TV-L1 optical flow for vector valued images

Lars Lau Rakêt; Lars Roholm; Mads Nielsen; Francois Bernard Lauze

doi:10.1007/978-3-642-23094-3_24

TV-L¹ optical flow for vector valued images

Lars Lau Rakêt, Lars Roholm, Mads Nielsen, Francois Bernard Lauze

Department of Computer Science

16 Citations (Scopus)

Abstract

The variational TV-L1 framework has become one of the most popular and successful approaches for calculating optical flow. One reason for the popularity is the very appealing properties of the two terms in the energy formulation of the problem, the robust L1-norm of the data fidelity term combined with the total variation (TV) regular- ization that smoothes the flow, but preserve strong discontinuities such as edges. Specifically the approach of Zach et al. [1] has provided a very clean and efficient algorithm for calculating TV-L1 optical flows between grayscale images. In this paper we propose a generalized algorithm that works on vector valued images, by means of a generalized projection step. We give examples of calculations of flows for a number of multi- dimensional constancy assumptions, e.g. gradient and RGB, and show how the developed methodology expands to any kind of vector valued images. The resulting algorithms have the same degree of parallelism as the case of one-dimensional images, and we have produced an efficient GPU implementation, that can take vector valued images with vectors of any dimension. Finally we demonstrate how these algorithms generally produce better flows than the original algorithm.

Original language	English
Title of host publication	Energy Minimization Methods in Computer Vision and Pattern Recognition : 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings
Editors	Yuri Boykov, Fredrik Kahl, Victor Lempitsky, Frank R. Schmidt
Number of pages	15
Publisher	Springer
Publication date	2011
Pages	329-343
ISBN (Print)	978-3-642-23093-6
ISBN (Electronic)	978-3-642-23094-3
DOIs	https://doi.org/10.1007/978-3-642-23094-3_24
Publication status	Published - 2011
Event	8th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition - Sankt Petersborg, Russian Federation Duration: 25 Jul 2011 → 27 Jul 2011 Conference number: 8

Conference

Conference	8th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition
Number	8
Country/Territory	Russian Federation
City	Sankt Petersborg
Period	25/07/2011 → 27/07/2011

Series	Lecture notes in computer science
Volume	6819
ISSN	0302-9743

Access to Document

10.1007/978-3-642-23094-3_24

Cite this

Rakêt, L. L., Roholm, L., Nielsen, M., & Lauze, F. B. (2011). TV-L¹ optical flow for vector valued images. In Y. Boykov, F. Kahl, V. Lempitsky, & F. R. Schmidt (Eds.), Energy Minimization Methods in Computer Vision and Pattern Recognition: 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings (pp. 329-343). Springer. https://doi.org/10.1007/978-3-642-23094-3_24

TV-L¹ optical flow for vector valued images. / Rakêt, Lars Lau; Roholm, Lars; Nielsen, Mads et al.
Energy Minimization Methods in Computer Vision and Pattern Recognition: 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings. ed. / Yuri Boykov; Fredrik Kahl; Victor Lempitsky; Frank R. Schmidt. Springer, 2011. p. 329-343 (Lecture notes in computer science, Vol. 6819).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Rakêt, LL, Roholm, L, Nielsen, M & Lauze, FB 2011, TV-L¹ optical flow for vector valued images. in Y Boykov, F Kahl, V Lempitsky & FR Schmidt (eds), Energy Minimization Methods in Computer Vision and Pattern Recognition: 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings. Springer, Lecture notes in computer science, vol. 6819, pp. 329-343, 8th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition, Sankt Petersborg, Russian Federation, 25/07/2011. https://doi.org/10.1007/978-3-642-23094-3_24

Rakêt LL, Roholm L, Nielsen M , Lauze FB. TV-L¹ optical flow for vector valued images. In Boykov Y, Kahl F, Lempitsky V, Schmidt FR, editors, Energy Minimization Methods in Computer Vision and Pattern Recognition: 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings. Springer. 2011. p. 329-343. (Lecture notes in computer science, Vol. 6819). doi: 10.1007/978-3-642-23094-3_24

Rakêt, Lars Lau ; Roholm, Lars ; Nielsen, Mads et al. / TV-L¹ optical flow for vector valued images. Energy Minimization Methods in Computer Vision and Pattern Recognition: 8th International Conference, EMMCVPR 2011, St. Petersburg, Russia, July 25-27, 2011. Proceedings. editor / Yuri Boykov ; Fredrik Kahl ; Victor Lempitsky ; Frank R. Schmidt. Springer, 2011. pp. 329-343 (Lecture notes in computer science, Vol. 6819).

@inproceedings{b473b96daa554cbb8855f08cb58d0e1d,

title = "TV-L1 optical flow for vector valued images",

abstract = "The variational TV-L1 framework has become one of the most popular and successful approaches for calculating optical flow. One reason for the popularity is the very appealing properties of the two terms in the energy formulation of the problem, the robust L1-norm of the data fidelity term combined with the total variation (TV) regular- ization that smoothes the flow, but preserve strong discontinuities such as edges. Specifically the approach of Zach et al. [1] has provided a very clean and efficient algorithm for calculating TV-L1 optical flows between grayscale images. In this paper we propose a generalized algorithm that works on vector valued images, by means of a generalized projection step. We give examples of calculations of flows for a number of multi- dimensional constancy assumptions, e.g. gradient and RGB, and show how the developed methodology expands to any kind of vector valued images. The resulting algorithms have the same degree of parallelism as the case of one-dimensional images, and we have produced an efficient GPU implementation, that can take vector valued images with vectors of any dimension. Finally we demonstrate how these algorithms generally produce better flows than the original algorithm.",

author = "Rak{\^e}t, {Lars Lau} and Lars Roholm and Mads Nielsen and Lauze, {Francois Bernard}",

year = "2011",

doi = "10.1007/978-3-642-23094-3_24",

language = "English",

isbn = "978-3-642-23093-6",

series = "Lecture notes in computer science",

publisher = "Springer",

pages = "329--343",

editor = "Yuri Boykov and Fredrik Kahl and Victor Lempitsky and Schmidt, {Frank R.}",

booktitle = "Energy Minimization Methods in Computer Vision and Pattern Recognition",

note = "8th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition ; Conference date: 25-07-2011 Through 27-07-2011",

}

TY - GEN

T1 - TV-L1 optical flow for vector valued images

AU - Rakêt, Lars Lau

AU - Roholm, Lars

AU - Nielsen, Mads

AU - Lauze, Francois Bernard

N1 - Conference code: 8

PY - 2011

Y1 - 2011

N2 - The variational TV-L1 framework has become one of the most popular and successful approaches for calculating optical flow. One reason for the popularity is the very appealing properties of the two terms in the energy formulation of the problem, the robust L1-norm of the data fidelity term combined with the total variation (TV) regular- ization that smoothes the flow, but preserve strong discontinuities such as edges. Specifically the approach of Zach et al. [1] has provided a very clean and efficient algorithm for calculating TV-L1 optical flows between grayscale images. In this paper we propose a generalized algorithm that works on vector valued images, by means of a generalized projection step. We give examples of calculations of flows for a number of multi- dimensional constancy assumptions, e.g. gradient and RGB, and show how the developed methodology expands to any kind of vector valued images. The resulting algorithms have the same degree of parallelism as the case of one-dimensional images, and we have produced an efficient GPU implementation, that can take vector valued images with vectors of any dimension. Finally we demonstrate how these algorithms generally produce better flows than the original algorithm.

AB - The variational TV-L1 framework has become one of the most popular and successful approaches for calculating optical flow. One reason for the popularity is the very appealing properties of the two terms in the energy formulation of the problem, the robust L1-norm of the data fidelity term combined with the total variation (TV) regular- ization that smoothes the flow, but preserve strong discontinuities such as edges. Specifically the approach of Zach et al. [1] has provided a very clean and efficient algorithm for calculating TV-L1 optical flows between grayscale images. In this paper we propose a generalized algorithm that works on vector valued images, by means of a generalized projection step. We give examples of calculations of flows for a number of multi- dimensional constancy assumptions, e.g. gradient and RGB, and show how the developed methodology expands to any kind of vector valued images. The resulting algorithms have the same degree of parallelism as the case of one-dimensional images, and we have produced an efficient GPU implementation, that can take vector valued images with vectors of any dimension. Finally we demonstrate how these algorithms generally produce better flows than the original algorithm.

U2 - 10.1007/978-3-642-23094-3_24

DO - 10.1007/978-3-642-23094-3_24

M3 - Article in proceedings

SN - 978-3-642-23093-6

T3 - Lecture notes in computer science

SP - 329

EP - 343

BT - Energy Minimization Methods in Computer Vision and Pattern Recognition

A2 - Boykov, Yuri

A2 - Kahl, Fredrik

A2 - Lempitsky, Victor

A2 - Schmidt, Frank R.

PB - Springer

T2 - 8th International Conference on Energy Minimization Methods in Computer Vision and Pattern Recognition

Y2 - 25 July 2011 through 27 July 2011