Awesome
ES-PTAM: Event-based Stereo Parallel Tracking and Mapping
Official repository for ES-PTAM: Event-based Stereo Parallel Tracking and Mapping, by Suman Ghosh, Valentina Cavinato and Guillermo Gallego, published at the European Conference on Computer Vision (ECCV) Workshops 2024 Milan, Italy.
<h2 align="left"> </h2>:sparkles: It was presented as an Oral Spotlight at the NeVi Workshop.
:sparkles: A live demo was also presented at ECCV 2024.
If you use this work in your research, please cite it as follows:
@InProceedings{Ghosh24eccvw,
author = {Suman Ghosh and Valentina Cavinato and Guillermo Gallego},
title = {{ES-PTAM}: Event-based Stereo Parallel Tracking and Mapping},
booktitle = {European Conference on Computer Vision (ECCV) Workshops},
year = {2024}
}
Data Processing Pipeline
Input
- Events from two or more cameras
- Camera calibration (intrinsic, extrinsic) parameters
Output
- Camera (i.e., sensor rig) poses
- Depth map
- Confidence map
- Point cloud
- Intermediate ray density maps / Disparity Space Images (DSI)
Code
- Installation
- Running examples on different datasets
- Running live with DAVIS cameras
- Parameter tuning
Results
The original ES-PTAM trajectories and GT poses for various sequences are available here.
They have been evaluted using this tool.
License
The license is available here.
Related works
- Event-based Stereo Depth Estimation: A Survey
- MC-EMVS: Multi-Event-Camera Depth Estimation and Outlier Rejection by Refocused Events Fusion
- EVO: Event based Visual Odometry