Procedural Animation of Conversational Gestures for Virtual Humans

虚拟人对话手势的程序动画

• 批准号: CCB21-2021-00146
• 负责人: Mowat, Vicki
• 金额: $ 72.85万
• 依托单位: Sheridan Institute of Technology and Advanced Learning
• 依托单位国家: 加拿大
• 项目类别: Applied Research and Technology Partnership Grants
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=729263
• 关键词: Procedural; Animation; Conversational; Gestures; Virtual

Two Ontario colleges and two Quebec cégeps are collaborating on co-creation of a prototype, open-source systems solution for the procedural animation of conversation gestures in Virtual Humans (VHs). Led by Sheridan College's Screen Industries Research and Training Centre (SIRT), Durham College's Hub for Applied Research in Artificial Intelligence for Business Solutions (AI Hub) and Mixed Reality Capture Studio (MRC), Cégep de Matane's Centre de développement et de recherche en intelligence numérique (CDRIN) and the Cégep de Rivière-du-Loup's Le Living Lab en innovation ouverte (LLio), the research team will conduct dual-language, cross-sectoral research and development focused on procedural animation workflows that incorporate advanced, data-driven analysis and leverage Artificial Intelligence (AI) and Machine Learning (ML) to allow for automation and procedural generation of complex VH movements, including hand, limb, body movement. A diverse group of researchers and students from Canadian institutions will conduct research and development using a human-centered design approach in collaboration with leading companies, including SpinVFX, Ubisoft, Transitional Forms, ReImagine AI, AMD Canada, etc. Canadian companies currently employ, or are in the process of developing, their own Virtual Humans pipelines, primarily for film & television productions, and realtime productions for interactive games and entertainment. As consumers expect increasingly realistic, believable, and interactive VHs within entertainment, and VHs are used more frequently within systems for training, simulation, healthcare, etc., applied research and development is needed to target the challenges of procedural generation to improve the intelligence of VH actions while reducing the amount of time needed to convincingly animate VHs. Research will provide companies with a significant competitive advantage by refining existing VH solutions to better capture customers needs, expanding VH solutions into new market segments, and by creating disruptive innovation with new products and services. An open-source gesture database will offer an effective resource to raise the state-of-the art for animation in Canada. Companies will be able to improve existing systems, processes, and products by integrating the open-source system and further develop and/or develop proprietary applications of the projects outcomes.

两所安大略大学和两所魁北克大学正在合作共同创建一个原型，开源系统解决方案，用于虚拟人（VH）中对话手势的程序动画。由谢里丹学院的屏幕产业研究和培训中心（SIRT）、达勒姆学院的人工智能商业解决方案应用研究中心（AI Hub）和混合现实捕捉工作室（MRC）、马塔内的数字智能发展与研究中心（CDRIN）和卢普河学院的创新生活实验室（LLio）牵头，该研究团队将进行双语、跨部门的研究和开发，重点关注程序动画工作流程，包括先进的数据驱动分析，并利用人工智能（AI）和机器学习（ML），以实现复杂VH运动的自动化和程序化生成，包括手、肢体和身体运动。来自加拿大机构的多元化研究人员和学生将与领先公司合作，使用以人为本的设计方法进行研究和开发，包括SpinVFX，Ubisoft，Transitional Forms，ReImagine AI，AMD Canada等加拿大公司目前正在使用或正在开发自己的虚拟人管道，主要用于电影和电视制作，以及用于互动游戏和娱乐的实时制作。由于消费者期望娱乐中越来越逼真、可信和交互式的虚拟主机，并且虚拟主机在用于培训、模拟、医疗保健等的系统中使用得越来越频繁，需要应用研究和开发来瞄准程序生成的挑战，以提高VH动作的智能，同时减少令人信服地动画化VH所需的时间量。研究将通过改进现有的VH解决方案来更好地捕捉客户需求，将VH解决方案扩展到新的细分市场，并通过新产品和服务创造颠覆性创新，为公司提供显著的竞争优势。一个开源的手势数据库将提供一个有效的资源，以提高在加拿大的动画艺术的国家。公司将能够通过整合开源系统来改进现有的系统、流程和产品，并进一步开发和/或开发项目成果的专有应用程序。