CrossGaussian: Enhancing Remote Collaboration through 3D Gaussian Splatting and Real-time 360° Streaming

Jaehyun Byun, Byunghoon Kang, Youghyun Gwon, Hongsong Choi, Yunseo Do, Eunho Kim, Sangkeun Park, Seungjae Oh
September 2025
PDF Cite

Abstract

Remote users often face significant challenges in remote collaboration systems when joining virtual scenes reconstructed from a local user’s environment. They are disadvantaged by the information asymmetry inherent in a shared virtual environment compared to local users. We present CrossGaussian, a VR collaboration system designed to address these limitations by providing remote users with a comprehensive 3D interactive view of the shared environment, created using 3DGS. It blends 360° video streaming and a large-scale 3DGS reconstruction via our automated pipeline. We then define the design space for visualization and interaction techniques that combine wide-coverage 3D and responsive 360° scenes.

Type
Conference-Extended-Abstract
Publication
In Proceedings of the 38th Annual ACM Symposium on User Interface Software and Technology
Remote Collaboration Virtual Reality 3D Gaussian Splatting ACM UIST
Jaehyun Byun
Jaehyun Byun
Undergraduate Researcher
Yunseo Do
Yunseo Do
Undergraduate Researcher
Seungjae Oh
Seungjae Oh
Assistant Professor

My research interests include interactive I/O technologies, Haptics, and game interaction techniques.