open-source data and tools
FOR GENERATIVE ANIMATION
About us
Mbody is an open-source, bilingual, applied research project aiming to accelerate the advancement and adoption of generative character animation technologies. It’s led by a diverse team of researchers from four post-secondary applied research centers, and funded by the National Sciences and Engineering Research Council of Canada (NSERC). The goal is to provide value to both industry and academia by respecting their needs.
What are we creating ?
For animators
A free, open-source generative animation tool will be developed, improving the efficiency of creating humanoid animations
For animation tool developpers
Free, open-source software systems will be created to streamline the integration of generative character performance models into industry-standard content creation tools
For researchers
A free, commercially usable dataset of multi-modal, multi-agent interactions will be developed to support training generative ML models with diverse, high-quality data
What’s in it for partners
Throughout the project, the research team will engage with industry companies and academic institutions to help guide the research work and ensure that the outputs fulfill the above-stated goals. Although all materials produced through this project work will be made openly available, industry and academic partners who engage with us can expect some additional benefits; for example:
The ability to help guide the project work to ensure their own business needs are considered
Early access to project outputs such as software tools and datasets so that early validation and integration work can be conducted
The opportunity to connect with other industry innovators to stay current on technological developments in this area
Research Team
This applied research project is being led by four Canadian research centers, each offering their own unique expertise critical for the success of this cross-disciplinary work.
Research centers
SIRT
Sheridan College’s Screen Industries Research and Training Centre (SIRT), based in Ontario, is providing a wealth of experience in the areas of performance capture, character animation, and interactive software development
CDRIN
Le Centre de développement et de recherche en intelligence numérique (CDRIN), du Cégep de Matane, based in Quebec, is contributing their expertise in artificial intelligence (AI) technologies for the entertainment industry
Durham
Durham College’s Mixed Reality Capture Studio (MRC) and AI Hub, based in Ontario, are bringing additional expertise in the areas of performance capture, interactive software development, and artificial intelligence
LLio
Le Laboratoire en innovation ouverte (LLio) of the Cégep de Rivière-du-Loup, based in Quebec, is supporting the collaboration processes both internal and external in this project with their expertise in user-centered design and other open innovation practices
Work Conducted So Far
The preliminary work on this project has focused on defining a clear project plan which ensures that deliverables will contribute real value to our target end-users. Toward this goal, the following high-level activities have been completed:
1
Interviews were conducted with professionals working in areas related to character animation to identify common pain points and needs that aren’t currently being addressed.
2
A broad environmental scan and literature review was conducted in the area of generative character animation in order to identify the current state of research and commercial adoption in this area.
3
Candidate applied research areas related to generative character animation were identified, and insights from the industry interviews and background research were analyzed to identify a project focus which is both technically feasible and provides real value to industry stakeholders.
4
Detailed plans were formulated for the development of initial deliverables in the three major project areas: generative character animation tools; software tools to ease machine learning integration into content creation tools; and a multi-agent, human motion dataset.
Future plans
General inquiries
Contact point
mbody@sheridancollege.ca