FURI | Spring 2023

How to Apply Volumetric Capture Technology to Bring Authentic Human Performances into Virtual Worlds by Using Multiple Cameras to Capture the Footage of a Person from Different Viewpoints in More Approachable Ways

This research proposes a methodology to apply volumetric capture technology for bringing authentic human performances into virtual worlds. The study builds on the advantages of existing models, SIMPLify and VIBE, and aims to simplify and optimize them using model pruning techniques. The output models will contain texture and can be opened in game engines. The process will be conducted using Google Colab, and an app will be created to visualize and edit the 3D reconstruction. The anticipated outcome is increased efficiency, reduced runtime, and the ability to bring authentic human performances into virtual worlds.