Local smoothness for global optical flow

Lars Lau Rakêt

doi:10.1109/ICIP.2012.6674231

Local smoothness for global optical flow

Lars Lau Rakêt

Department of Computer Science

Abstract

We consider the problem of estimating the “smoothness parameter” that controls the tradeoff between data fidelity and regularity in optical flow estimation. We start by reviewing the problem of global estimation using the Optimal Prediction Principle (OPP) by Zimmer et al. Inspired by this technique and work on local-global optical flow we propose a simple method for fusing optical flow estimates of different smoothness by evaluating interpolation quality locally by means of L1 block match on the corresponding set of gradient images. We illustrate the method in a setting where optical flows are estimated by a TV-L1 energy. On average this procedure reduces the average endpoint error by 15% over flows estimated using the OPP, and gives flow fields that are consistently better than the single best flows with a fixed smoothness parameter.

Original language	English
Title of host publication	2012 19th IEEE International Conference on Image Processing (ICIP)
Number of pages	4
Publisher	IEEE
Publication date	2012
ISBN (Print)	978-1-4673-2534-9
ISBN (Electronic)	978-1-4673-2533-2
DOIs	https://doi.org/10.1109/ICIP.2012.6674231
Publication status	Published - 2012
Event	2012 19th IEEE International Conference on Image Processing - Orlando , United States Duration: 30 Sept 2012 → 3 Oct 2012 Conference number: 19

Conference

Conference	2012 19th IEEE International Conference on Image Processing
Number	19
Country/Territory	United States
City	Orlando
Period	30/09/2012 → 03/10/2012

Access to Document

10.1109/ICIP.2012.6674231

Cite this

@inproceedings{44fef4e33a5d425b989ca3243ce64f8a,

title = "Local smoothness for global optical flow",

abstract = "We consider the problem of estimating the “smoothness parameter” that controls the tradeoff between data fidelity and regularity in optical flow estimation. We start by reviewing the problem of global estimation using the Optimal Prediction Principle (OPP) by Zimmer et al. Inspired by this technique and work on local-global optical flow we propose a simple method for fusing optical flow estimates of different smoothness by evaluating interpolation quality locally by means of L1 block match on the corresponding set of gradient images. We illustrate the method in a setting where optical flows are estimated by a TV-L1 energy. On average this procedure reduces the average endpoint error by 15% over flows estimated using the OPP, and gives flow fields that are consistently better than the single best flows with a fixed smoothness parameter.",

author = "Rak{\^e}t, {Lars Lau}",

year = "2012",

doi = "10.1109/ICIP.2012.6674231",

language = "English",

isbn = "978-1-4673-2534-9",

booktitle = "2012 19th IEEE International Conference on Image Processing (ICIP)",

publisher = "IEEE",

note = "2012 19th IEEE International Conference on Image Processing ; Conference date: 30-09-2012 Through 03-10-2012",

}

TY - GEN

T1 - Local smoothness for global optical flow

AU - Rakêt, Lars Lau

N1 - Conference code: 19

PY - 2012

Y1 - 2012

N2 - We consider the problem of estimating the “smoothness parameter” that controls the tradeoff between data fidelity and regularity in optical flow estimation. We start by reviewing the problem of global estimation using the Optimal Prediction Principle (OPP) by Zimmer et al. Inspired by this technique and work on local-global optical flow we propose a simple method for fusing optical flow estimates of different smoothness by evaluating interpolation quality locally by means of L1 block match on the corresponding set of gradient images. We illustrate the method in a setting where optical flows are estimated by a TV-L1 energy. On average this procedure reduces the average endpoint error by 15% over flows estimated using the OPP, and gives flow fields that are consistently better than the single best flows with a fixed smoothness parameter.

AB - We consider the problem of estimating the “smoothness parameter” that controls the tradeoff between data fidelity and regularity in optical flow estimation. We start by reviewing the problem of global estimation using the Optimal Prediction Principle (OPP) by Zimmer et al. Inspired by this technique and work on local-global optical flow we propose a simple method for fusing optical flow estimates of different smoothness by evaluating interpolation quality locally by means of L1 block match on the corresponding set of gradient images. We illustrate the method in a setting where optical flows are estimated by a TV-L1 energy. On average this procedure reduces the average endpoint error by 15% over flows estimated using the OPP, and gives flow fields that are consistently better than the single best flows with a fixed smoothness parameter.

U2 - 10.1109/ICIP.2012.6674231

DO - 10.1109/ICIP.2012.6674231

M3 - Article in proceedings

SN - 978-1-4673-2534-9

BT - 2012 19th IEEE International Conference on Image Processing (ICIP)

PB - IEEE

T2 - 2012 19th IEEE International Conference on Image Processing

Y2 - 30 September 2012 through 3 October 2012

ER -

Local smoothness for global optical flow

Abstract

Conference

Access to Document

Fingerprint

Cite this