Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
Paper • 1703.06907 • Published
Error code: UnexpectedError
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.
image image | label class label |
|---|---|
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off | |
0off |
Synthetic dataset for training phone screen ON/OFF binary classifiers.
on/ — 2000 images of screens in ON state (various UI layouts)off/ — 2000 images of screens in OFF state (dark glass with reflections, glare, fingerprints)Generated using domain randomization (arxiv:1703.06907 principles):
Images are 128x128 RGB PNG. Training resizes to 64x64 grayscale.