Details
| Original language | English |
|---|---|
| Title of host publication | Articulated Motion and Deformable Objects - 5th International Conference, AMDO 2008, Proceedings |
| Pages | 102-111 |
| Number of pages | 10 |
| Publication status | Published - 2008 |
| Externally published | Yes |
| Event | 5th International Conference on Articulated Motion and Deformable Objects, AMDO 2008 - Port d'Andratx, Mallorca, Spain Duration: 9 Jul 2008 → 11 Jul 2008 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 5098 LNCS |
| ISSN (Print) | 0302-9743 |
| ISSN (electronic) | 1611-3349 |
Abstract
Self-occlusion is a common problem in silhouette based motion capture, which often results in ambiguous pose configurations. In most works this is compensated by a priori knowledge about the motion or the scene, or by the use of multiple cameras. Here we suggest to overcome this problem by splitting the surface model of the object and tracking the silhouette of each part rather than the whole object. The splitting can be done automatically by comparing the appearance of the different parts with the Jensen-Shannon divergence. Tracking is then achieved by maximizing the appearance differences of all involved parts and the background simultaneously via gradient descent. We demonstrate the improvements with tracking results from simulated and real world scenes.
Keywords
- Computer vision, Human motion analysis, Kinematic chain, Pose estimation, Tracking
ASJC Scopus subject areas
- Mathematics(all)
- Theoretical Computer Science
- Computer Science(all)
- General Computer Science
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Articulated Motion and Deformable Objects - 5th International Conference, AMDO 2008, Proceedings. 2008. p. 102-111 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5098 LNCS).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Dealing with self-occlusion in region based motion capture by means of internal regions
AU - Schmaltz, Christian
AU - Rosenhahn, Bodo
AU - Brox, Thomas
AU - Weickert, Joachim
AU - Wietzke, Lennart
AU - Sommer, Gerald
N1 - Funding information: We acknowledge funding by the German Research Foundation under the projects We 2602/5-1 and SO 320/4-2, and by the Max-Planck Center for Visual Computing and Communication.
PY - 2008
Y1 - 2008
N2 - Self-occlusion is a common problem in silhouette based motion capture, which often results in ambiguous pose configurations. In most works this is compensated by a priori knowledge about the motion or the scene, or by the use of multiple cameras. Here we suggest to overcome this problem by splitting the surface model of the object and tracking the silhouette of each part rather than the whole object. The splitting can be done automatically by comparing the appearance of the different parts with the Jensen-Shannon divergence. Tracking is then achieved by maximizing the appearance differences of all involved parts and the background simultaneously via gradient descent. We demonstrate the improvements with tracking results from simulated and real world scenes.
AB - Self-occlusion is a common problem in silhouette based motion capture, which often results in ambiguous pose configurations. In most works this is compensated by a priori knowledge about the motion or the scene, or by the use of multiple cameras. Here we suggest to overcome this problem by splitting the surface model of the object and tracking the silhouette of each part rather than the whole object. The splitting can be done automatically by comparing the appearance of the different parts with the Jensen-Shannon divergence. Tracking is then achieved by maximizing the appearance differences of all involved parts and the background simultaneously via gradient descent. We demonstrate the improvements with tracking results from simulated and real world scenes.
KW - Computer vision
KW - Human motion analysis
KW - Kinematic chain
KW - Pose estimation
KW - Tracking
UR - http://www.scopus.com/inward/record.url?scp=49049103772&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-70517-8_11
DO - 10.1007/978-3-540-70517-8_11
M3 - Conference contribution
AN - SCOPUS:49049103772
SN - 3540705163
SN - 9783540705161
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 102
EP - 111
BT - Articulated Motion and Deformable Objects - 5th International Conference, AMDO 2008, Proceedings
T2 - 5th International Conference on Articulated Motion and Deformable Objects, AMDO 2008
Y2 - 9 July 2008 through 11 July 2008
ER -