NerfBaselines documentation

NerfBaselines is a framework for evaluating and comparing existing NeRF and 3DGS methods. Currently, most official implementations use different dataset loaders, evaluation protocols, and metrics, which renders benchmarking difficult. Therefore, this project aims to provide a unified interface for running and evaluating methods on different datasets in a consistent way using the same metrics. But instead of reimplementing the methods, we use the official implementations and wrap them so that they can be run easily using the same interface.

Please visit the project page to see the results of implemented methods on dataset benchmarks.

Web   |   GitHub   |   Paper

Main features

• Unified interface: All methods can be run using the same interface.

• Consistent evaluation: All methods are evaluated using the same metrics and protocols.

• Reproducibility: All methods are run using the official implementations.

• Easy to use: The CLI is easy to use and requires minimal setup.

• Extensible: New methods can be added easily by wrapping the official implementation.

• Public benchmarks: The results of all methods (and checkpoints) are available on the website.

For the full list of implemented methods, see the methods section. For the full list of available datasets (datasets which support automatic download), see the datasets section.

Tip

NerfBaselines now supports online demos for 3DGS-based methods. Check out the demos on the benchmark page!

Contents

The documentation is organized into several sections with increasing level of detail and difficulty:

• Installation
• Quickstart
• Using custom data
• Python tutorial

Advanced

• Adding new methods
• Custom dataset loader
• Adding new datasets
• Isolated backends
• Creating web demos
• Exporting meshes
• Custom web benchmark

Reference

• Methods
• Datasets
• CLI
• API

Tip

The documentation is available for all released versions of the project. You can switch between versions using the version selector in the bottom left corner.

Acknowledgements

A big thanks to the authors of all implemented methods for the great work they have done. We would also like to thank the authors of NerfStudio, especially Brent Yi, for viser - a great framework powering the viewer. This work was supported by the Czech Science Foundation (GAČR) EXPRO (grant no. 23-07973X), the Grant Agency of the Czech Technical University in Prague (grant no. SGS24/095/OHK3/2T/13), and by the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID:90254).

Citation

If you use this project in your research, please cite the following paper:

@article{kulhanek2024nerfbaselines,
  title={{N}erf{B}aselines: Consistent and Reproducible Evaluation of Novel View Synthesis Methods},
  author={Jonas Kulhanek and Torsten Sattler},
  year={2024},
  journal={arXiv},
}