Gaussian Splatting vs. Classical Photogrammetry
A comparison of Gaussian Splatting and classical photogrammetry for creating virtual backdrops.
2025
This paper describes and evaluates three end-to-end workflows from capture to reconstruction including photogrammetry, 3D Gaussian splatting (GS), and extraction of textured meshes from splats with the explicit aim of individuals unfamiliar with 3D reconstruction creating backdrops for virtual production with a central presenter by using a smartphone as capture device. We show that GS significantly outperforms the alternatives in our experiments.
Videos Used in 2AFC Task
| Scene | Gaussian Splatting | 2DGS | Photogrammetry |
|---|---|---|---|
| Chapel | |||
| Lecture Room | |||
| Lecture Hall | |||
| Meeting Room | |||
| Office Space | |||
| Outdoor Scene | |||
| Cluttered Space |
3D Models Used in 2AFC Task
Note: The 3D models will be available for download soon.
Raw Capture Data
The dataset is available for download from Hugging Face.