---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: kpts
    struct:
    - name: depth
      list: float64
    - name: keypoints2d
      list:
        list: float64
    - name: keypoints3d
      list:
        list: float64
  splits:
  - name: train
    num_bytes: 8215932300
    num_examples: 100000
  download_size: 8267044940
  dataset_size: 8215932300
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
task_categories:
- object-detection
pretty_name: SynthesiaSet
size_categories:
- 100K<n<1M
---

<div align="center">
<h1>SynthesiaSet</h1>
Matan Davidi
&nbsp;&nbsp;&nbsp;&nbsp;
Cyril Moser
&nbsp;&nbsp;&nbsp;&nbsp;
Gent Serifi
&nbsp;&nbsp;&nbsp;&nbsp;
Nicola Studer
&nbsp;&nbsp;&nbsp;&nbsp;

ETH Zürich, Switzerland

[[Project on GitHub]](https://github.com/MixedRealityETHZ/AR_Musician)
</div>

Synthetic dataset containing `100K` Piano images with 2D+3D keypoint annotations (the 4 corners), rendered with [Mitsuba 3](https://www.mitsuba-renderer.org/).

Created as part of our [course project](https://github.com/MixedRealityETHZ/AR_Musician) for [Mixed Reality](https://cvg.ethz.ch/lectures/Mixed-Reality/) at ETH Zürich.

- **image**: RGB, 224x304
- **kpts**: dictionary containing:
  - keypoints3d: (4, 3) 3D positions for all 4 keypoints (in clockwise order, starting from the top-left)
  - keypoints2d: (4, 2) 2D keypoint projections
  - depth: (4,) Distance to the camera

### References
```
@software{jakob2022mitsuba3,
    title = {Mitsuba 3 renderer},
    author = {Wenzel Jakob and Sébastien Speierer and Nicolas Roussel and Merlin Nimier-David and Delio Vicini and Tizian Zeltner and Baptiste Nicolet and Miguel Crespo and Vincent Leroy and Ziyi Zhang},
    note = {https://mitsuba-renderer.org},
    version = {3.0.1},
    year = 2022,
}
```