基于Stable Diffusion的虚拟人形象预设计的应用与研究
Application and Research on Virtual Human Image Pre-design Based on Stable Diffusion
曾建勇 1沈晓萍1
作者信息
- 1. 嘉兴职业技术学院 互联网学院,浙江 嘉兴 314036
- 折叠
摘要
对当前AIGC在虚拟形象预设计方面的现状及影响进行了分析和探讨.以Stable Diffusion为例,详细介绍了工程构建和实现,对相关模块的作用、运行环境、使用方法及其指令等多个方面进行了综合叙述、分析和探讨,针对使用不同采样方法、不同采样参数及不同训练模型生成图片效果的优劣进行了说明及展示.随后,通过项目实例,完整地展示了人物形象预设计的过程.最后,对AIGC等新技术可能带来的社会影响进行了预测和总结.
Abstract
This paper analyzes and explores the current status and impact of AIGC(AI Generated Content)in virtual image pre-design.Taking Stable Diffusion as an example,this paper provides a detailed introduction to engineering construction and implementation,and comprehensively describes,analyzes,and explores the functions,operating environment,usage methods,and instructions of relevant modules.It also explains and demonstrates the advantages and disadvantages of the effect of using different sampling methods,sampling parameters,and training models to generate images.Subsequently,the process of the human image pre-design is fully demonstrated through project examples.Finally,the potential social impacts of new technologies such as AIGC are predicted and summarized.
关键词
AI生成内容/Stable/Diffusion/生成对抗网络/提示词/虚拟人形象Key words
AIGC/Stable Diffusion/GAN/prompt/virtual human image引用本文复制引用
基金项目
2022年浙江省大学生科技创新活动计划暨新苗人才计划(2022R470A009)
出版年
2024
国家科技期刊平台
NSTL
万方数据