Gradient-enhanced particle filter for vision-based motion capture
(2007) 2nd Workshop on Human Motion Understanding, Modeling, Capture and Animation In Lecture Notes in Computer Science 4814. p.28-41- Abstract
Tracking of rigid and articulated objects is usually addressed within a particle filter framework or by correspondence based gradient descent methods. We combine both methods, such that (a) the correspondence based estimation gains the advantage of the particle filter and becomes able to follow multiple hypotheses while (b) the particle filter becomes able to propagate the particles in a better manner and thus gets by with a smaller number of particles. Results on noisy synthetic depth data show that the new method is able to track motion correctly where the correspondence based method fails. Further experiments with real-world stereo data underline the advantages of our coupled method.
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/aab1d899-c404-40d8-aa44-b192a4de49f9
- author
- Grest, Daniel
and Krüger, Volker
LU
- publishing date
- 2007-12-01
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Human Motion - Understanding, Modeling, Capture and Animation - Second Workshop, Human Motion 2007, Proceedings
- series title
- Lecture Notes in Computer Science
- volume
- 4814
- pages
- 14 pages
- publisher
- Springer
- conference name
- 2nd Workshop on Human Motion Understanding, Modeling, Capture and Animation
- conference location
- Rio de Janeiro, Brazil
- conference dates
- 2007-10-20 - 2007-10-20
- external identifiers
-
- scopus:38149119843
- ISSN
- 1611-3349
- 0302-9743
- ISBN
- 9783540757023
- DOI
- 10.1007/978-3-540-75703-0_3
- language
- English
- LU publication?
- no
- id
- aab1d899-c404-40d8-aa44-b192a4de49f9
- date added to LUP
- 2019-07-08 21:17:06
- date last changed
- 2025-04-04 15:19:23
@inproceedings{aab1d899-c404-40d8-aa44-b192a4de49f9, abstract = {{<p>Tracking of rigid and articulated objects is usually addressed within a particle filter framework or by correspondence based gradient descent methods. We combine both methods, such that (a) the correspondence based estimation gains the advantage of the particle filter and becomes able to follow multiple hypotheses while (b) the particle filter becomes able to propagate the particles in a better manner and thus gets by with a smaller number of particles. Results on noisy synthetic depth data show that the new method is able to track motion correctly where the correspondence based method fails. Further experiments with real-world stereo data underline the advantages of our coupled method.</p>}}, author = {{Grest, Daniel and Krüger, Volker}}, booktitle = {{Human Motion - Understanding, Modeling, Capture and Animation - Second Workshop, Human Motion 2007, Proceedings}}, isbn = {{9783540757023}}, issn = {{1611-3349}}, language = {{eng}}, month = {{12}}, pages = {{28--41}}, publisher = {{Springer}}, series = {{Lecture Notes in Computer Science}}, title = {{Gradient-enhanced particle filter for vision-based motion capture}}, url = {{http://dx.doi.org/10.1007/978-3-540-75703-0_3}}, doi = {{10.1007/978-3-540-75703-0_3}}, volume = {{4814}}, year = {{2007}}, }