8 May 2024
Animal Logic’s Research and Development Department is always looking for new ways to promote innovation and foster collaboration, and in 2023, an exciting opportunity came up that saw the team initiate a sponsored research project with the University of British Columbia located in Vancouver, Canada.
The project was led by Engineering Manager Curtis Andrus, with assistance from developers Guillaume Pernin and Soorya Narayan, with Manuel Oliveira, a PhD student, the main contributor to the project. Over four months, Manuel collaborated with the Procedural Geometry & Simulation team on the proposed problem and received mentorship from his Professor Alla Sheffer (University of British Columbia). While undergoing the set-up and application process, Animal Logic received incredible assistance from the Innovation Partnerships Team, Sabrina Hauser, and Tina Lee, which led to the team being accepted into the Mitacs Accelerate program. The Mitacs Accelerate program provides short and long-term funding internships for master’s students, PhD candidates, and postdoctoral fellows.
The research project, Grooming – Guide Hair Interpolation, looked at Alfro, Animal Logic’s hair and fur grooming system, which exists largely as a set of nodes in Houdini. The typical artist workflow involves manually grooming a small set of guide hairs (thousands) and interpolating between those hairs to get the full-density groom. The interpolation step needs to build the shape of a groom curve from the shapes of nearby guide curves while avoiding guides on entirely different parts of the skin. Before the research project, this was achieved by generating a triangle mesh from the guide curves’ root points (called a retile mesh) and doing closest point lookups on that mesh to determine interpolation weights. However, this approach still had limitations with self-intersection surfaces or small numbers of guide curves. By the end of this project, a new approach was discovered that enabled teams to interpolate an even fewer number of guides directly on the original surface, avoiding the problems with the retile mesh approach!
The project was extremely beneficial to both parties, providing insightful research results and successful collaboration across disciplines.
Click here to read the technical paper, written by Soorya Narayan sourced from Manuel Oliveira’s research work.