license: cc-by-4.0
Monado SLAM Datasets
The Monado SLAM datasets (MSD), are challenging egocentric visual-inertial (VI) datasets recorded from VR headsets to improve state-of-the-art VI-SLAM systems. They were originally developed to improve the open-source inside-out tracking component of the Monado project but its applicability is not limited to the XR field. The dataset has a permissive license CC-BY 4.0, meaning you can use them for any purpose you want, and only a mention of the original project is required. The creation of the dataset was supported by Collabora. Further work for publication was done in the Computer Vision Group from the Technical University of Munich.
Monado is an open-source OpenXR runtime that you can use to make devices OpenXR compatible. It also provides drivers for different existing hardware thanks to contributors in the community creating drivers for it. Monado provides useful modules that these drivers can use, including an inside-out head tracking component. While it allows to use different tracking systems, the preferred system is a fork of Basalt. Creating a good tracking solution requires a solid measurement pipeline to understand how changes in the system affect tracking quality. This dataset is published to support this purpose, and in the process, highlights problems with state-of-the-art real-time VI tracking systems that extend beyond the XR use case, especially in humanoid robotics.
For questions or comments, the preferred way is to use the Hugging Face
community discussions
tab. You might also join Monado's Discord server
and ask in the #slam channel.
For additional documentation see:
- Valve Index documentation
- HP Reverb G2 documentation
- Samsung Odyssey+ documentation
- Post-processing walkthrough video
Sequences
Categories
- MI_valve_index: Valve Index datasets (2x960x960 @54fps, IMU@1000Hz)
- MIC_calibration: Recordings of target with 3cm sides for camera and IMU calibration.
- MIO_others: Miscelaneous datasets including specific challenges.
- MIP_playing: Datasets in which the user is playing a game.
- MIPB_beat_saber: Music game in which the player hits song notes with two swords.
- MIPP_pistol_whip: Music game in which the player uses a pistol to target enemies in sync with the song rythm.
- MIPT_thrill_of_the_fight: Boxing simulator.
- MG_reverb_g2: (4x640x480 @30fps, IMU@1000Hz)
- MGC_calibration: Same as
MIC, includes pairwise recordings for each camera pair and magnetometer calibrations. - MGO_others: Miscelaneous recordings with different challenges.
- MGC_calibration: Same as
- MO_odyssey_plus: (2x640x480 @30fps, IMU@1000Hz)
- MOC_calibration: Same as previous calibration datasets. Includes a long MOC13 dataset for measuring IMU allan variance.
- MOO_others: Miscelaneous recordings with different challenges.
Table of sequences
Table of calibration sequences
Click to show calibration sequences
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
