ViHASi: virtual human action silhouette data for the performance evaluation of silhouette-based action recognition methods

ABSTRACT

We introduce a large body of virtual human action silhouette (ViHASi) data generated recently for the purpose of evaluating a family of action recognition methods. These are the silhouette-based human action recognition methods. This synthetic multi-camera video data-set consists of 20 action classes, 9 actors and up to 40 synchronized perspective cameras. The data-set has been made available online for other researchers to use. In order to demonstrate the usefulness of the ViHASi data we make use of our recent action recognition method that is simple and relatively fast. Moreover, to deal with long video sequences containing several action samples, a practical temporal segmentation algorithm is introduced and tested that is tightly coupled with the action recognition method used. Our experimental methodologies provides a reasonable platform for quantitatively comparing silhouette-based action recognition methods.

REFERENCES

• 1: Aaron F. Bobick , James W. Davis, The Recognition of Human Movement Using Temporal Templates, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.23 n.3, p.257-267, March 2001 [doi>10.1109/34.910878]
• 2: Alexei A. Efros , Alexander C. Berg , Greg Mori , Jitendra Malik, Recognizing Action at a Distance, Proceedings of the Ninth IEEE International Conference on Computer Vision, p.726, October 13-16, 2003
• 3: Lena Gorelick , Moshe Blank , Eli Shechtman , Michal Irani , Ronen Basri, Actions as Space-Time Shapes, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.29 n.12, p.2247-2253, December 2007 [doi>10.1109/TPAMI.2007.70711]
• 4: I. Laptev and P. Perez, "Retrieving Actions in Movies," IEEE ICCV, 2007: 1--8.
• 5: F. Lv and R. Nevatia, "Single View Human Action Recognition using Key Pose Matching and Viterbi Path Searching," IEEE CVPR, 2007: 1--8.
• 6: MuHAVi-MAS: Multicamera Human Action Video and Manually Annotated Silhouette Data, http://dipersec.king.ac.uk/MuHAVi-MAS/.
• 7: J. C. Niebles and L. Fei-Fei, "A Hierarchical Model of Shape and Appearance for Human Action Classification," IEEE CVPR, 2007: 1--8.
• 8: A. Ogale, A. Karapurkar and Y. Aloimonos, "View-Invariant Modeling and Recognition of Human Actions Using Grammars," Dynamical Vision, LNCS 4358, 2006: 115--126.
• 9: F. Qureshi and D. Terzopoulos, "Surveillance in Virtual Reality: System Design and Multicamera Control," IEEE CVPR, 2007: 1--8.
• 10: H. Ragheb, S. Velastin, P. Remagnino and T. Ellis, "Human Action Recognition using Robust Power Spectrum Features," IEEE ICIP, 2008, to appear.
• 11: Cen Rao , Alper Yilmaz , Mubarak Shah, View-Invariant Representation and Recognition of Actions, International Journal of Computer Vision, v.50 n.2, p.203-226, November 2002 [doi>10.1023/A:1020350100748]
• 12: Liu Ren , Alton Patrick , Alexei A. Efros , Jessica K. Hodgins , James M. Rehg, A data-driven approach to quantifying natural human motion, ACM Transactions on Graphics (TOG), v.24 n.3, July 2005
• 13: Liu Ren , Gregory Shakhnarovich , Jessica K. Hodgins , Hanspeter Pfister , Paul Viola, Learning silhouette features for control of human motion, ACM Transactions on Graphics (TOG), v.24 n.4, p.1303-1331, October 2005 [doi>10.1145/1095878.1095882]
• 14: G. R. Taylor, A. J. Chosak and P. C. Brewer, "OVVV: Using Virtual Worlds to Design and Evaluate Surveillance Systems," IEEE CVPR, 2007: 1--8.
• 15: L. Wang and D. Suter, "Learning and Matching of Dynamic Shape Manifolds for Human Action Recognition," IEEE Trans. Im. Proc., 16(6), 2007: 1646--1661.
• 16: D. Weinland, E. Boyer and R. Ronfard, "Action Recognition from Arbitrary Views using 3D Exemplars," IEEE ICCV, 2007: 1--7.
• 17: Daniel Weinland , Remi Ronfard , Edmond Boyer, Free viewpoint action recognition using motion history volumes, Computer Vision and Image Understanding, v.104 n.2, p.249-257, November 2006 [doi>10.1016/j.cviu.2006.07.013]
• 18: Alper Yilmaz , Mubarak Shah, Actions Sketch: A Novel Action Representation, Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 1, p.984-989, June 20-26, 2005 [doi>10.1109/CVPR.2005.58]
• 19: ViHASi: Virtual Human Action Silhouette Data, http://dipersec.king.ac.uk/VIHASI/.

INDEX TERMS

Primary Classification:
  J. Computer Applications
  J.m MISCELLANEOUS

General Terms:
Experimentation

Keywords:
action recognition, multiple cameras, performance evaluation, silhouette, synthetic data, video