Feasibility study on improving video production process using data-efficient Generative-AI models

使用数据高效的生成人工智能模型改进视频制作流程的可行性研究

• 批准号: 10081373
• 金额: $ 5.61万
• 依托单位: DEEPREEL LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2023
• 资助国家: 英国
• 起止时间: 2023 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10081373
• 关键词: Feasibility; study; improving; video; production

AI avatars are now increasingly used in creative industries, from areas as diverse as AI-generated videos for rapid video production, visually dubbing film & tv content, making novel music videos, or producing the live ABBA concert in London.While these avatars have become very high quality and photorealistic, the process of developing these avatars is cumbersome, time-consuming and expensive.This feasibility study is a collaboration between DeepReel and University of Bath to overcome the aforementioned problems in avatar creation to bring unprecedented productivity into DeepReel's business processes of creating AI-generated videos and democratise AI avatar technology as a result.If successful, the outputs of this study have the potential to transform video production in creative industries in the UK and around the world.

人工智能化身现在越来越多地用于创意产业，从人工智能生成的视频用于快速视频制作，视觉配音电影和电视内容，制作新颖的音乐视频或制作伦敦的现场ABBA音乐会等不同领域。虽然这些化身已经变得非常高的质量和照片逼真，但开发这些化身的过程很繁琐，这项可行性研究是DeepReel和巴斯大学之间的合作，旨在克服化身创建中的上述问题，为DeepReel创建AI的业务流程带来前所未有的生产力。如果成功的话，这项研究的成果有可能改变英国和世界各地创意产业的视频制作。