With the market launch of the Microsoft Kinect and ASUS Xtion Pro Live, the availability of low-cost RGB-D cameras has led to a significant impact on robotics and computer vision research. Since then, we have witnessed a large grow of novel approaches using such cameras including mapping, localization, segmentation, object recognition, tracking, 3D reconstruction, human pose estimation and human robot interaction. The goal of this workshop is to establish an overview on the state-of-the-art in 3D perception from RGB-D sensors and to provide to the participants a forum for the presentation and discussion of their recent works. This will be achieved by a set of high-quality invited talks and a selection of contributed papers that will undergo a strict peer-review process. Both for the talks and the poster presentations, we will encourage the authors to provide live demonstrations of their work.

We solicit the submission of extended abstracts (2 pages) and full papers (6 pages). We expect between 10-20 submissions, of which we will select 6-8 for oral presentations. All papers will undergo a strict review process by the program committee. Furthermore, we will organize an interactive poster and demo session to stimulate discussion between the participants.

Topics for contributed papers include, but are not limited to:

• sensors and calibration
• feature descriptors and feature matching
• object segmentation, detection and tracking
• visual odometry
• localization, mapping and navigation
• 3D reconstruction
• scene understanding and sematic perception
• human pose recognition and tracking
• human robot interaction
• benchmarking and evaluation

• Andrew Davison*, Imperial College (3D reconstruction)
• Markus Vinzce*, Vienna University of Technology (object perception and grasping)
• Martial Herbert*, Carnegie Mellon University (3D-based scene understanding)
• Lennart Witzke*, Raytrix GmbH (lightfield cameras)
• Kurt Konolige, Industrial Perception/Willow Garage (visual SLAM)
• Thomas Pock, Technical University of Graz (3D reconstruction from monocular vision)

*: confirmed

The workshop is planned as a full-day event.

• 4-6 invited talks, 30-45 minutes each
• 6-8 talks of contributed papers, 20 minutes each
• poster and demo session (up to 10 posters, 45 minutes)

The proceedings of the workshop will be published on our workshop website. Furthermore, we currently consider to propose a special issue on RGB-D perception in an internationally renowed computer vision or robotics journal.