　要：随着算法时代的到来,“算法依赖”影响着人类的文化认知,数字化的生活方式削弱了人们的生态关注。本文以生成对抗网络为媒介,结合未来的社会环境及自然环境发展趋势,对神兽形象进行创新设计。旨在探究生成对抗网络在设计领域的新应用,探索中华传统神兽文化的新发展,构建对于算法时代的技术与生态的设计思辨。本文阐明了神兽形象的基本概念,分析总结了神兽形象的三种造型规律。探索了生成对抗网络在设计领域的应用,归纳得出中间态和最终态两类设计辅助方法。With the advent of Algorithm Era, “algorithm dependence” has impacts on human cultural cognition,while digital lifestyle weakens people's ecological concern. This article regards Generative Antagonistic Network as the medium, echoes with the future social environment and natural environment development trend, and presents innovative design for the image of mythical creatures. This article is aimed to explore the new applications of generative adversarial network in the design sphere, find the way for new development of traditional Chinese mythical creature culture, and make intellectual enquiries for the design of technology and ecosystem in Algorithm Era.

关 键 词：生成对抗网络 形象设计 创新设计 神兽 中间态 文化认知 造型规律 数字化 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]