Real-Time Pose Estimation Using Constrained Dynamics

Pose estimation in the context of human motion analysis is the process of approximating the body configuration in each frame of a motion sequence. We propose a novel pose estimation method based on fitting a skeletal model to tree structures built from skeletonised visual hulls reconstructed from multi-view video. The pose is estimated independently in each frame, hence the method can recover from errors in previous frames, which overcomes some problems of tracking. Publically available datasets were used to evaluate the method. On real data the method performs at a framerate of $\sim\!14$ fps. Using synthetic data the positions of the joints were determined with a mean error of $\sim\!6$ cm.

[1]  Gutemberg Guerra-Filho.  A General Motion Representation - Exploring the Intrinsic Viewpoint of a Motion , 2012, GRAPP/IVAPP.

[2]  Yen-Lin Chen,et al.  3D Reconstruction of Human Motion and Skeleton from Uncalibrated Monocular Video , 2009, ACCV.

[3]  Ioannis Pitas,et al.  The i3DPost Multi-View and 3D Human Action/Interaction Database , 2009, 2009 Conference for Visual Media Production.

[4]  Adrian Hilton,et al.  A survey of advances in vision-based human motion capture and analysis , 2006, Comput. Vis. Image Underst..

[5]  Ronald Poppe,et al.  Vision-based human motion analysis: An overview , 2007, Comput. Vis. Image Underst..

[6]  Andrea Fusiello,et al.  Tracking Human Motion with Multiple Cameras Using an Articulated Model , 2009, MIRAGE.

[7]  Adrian Hilton,et al.  Real-time Pose Estimation using Tree Structures Built from Skeletonised Volume Sequences , 2012, VISAPP.

[8]  Hongbin Zha,et al.  Computer Vision - ACCV 2009, 9th Asian Conference on Computer Vision, Xi'an, China, September 23-27, 2009, Revised Selected Papers, Part III , 2010, Asian Conference on Computer Vision.

[9]  Marc Cavazza,et al.  A constrained inverse kinematics technique for real-time motion capture animation , 1999, The Visual Computer.

[10]  Bruno Raffin,et al.  3D Skeleton-Based Body Pose Recovery , 2006, Third International Symposium on 3D Data Processing, Visualization, and Transmission (3DPVT'06).

[11]  Thomas Kastenmeier,et al.  Numerical robot kinematics based on stochastic and molecular simulation methods , 1996, Robotica.

[12]  Horst Bischof,et al.  Skeletal Graph Based Human Pose Estimation in Real-Time , 2011, BMVC.

[13]  Saïda Bouakaz,et al.  Real-Time and Markerless 3D Human Motion Capture Using Multiple Views , 2007, Workshop on Human Motion.

[14]  Rama Chellappa,et al.  Model Driven Segmentation of Articulating Humans in Laplacian Eigenspace , 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[15]  Rune Havnung Bakken,et al.  Using Synthetic Data for Planning, Development and Evaluation of Shape-from-Silhouette Based Human Motion Capture Methods , 2012, ISVC.

[16]  Toby Howard,et al.  Real-time 3-D human body tracking using learnt models of behaviour , 2008, Comput. Vis. Image Underst..

[17]  S. Buss.  Introduction to Inverse Kinematics with Jacobian Transpose , Pseudoinverse and Damped Least Squares methods , 2004 .

[18]  Bodo Rosenhahn,et al.  Human Motion - Understanding, Modeling, Capture and Animation, Second Workshop, Human Motion 2007, Rio de Janeiro, Brazil, October 20, 2007, Proceedings , 2007, Workshop on Human Motion.