SynthRender and IRIS: Open-Source Framework and Dataset for Bidirectional Sim-Real Transfer in Industrial Object Perception
Paper • 2602.21141 • Published
Datasets:
Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
Job has been terminated due to a temporary spike in resource usage and may be restarted later.
Error code: JobManagerCrashedError
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
text string |
|---|
0 0.642578 0.427734 0.050781 0.082031 |
2 0.335449 0.559570 0.090820 0.068359 |
8 0.554688 0.300293 0.134766 0.104492 |
9 0.351562 0.654785 0.169922 0.071289 |
5 0.333496 0.291016 0.206055 0.160156 |
4 0.769043 0.515137 0.102539 0.092773 |
3 0.573242 0.617676 0.109375 0.133789 |
0 0.891113 0.216309 0.127930 0.075195 |
0 0.496094 0.725098 0.093750 0.151367 |
1 0.595703 0.500488 0.099609 0.194336 |
1 0.816406 0.655762 0.222656 0.084961 |
2 0.443359 0.520996 0.054688 0.090820 |
2 0.590332 0.754883 0.096680 0.056641 |
6 0.722168 0.481445 0.149414 0.142578 |
6 0.911133 0.345215 0.169922 0.129883 |
9 0.480957 0.347168 0.100586 0.127930 |
5 0.895508 0.734375 0.208984 0.101562 |
5 0.388184 0.420898 0.090820 0.232422 |
4 0.863281 0.523926 0.117188 0.114258 |
4 0.807617 0.265625 0.115234 0.113281 |
3 0.651367 0.145508 0.119141 0.119141 |
3 0.686035 0.260742 0.081055 0.117188 |
0 0.653809 0.628418 0.110352 0.084961 |
1 0.395508 0.538574 0.148438 0.198242 |
2 0.440918 0.426270 0.067383 0.086914 |
8 0.933105 0.656738 0.133789 0.079102 |
6 0.529297 0.627930 0.160156 0.144531 |
9 0.386230 0.743652 0.145508 0.122070 |
5 0.887695 0.163574 0.144531 0.223633 |
4 0.721191 0.356445 0.075195 0.109375 |
3 0.571777 0.822754 0.135742 0.147461 |
0 0.568359 0.339844 0.099609 0.113281 |
0 0.745117 0.617676 0.105469 0.110352 |
1 0.607422 0.495605 0.138672 0.190430 |
1 0.343262 0.687012 0.211914 0.125977 |
2 0.367676 0.366699 0.090820 0.053711 |
2 0.263672 0.613281 0.091797 0.054688 |
8 0.305664 0.490723 0.144531 0.086914 |
7 0.498047 0.298340 0.167969 0.188477 |
6 0.408203 0.022949 0.148438 0.045898 |
6 0.676758 0.741699 0.189453 0.149414 |
9 0.466797 0.784180 0.109375 0.166016 |
9 0.303711 0.290039 0.134766 0.091797 |
5 0.134277 0.456543 0.088867 0.243164 |
5 0.819336 0.597656 0.097656 0.246094 |
4 0.264648 0.833496 0.126953 0.124023 |
4 0.224609 0.428711 0.099609 0.107422 |
3 0.665039 0.422852 0.111328 0.128906 |
0 0.631836 0.436523 0.093750 0.066406 |
1 0.830566 0.434082 0.069336 0.139648 |
2 0.425781 0.569824 0.054688 0.059570 |
8 0.452148 0.444336 0.099609 0.056641 |
9 0.248535 0.683594 0.071289 0.115234 |
5 0.729980 0.693359 0.178711 0.064453 |
4 0.458984 0.510254 0.076172 0.069336 |
3 0.607910 0.491211 0.096680 0.093750 |
0 0.728516 0.484863 0.097656 0.061523 |
0 0.689453 0.532715 0.066406 0.096680 |
1 0.541504 0.299316 0.086914 0.116211 |
1 0.532227 0.578613 0.119141 0.100586 |
2 0.639160 0.388184 0.061523 0.043945 |
2 0.485352 0.742676 0.048828 0.053711 |
8 0.569824 0.339355 0.047852 0.096680 |
8 0.657227 0.778809 0.064453 0.112305 |
7 0.788574 0.378418 0.104492 0.131836 |
6 0.749512 0.598633 0.102539 0.107422 |
9 0.647461 0.583496 0.097656 0.081055 |
9 0.484375 0.394043 0.109375 0.051758 |
5 0.914551 0.561035 0.159180 0.100586 |
4 0.549805 0.726562 0.041016 0.076172 |
4 0.560059 0.409180 0.071289 0.054688 |
3 0.923828 0.199707 0.091797 0.081055 |
3 0.898438 0.317871 0.093750 0.083008 |
0 0.506348 0.415039 0.059570 0.074219 |
1 0.669434 0.466797 0.067383 0.128906 |
2 0.821289 0.497070 0.052734 0.044922 |
8 0.787109 0.448242 0.062500 0.091797 |
6 0.680664 0.375488 0.091797 0.086914 |
9 0.621094 0.592773 0.083984 0.083984 |
4 0.716797 0.594727 0.076172 0.074219 |
3 0.384766 0.775879 0.080078 0.090820 |
0 0.626465 0.550293 0.053711 0.088867 |
0 0.635254 0.291504 0.049805 0.083008 |
1 0.634766 0.685547 0.132812 0.056641 |
1 0.432617 0.490723 0.072266 0.127930 |
2 0.606934 0.490723 0.057617 0.038086 |
2 0.689453 0.545410 0.058594 0.032227 |
8 0.563965 0.238770 0.051758 0.090820 |
8 0.507324 0.527344 0.065430 0.085938 |
7 0.413086 0.384277 0.121094 0.094727 |
6 0.973633 0.604980 0.052734 0.092773 |
9 0.599121 0.413086 0.045898 0.099609 |
9 0.392090 0.625000 0.065430 0.101562 |
4 0.656738 0.217285 0.067383 0.067383 |
4 0.476562 0.461914 0.064453 0.042969 |
3 0.544434 0.450195 0.079102 0.064453 |
3 0.535645 0.337402 0.081055 0.084961 |
0 0.630371 0.573730 0.057617 0.073242 |
1 0.778809 0.682617 0.083008 0.123047 |
2 0.538086 0.678711 0.056641 0.033203 |
End of preview.
SynthRender_Robotics: Synthetic Training Sets for the Robotics Sim-to-Real Benchmark
This repository hosts the synthetic training sets generated with SynthRender and used to benchmark sim-to-real transfer on the public Robotics dataset (Horváth et al., 2022), as reported in "SynthRender and IRIS: Open-Source Framework and Dataset for Bidirectional Sim-Real Transfer in Industrial Object Perception" (arXiv:2602.21141).
Sample Images (720x720)
 |
 |
 |
 |
Dataset Summary
SynthRender is an open-source, scriptable Domain Randomization (DR) engine built on BlenderProc, used in the paper to systematically ablate rendering design choices (physics-based placement, exponential light sampling, RGB lighting, material randomization) and quantify their effect on sim-to-real transfer. The synthetic sets in this repository correspond to the configurations used to train detectors evaluated against the Robotics benchmark (10 classes, 190 real test images, 920 annotated instances), on which the proposed framework reached 99.1% mAP@50.
The repository contains two synthetic datasets, generated at different rendering resolutions:
| Resolution | Total Images | Train (80%) | Validation (20%) |
|---|---|---|---|
| 1024x1024 | 4,000 | 3,200 | 800 |
| 720x720 | 4,000 | 3,200 | 800 |
- Annotation format: YOLO-style bounding box labels (
class x_center y_center width height, normalized) - Modality: RGB images with text-based annotation files
- Classes: 10, matching the target Robotics benchmark
Relation to the Paper
The paper's contribution statement notes that multiple synthetic training sets of 4,000 domain-randomized images each were generated to support the ablation studies. This repository provides two such sets, at 1024x1024 and 720x720 resolution, used for the Robotics benchmark specifically. It is a companion resource to the IRIS dataset and to the SynthRender framework itself.
Intended Use
For research on synthetic-to-real domain transfer, domain randomization ablations, and object detection benchmarking against the public Robotics dataset. Not intended for commercial use.
License
Released under CC BY-NC 3.0. Commercial use is not permitted; attribution is required.
Citation
@article{araya2026synthrender,
title={SynthRender and IRIS: Open-Source Framework and Dataset for Bidirectional Sim-Real Transfer in Industrial Object Perception},
author={Araya-Martinez, Jose Moises and Tom, Thushar and Sanchis Reig, Adri{\'a}n and Rey Valiente, Pablo and Lambrecht, Jens and Kr{\"u}ger, J{\"o}rg},
journal={arXiv preprint arXiv:2602.21141},
year={2026}
}
J. M. Araya-Martinez, T. Tom, A. S. Reig, P. R. Valiente, J. Lambrecht, and J. Krüger, “Synthrender and iris: Open-source framework and dataset for bidirectional sim-real transfer in industrial object perception,” 2026. [Online]. Available: https://arxiv.org/abs/2602.21141
Contact
For questions about the dataset, please open a discussion on this repository.
- Downloads last month
- 188