Optical flow using textures

M. A. Arredondo, K. Lebart, D. Lane

Research output: Contribution to journalArticle

Abstract

Motion estimation is a key problem in the analysis of image sequences. From a sequence of images we can only estimate an approximation of the image motion field called optical flow. We propose to improve optical flow estimation by including information from images of textural features. We compute the optical flow from intensity and textural images from first-order derivatives, then combine estimates using the spatial gradient as confidence measure. Experimental results with images for which the ground-truth optical flow is known show clearly that the estimate improves by including estimates from textural images. Experiments with several underwater images also show a qualitative improvement. © 2003 Elsevier B.V. All rights reserved.

Original languageEnglish
Pages (from-to)449-457
Number of pages9
JournalPattern Recognition Letters
Volume25
Issue number4
DOIs
Publication statusPublished - Mar 2004

Fingerprint

Optical flows
Textures
Motion estimation
Derivatives
Experiments

Keywords

  • Assumption
  • Brightness constancy
  • Differential approach
  • Optical flow
  • Texture
  • Underwater images

Cite this

Arredondo, M. A. ; Lebart, K. ; Lane, D. / Optical flow using textures. In: Pattern Recognition Letters. 2004 ; Vol. 25, No. 4. pp. 449-457.
@article{f4e0b3a31b8a4133b9b583936a99a273,
title = "Optical flow using textures",
abstract = "Motion estimation is a key problem in the analysis of image sequences. From a sequence of images we can only estimate an approximation of the image motion field called optical flow. We propose to improve optical flow estimation by including information from images of textural features. We compute the optical flow from intensity and textural images from first-order derivatives, then combine estimates using the spatial gradient as confidence measure. Experimental results with images for which the ground-truth optical flow is known show clearly that the estimate improves by including estimates from textural images. Experiments with several underwater images also show a qualitative improvement. {\circledC} 2003 Elsevier B.V. All rights reserved.",
keywords = "Assumption, Brightness constancy, Differential approach, Optical flow, Texture, Underwater images",
author = "Arredondo, {M. A.} and K. Lebart and D. Lane",
year = "2004",
month = "3",
doi = "10.1016/j.patrec.2003.11.007",
language = "English",
volume = "25",
pages = "449--457",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier",
number = "4",

}

Optical flow using textures. / Arredondo, M. A.; Lebart, K.; Lane, D.

In: Pattern Recognition Letters, Vol. 25, No. 4, 03.2004, p. 449-457.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Optical flow using textures

AU - Arredondo, M. A.

AU - Lebart, K.

AU - Lane, D.

PY - 2004/3

Y1 - 2004/3

N2 - Motion estimation is a key problem in the analysis of image sequences. From a sequence of images we can only estimate an approximation of the image motion field called optical flow. We propose to improve optical flow estimation by including information from images of textural features. We compute the optical flow from intensity and textural images from first-order derivatives, then combine estimates using the spatial gradient as confidence measure. Experimental results with images for which the ground-truth optical flow is known show clearly that the estimate improves by including estimates from textural images. Experiments with several underwater images also show a qualitative improvement. © 2003 Elsevier B.V. All rights reserved.

AB - Motion estimation is a key problem in the analysis of image sequences. From a sequence of images we can only estimate an approximation of the image motion field called optical flow. We propose to improve optical flow estimation by including information from images of textural features. We compute the optical flow from intensity and textural images from first-order derivatives, then combine estimates using the spatial gradient as confidence measure. Experimental results with images for which the ground-truth optical flow is known show clearly that the estimate improves by including estimates from textural images. Experiments with several underwater images also show a qualitative improvement. © 2003 Elsevier B.V. All rights reserved.

KW - Assumption

KW - Brightness constancy

KW - Differential approach

KW - Optical flow

KW - Texture

KW - Underwater images

UR - http://www.scopus.com/inward/record.url?scp=0442280826&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2003.11.007

DO - 10.1016/j.patrec.2003.11.007

M3 - Article

VL - 25

SP - 449

EP - 457

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

IS - 4

ER -