ONERA.ROOM

We introduce a new robotic RGBD dataset with difficult luminosity conditions: ONERA.ROOM. It comprises RGB-D data (as pairs of images) and corresponding annotations in PASCAL VOC format (xml files)

It aims at People detection, in (mostly) indoor and outdoor environments. People in the field of view can be standing, but also lying on the ground as after a fall.

Different sets are made available (Easy, Average, Hard, Search’n’Rescue, dark conditions, sunlight interference, etc…) to allow the community to test their RGBD algorithms on these challenging data. Indeed, transitions from indoor to outdoor exist in the dataset, which requires to use both modalities to detect people.

If you use ONERA.ROOM, please consider citing:

@inproceedings{guerry2017lato,

  title={{"Look At This One" Detection sharing between modality-independent classifiers for robotic people discovery}},

  author={Guerry, Joris and {Le Saux}, Bertrand and Filliat, David},

  booktitle={Proc. Eur. Conf. on Mobile Robotics (ECMR)},

  year={2017}

}

ONERA.ROOM of Joris GUERRY and Bertrand LE SAUX is available under BY-NC-SA 4.0 license.