Researchers from the National University of Singapore and Tencent BCG Business School have developed a new method called HOSNeRF (Human-Object-Scene Neural Radiance Fields) that can create 360° free-viewpoint renderings of dynamic scenes with human-environment interactions from just a single video.