Researchers From Max Plank Propose MIME: A Generative AI Model that Takes 3D Human Motion Capture and Generates Plausible 3D Scenes that are Consistent with the Motion

Researchers From Max Plank Propose MIME: A Generative AI Model that Takes 3D Human Motion Capture and Generates Plausible 3D Scenes that are Consistent with the Motion

By Aneesh Tickoo- June 20, 2023      Markettechpost

https://arxiv.org/abs/2212.04360

Humans are always interacting with their surroundings. They move about a space, touch things, sit on chairs, or sleep on beds. These interactions detail how the scene is set up and where the objects are. A mime is a performer who uses their comprehension of such relationships to create a rich, imaginative, 3D environment with nothing more than their body movements. Can they teach a computer to mimic human actions and make the appropriate 3D scene? Numerous fields, including architecture, gaming, virtual reality, and the synthesis of synthetic data, might benefit from this technique. For instance, there are substantial datasets of 3D human motion, such as AMASS, but these datasets seldom include details on the 3D setting in which they were collected.   .... '