人工智能生成内容(AIGC)驱动的电影虚拟角色面部特效研究
Research on AIGC-driven facial animation for film virtual characters
吴方强 1徐沁雪 1周冰2
作者信息
- 1. 北京电影学院摄影系,北京 100088
- 2. 北京电影学院表演学院,北京 100088
- 折叠
摘要
为研究AIGC技术对特效领域中虚拟非人角色表情制作带来的变革,本文结合当下主流国产AI大模型,对影视级虚拟角色表情特效制作开展实验,并提出制作思路和优化的制作流程.实验结果基本满足影视作品虚拟预演(PreViz)快速制作的技术要求,可用于电影特效镜头的辅助制作.本文最后总结了该实验方法的可拓展性及迭代方向.
Abstract
Aiming at exploring the revolution that artificial intelligence generated content technology,also known as AIGC technology,brought to the production of facial animation for virtual non-human characters in the field of VFX,integrating with current leading domestic AI large language models,this paper conducts experiments towards the professional produc-tion of facial animation for virtual characters,then proposes an optimized production workflow.The results could meet the technical requirements for rapid previsualisation of virtual characters in film productions and could assisting production for special effects shots.For conclusion,this article summerises the expansibility and elevating directions of the proposed methodology.
关键词
AIGC/动作捕捉/文生视频/表情特效/国产大模型Key words
AIGC/Motion Capture/Text-to-Video/Facial Animation/Domestic Large Language Models引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据