Mobility Aids Dataset

This dataset contains over 17'000 annotated RGB-D images, containing people categorized according to the mobility aids they use. It was collected in the facilities of the Faculty of Engineering of the University of Freiburg and in a hospital in Frankfurt. The dataset includes category labels as well as 2D image bounding box labels and 3D centroid depth labels for each bounding box. 

The dataset is subdevided into sets for training (more DepthJet training examples than RGB because of privacy concerns in the hospital) and 2 different test sets: test set 1 and test set 2. Test set 2 contains 4 sequences and includes occluded objects.   

This dataset is provided for research purposes only. Any commercial use is prohibited. If you use the dataset please cite our paper

@INPROCEEDINGS{vasquez17ecmr,
  author = {Andres Vasquez and Marina Kollmitz and Andreas Eitel and Wolfram Burgard},
  title = {Deep Detection of People and their Mobility Aids for a Hospital Robot},
  booktitle={Proceedings of the IEEE European Conference on Mobile Robotics (ECMR)},
  year = 2017
}

Folders:

# Images_RGB #
- RGB images, 3 channels, 960x540 pixel

# Images_Depth #
- depth images, 1 channel, 960x540 pixel

# Images_DepthJet #
- color-encoded depth images, 3 channels, 960x540 pixel

# Annotations_RGB #
- annotations for the RGB images, train set and test set 1

# Annotations_RGB_TestSet2 #
- annotations for the RGB images, test set 2
- includes occluded objects

# Annotations_DepthJet #
- annotations for the DepthJet images, train set and test set 1. 
- labels are generally different from RGB because some people are not visible in DepthJet, due to the limited depth range of the camera  

# Annotations_DepthJet_TestSet2 #
- annotations for the DepthJet images, test set 2
- labels are generally different from RGB because some people are not visible in DepthJet, due to the limited depth range of the camera 
- includes occluded objects

# ImageSets #
- list of files for the train and test sets

# odometry_TestSet2 #
- odometry information for test set 2, can be used for tracking.
- transformation between odometry frame and robot base frame

Files:

# camera_calibration.txt #
- intrinsic camera calibration values

# trafo_base_to_cam.yml #
- (static) transformation between robot base and camera