A team from the Massachusetts Institute of Technology (MIT) has created a technique that allows animators to have a more significant scale of control over their works. The researchers have developed a method that produces mathematical functions known as “barycentric coordinates,” which indicate how 2D and 3D shapes can move, stretch, and contour in space. This tool will permit animators to determine functions that align with their vision for how their animated characters should look and behave.
Previously, barycentric coordinate functions have generally been rigid, with only a single approach available for each respective character. This lack of flexibility often forced animators to begin afresh with another style whenever they wanted to adjust the appearance of their animations, a time-consuming and frustrating process. The new research aims to address this issue and provide a more versatile approach to animation. While its primary focus is on animation, the technique could have potential applications in other fields like medical imaging, architecture, virtual reality, and computer vision for robots.
For the study, the MIT team took the traditionally complex equations used in ‘cage-based’ animation, where a simpler shape outlines the character, and reworked these into a more generalized approach. This allowed for the creation of ‘smoothness energies’ for any shape, giving animators a selection of different smoothness energies to choose from until they identify the one that works best for their design.
In driving this change, the scholars used an innovative type of neural network to model the complicated barycentric coordinate functions that essentially control how each corner of a shape manipulates the shape’s interior. The neural network helps create a function satisfying all necessary limits while maintaining smoothness.
This new approach allows an animator to test a function, review the final outcome, and then adjust the coordinates to generate various animations until they produce the specific one they desire. By enabling more natural-looking animations, the team demonstrated an increase in flexibility not previously available.
Looking ahead, the researchers hope to speed up the neural network and incorporate this method into an interactive system that would facilitate an artist to easily modify animations in real time. They believe that the significant impact of their research lies in the flexibility that the neural networks provide, which were not previously accessible. The research was funded by several organizations, including the U.S. Army Research Office, the U.S. Air Force Office of Scientific Research, the U.S. National Science Foundation, and the Singapore Defense Science and Technology Agency. The findings of the research were presented at SIGGRAPH Asia.
