Details
| Original language | English |
|---|---|
| Title of host publication | Pattern Recognition |
| Subtitle of host publication | 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings |
| Editors | Bastian Leibe, Juergen Gall, Peter Gehler |
| Publisher | Springer Verlag |
| Pages | 16-28 |
| Number of pages | 13 |
| ISBN (electronic) | 978-3-319-24947-6 |
| ISBN (print) | 9783319249469 |
| Publication status | Published - 3 Nov 2015 |
| Event | 37th German Conference on Pattern Recognition, GCPR 2015 - Aachen, Germany Duration: 7 Oct 2015 → 10 Oct 2015 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 9358 |
| ISSN (Print) | 0302-9743 |
| ISSN (electronic) | 1611-3349 |
Abstract
We propose to look at large-displacement optical flow from a discrete point of view. Motivated by the observation that sub-pixel accuracy is easily obtained given pixel-accurate optical flow, we conjecture that computing the integral part is the hardest piece of the problem. Consequently, we formulate optical flow estimation as a discrete inference problem in a conditional random field, followed by sub-pixel refinement. Naive discretization of the 2D flow space, however, is intractable due to the resulting size of the label set. In this paper, we therefore investigate three different strategies, each able to reduce computation and memory demands by several orders of magnitude. Their combination allows us to estimate large-displacement optical flow both accurately and efficiently and demonstrates the potential of discrete optimization for optical flow. We obtain state-of-the-art performance on MPI Sintel and KITTI.
ASJC Scopus subject areas
- Mathematics(all)
- Theoretical Computer Science
- Computer Science(all)
- General Computer Science
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Pattern Recognition : 37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings. ed. / Bastian Leibe; Juergen Gall; Peter Gehler. Springer Verlag, 2015. p. 16-28 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9358).
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Discrete Optimization for Optical Flow
AU - Menze, Moritz
AU - Heipke, Christian
AU - Geiger, Andreas
PY - 2015/11/3
Y1 - 2015/11/3
N2 - We propose to look at large-displacement optical flow from a discrete point of view. Motivated by the observation that sub-pixel accuracy is easily obtained given pixel-accurate optical flow, we conjecture that computing the integral part is the hardest piece of the problem. Consequently, we formulate optical flow estimation as a discrete inference problem in a conditional random field, followed by sub-pixel refinement. Naive discretization of the 2D flow space, however, is intractable due to the resulting size of the label set. In this paper, we therefore investigate three different strategies, each able to reduce computation and memory demands by several orders of magnitude. Their combination allows us to estimate large-displacement optical flow both accurately and efficiently and demonstrates the potential of discrete optimization for optical flow. We obtain state-of-the-art performance on MPI Sintel and KITTI.
AB - We propose to look at large-displacement optical flow from a discrete point of view. Motivated by the observation that sub-pixel accuracy is easily obtained given pixel-accurate optical flow, we conjecture that computing the integral part is the hardest piece of the problem. Consequently, we formulate optical flow estimation as a discrete inference problem in a conditional random field, followed by sub-pixel refinement. Naive discretization of the 2D flow space, however, is intractable due to the resulting size of the label set. In this paper, we therefore investigate three different strategies, each able to reduce computation and memory demands by several orders of magnitude. Their combination allows us to estimate large-displacement optical flow both accurately and efficiently and demonstrates the potential of discrete optimization for optical flow. We obtain state-of-the-art performance on MPI Sintel and KITTI.
UR - http://www.scopus.com/inward/record.url?scp=84952312999&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-24947-6_2
DO - 10.1007/978-3-319-24947-6_2
M3 - Conference contribution
AN - SCOPUS:84952312999
SN - 9783319249469
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 16
EP - 28
BT - Pattern Recognition
A2 - Leibe, Bastian
A2 - Gall, Juergen
A2 - Gehler, Peter
PB - Springer Verlag
T2 - 37th German Conference on Pattern Recognition, GCPR 2015
Y2 - 7 October 2015 through 10 October 2015
ER -