Dataset Open Access

UNICITY: A depth maps database for people detection in security airlocks

Joël Dumoulin; Olivier Canévet; Michael Villamizar; Hugo Nunes; Omar Abou Khaled; Elena Mugellini; Fabrice Moscheni; Jean-Marc Odobez

UNICITY: A depth maps database for people detection in security airlocks.

UNICITY consists of 58k images collected from 65 recorded sequences with one or two people performing different behaviors including attacks and trickeries, like for instance tailgating (when a person walks very close to another to get into a restricted area). It also provides full annotation of people such as the location of head and shoulders. As as result, UNICITY is perfectly suited for training and adapting machine learning algorithms for video surveillance applications.

Main Features:

  • UNICITY consists of 58k images using two depth sensors.
  • 65 recorded sequences with one or two people performing different behaviors such as attacks and tailgating.
  • UNICITY also provides code for evaluation and visualization, and full annotation of people such as the location of head and shoulders.
  • This new dataset is perfectly suited for training and adapting machine learning algorithms for video surveillance applications.

Citation:

Please cite the following paper if you use the UNICITY dataset in your work (papers, articles, reports, books, software, etc):

  • UNICITY: A depth maps database for people detection in security airlocks. J. Dumoulin, O. Canevet, M. Villamizar, H. Nunes, O.A. Khaled, E. Mugellini, F. Moscheni, and J.M Odobez. International Conference on Advanced Video and Signal-based Surveillance Workshop (AVSSW). November 2018.

Contributors:

  • Joël Dumoulin, HumanTech Institute, HES-SO Fribourg, Switzerland.
  • Olivier Canévet, Idiap Research Institute, Martigny, Switzerland.
  • Michael Villamizar, Idiap Research Institute, Martigny, Switzerland.
  • Hugo Nunes, Fastcom Technology SA, Lausanne, Switzerland.
  • Omar Abou Khaled, HumanTech Institute, HES-SO Fribourg, Switzerland.
  • Elena Mugellini, HumanTech Institute, HES-SO Fribourg, Switzerland.
  • Fabrice Moscheni, Fastcom Technology SA, Lausanne, Switzerland.
  • Jean-Marc Odobez, Idiap Research Institute, Martigny, Switzerland.

Acknowledgement:

The work was supported by Innosuisse, the Swiss innovation agency, through the UNICITY (3D scene understanding through machine learning to secure entrance zones) project.

Links:

Next links contain additional information about the dataset:

  • Innosuisse UNICITY project: [link]
  • Paper describing the dataset: [link]
  • Video presenting the dataset: [link]
  • Paper using the dataset for counting people and detecting intrusions: [link]
    • WatchNet: Efficient and Depth-based Network for People Detection in Video Surveillance Systems.
      M. Villamizar, A. Martinez-Gonzalez, O. Canevet and J-M. Odobez.
      International Conference on Advanced Video and Signal-based Surveillance (AVSS) - 2018.

Contact:

For any questions, please contact:

  • Michael Villamizar, Idiap Research Institute, Martigny -Switzerland

Files (13.6 GB)
Name Size
code.zip
md5:87c46855cfec64d6a0bc5437c1f7705e
55.2 kB Download
data.zip
md5:044ca71034bb35d6d25152c1f254638a
13.6 GB Download
imgs.zip
md5:934028aad23a39387e0a75a6a9ce92d9
2.8 MB Download
README.md
md5:9230af16464e6f1d0a81187cf62b05f9
10.5 kB Download
requirements.txt
md5:5ff4981dcc6a41b83f21241e67ed1ade
35 Bytes Download
614
489
views
downloads
All versions This version
Views 614618
Downloads 489489
Data volume 5.1 TB5.1 TB
Unique views 560563
Unique downloads 129129

Share

Cite as