The Image Group – University of Copenhagen

The Image Group
Søren Hauberg

Søren Hauberg, Post doc.

Photograph of Søren Hauberg Photograph of my pig Output of our Stick tracker
Email: hauberg@diku.dk · Google Scholar · Phone: (+45) 24 84 53 51 · Office: E04.015B

I Have Moved!

Since March 1st, 2012, I have been a post doc at the Max Planck Institute for Intelligent Systems. This page is no longer updated! Please see my MPI web page for updates.

Ego

I am a post doc at the Department of Computer Science at the University of Copenhagen in Denmark. I am a member of the Image Group.

I work on articulated tracking. More specifically, I work on general frameworks for modelling articulalted human motion spatially. This work is in stark contrast to most modern approaches, where intrinsic properties, such as joint angles, are modelled. We take a more extrinsic approach and model spatial positions of bone end-points. As the world we live in is inherently spatial, this approach allows to model interactions with the surroundings trivially. We can also do data-driven predictions in a straight-forward manner. For more details see our publications or contact me directly.

Publications

A full list of accepted publications is available below. For more details I also have a Curriculum Vitae available. I try to keep both up to date, though some delay is to be expected.

Journal Papers

  • Søren Hauberg, Stefan Sommer and Kim Steenstrup Pedersen: Natural Metrics and Least-Committed Priors for Articulated Tracking. In Image and Vision Computing (IMAVIS) 2011.
    Publishers website · PDF · BibTeX · Code.
  • Søren Hauberg and Kim Steenstrup Pedersen: Predicting Articulated Human Motion from Spatial Processes. In International Journal of Computer Vision (IJCV) 2011.
    Publishers website · Paper website · PDF · BibTeX · Code.
  • Mads Nielsen, Peter Johansen, Andrew Jackson, Benny Lautrup and Søren Hauberg: Brownian Warps for Non-Rigid Registration. In Journal of Mathematical Imaging and Vision (JMIV) 2008.
    Publishers website · PDF · BibTeX.
  • Søren Hauberg and Jakob Sloth: An Efficient Algorithm for Modelling Duration in Hidden Markov Models, with a Dramatic Application. In Journal of Mathematical Imaging and Vision (JMIV) 2008.
    Publishers website · Paper website · PDF · BibTeX.

Conference Papers

  • Aasa Feragen, Søren Hauberg, Mads Nielsen and Francois Lauze: Means in spaces of tree-like shapes. In International Conference on Computer Vision (ICCV) 2011 (to appear).
    PDF · Supplementary Material · BibTeX.
  • Søren Hauberg and Kim Steenstrup Pedersen: Data-Driven Importance Distributions for Articulated Tracking. In Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR) 2011.
    PDF · Supplementary Material · BibTeX · Code.
  • Peter Mysling, Søren Hauberg and Kim Steenstrup Pedersen: An Empirical Study on the Performance of Spectral Manifold Learning Techniques. International Conference on Artificial Neural Networks (ICANN) 2011.
    PDF · BibTeX.
  • Anders Boesen Lindbo Larsen, Søren Hauberg and Kim Steenstrup Pedersen: Unscented Kalman Filtering for Articulated Human Tracking. At Scandinavian Conference on Image Analysis (SCIA) 2011.
    Paper website · PDF · BibTeX.
  • Søren Hauberg and Kim Steenstrup Pedersen: Stick It! Articulated Tracking using Spatial Rigid Object Priors. In Asian Conference on Computer Vision (ACCV) 2010.
    Paper website · PDF · BibTeX · Code.
  • Stefan Sommer, Francois Lauze, Søren Hauberg and Mads Nielsen: Manifold Valued Statistics, Exact Principal Geodesic Analysis and the Effect of Linear Approximations. In European Conference on Computer Vision (ECCV) 2010.
    PDF · BibTeX.
  • Søren Hauberg, Stefan Sommer and Kim Steenstrup Pedersen: Gaussian-like Spatial Priors for Articulated Tracking. In European Conference on Computer Vision (ECCV) 2010.
    Paper website · PDF · BibTeX · Code.
  • Rune Møllegaard Friborg, Søren Hauberg and Kenny Erleben: GPU Accelerated Likelihoods for Stereo-Based Articulated Tracking. At the CVGPU workshop at European Conference on Computer Vision (ECCV) 2010.
    Workshop website · PDF · BibTeX.
  • Morten Engell-Nørregård, Søren Hauberg, Jerome Lapuyade, Kenny Erleben and Kim Steenstrup Pedersen: Interactive Inverse Kinematics for Monocular Motion Estimation. At the 6th Workshop on Virtual Reality Interaction and Physical Simulation (VRIPHYS), 2009.
    Conference website · Paper website · BibTeX.
  • Søren Hauberg, Jerome Lapuyade, Morten Engell-Nørregård, Kenny Erleben and Kim Steenstrup Pedersen: Three Dimensional Monocular Human Motion Analysis in End-Effector Space. In Energy Minimization Methods in Computer Vision and Pattern Recognition (EMMCVPR) 2009.
    Publishers website · Paper website · PDF · BibTeX.

Books

  • John W. Eaton, David Bateman and Søren Hauberg: GNU Octave Manual Version 3. Network Theory Ltd, ISBN-13: 978-0954612061.
    Publishers website · PDF is distributed with GNU Octave · BibTeX.

Other Publications

  • Søren Hauberg and Kim Steenstrup Pedersen: HUMIM Software for Articulated Tracking. Technical report, January 2012.
    PDF · BibTeX.
  • Søren Hauberg: Spatial Models of Human Motion (PhD Thesis), 2011.
    The thesis is currently under review by the PhD commitee; please contact me if you want a copy.
    BibTeX.
  • Søren Hauberg and Kim Steenstrup Pedersen: A Physically Natural Metric for Human Motion and the Associated Brownian Motion Model. At the 1st IEEE Workshop on Kernels and Distances for Computer Vision (ICCV workshop), 2011.
    Workshop website · BibTeX.
  • Aasa Feragen, Mads Nielsen, Søren Hauberg, Pechin Lo, Marleen de Bruijne and François Lauze: A geometric framework for statistics on trees. Technical report, April 2011.
    PDF · BibTeX.
  • Søren Hauberg, Bente Rona Jensen, Morten Engell-Nørregård, Kenny Erleben and Kim Steenstrup Pedersen: Dense Marker-less Three Dimensional Motion Capture. At Virtual Vistas; Eleventh International Symposium on the 3D Analysis of Human Movement, 2010.
    Conference website · BibTeX.

Projects and Software

I am a member of the HUMIM research project. My contributions to this project is mainly in the development of an articulated tracking system. Our main goal is to build a system that is easy to use, i.e. should require no calibration, be robust with respect to light and so forth. Our software is available here.

In the development of the tracking system we are heavily depending on the OpenTissue C++ library. This provides us with basic data structures for representating skeletons, algorithms for solving both forward and inverse kinematics and much more.

I am also a developer of GNU Octave, which provides an excellent tool for rapid prototyping. I highly recommend using this programming language if you think science should be build on software that can be legally studied and modified.