Incremental Gaussian splatting: Gradual 3D reconstruction from a monocular camera following physical world changes

Author

Minamida, Keigo and Rekimoto, Jun

Abstract

Real-time 3D reconstruction is crucial for remote collaboration, but adapting to dynamic environments remains challenging. We introduce Incremental Gaussian Splatting (I-GS), a novel approach for efficient 3D reconstruction in changing scenes. I-GS focuses on human-induced changes, using images from an operator-worn monocular camera. By processing only regions consistent with the current state, I-GS achieves sequential updates. Our evaluation demonstrates that I-GS outperforms conventional Gaussian Splatting, providing accurate spatial representations resilient to moving objects. This advance significantly improves the quality of remote collaboration in dynamic environments.