Event-based Mosaicing Bundle Adjustment

Official repository for Event-based Mosaicing Bundle Adjustment, ECCV 2024 paper, by Shuang Guo and Guillermo Gallego.

PDF | Poster | Video | Dataset

Citation

If you use this work in your research, please cite it as follows:

@InProceedings{Guo24eccv,
  author        = {Shuang Guo and Guillermo Gallego},
  title         = {Event-based Mosaicing Bundle Adjustment},
  booktitle     = {European Conference on Computer Vision (ECCV)},
  pages         = {479--496},
  doi           = {10.1007/978-3-031-72624-8_27},
  year          = 2024
}

---

Setup

High-level Input-Output

Input:

• Events.
• Camera calibration.
• Initial trajectory from other front-end methods.
• (Optional) Initial panoramic gradient maps.

Output:

• Refined rotations of the event camera.
• Refined intensity panorama.

Usage

• Installation
• Execution
• Parameter Guide

Event Camera Rotation Dataset (ECRot)

We test EMBA on the newly-released ECRot dataset, see and download it at here. The instructions of running EMBA on the ECRot datasets can be found here.

Acknowledgements

This code leverages the following repository for computing the derivative of Lie Group B-splines:

• Basalt Headers

Additional Resources

• CMax-SLAM
• Research page (TU Berlin RIP lab)
• Course at TU Berlin
• Survey paper
• List of Resources