Variational method for super-resolution optical flow

Yoshihiko Mochizuki, Yusuke Kameda, Atsushi Imiya, Tomoya Sakai, Takashi Imaizumi

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

The motion fields in an image sequence observed by a car-mounted imaging system depend on the positions in the imaging plane. Since the motion displacements in the regions close to the camera centre are small, for accurate optical flow computation in this region, we are required to use super-resolution of optical flow fields. We develop an algorithm for super-resolution optical flow computation. Super-resolution of images is a technique for recovering a high-resolution image from a low-resolution image and/or image sequence. Optical flow is the appearance motion of points on the image. Therefore, super-resolution optical flow computation yields the appearance motion of each point on the high-resolution image from a sequence of low-resolution images. We combine variational super-resolution and variational optical flow computation in super-resolution optical flow computation. Our method directly computes the gradient and spatial difference of high-resolution images from those of low-resolution images, without computing any high-resolution images used as intermediate data for the computation of optical flow vectors of the high-resolution image.

Original languageEnglish
Pages (from-to)1535-1567
Number of pages33
JournalSignal Processing
Volume91
Issue number7
DOIs
Publication statusPublished - 2011 Jul
Externally publishedYes

Fingerprint

Optical flows
Image resolution
Optical resolving power
Imaging systems
Flow fields
Railroad cars
Cameras
Imaging techniques

Keywords

  • Gauss pyramid
  • Numerical analysis
  • Optical flow
  • Pyramid transform
  • Small motion displacement
  • Super-resolution
  • Variational method
  • Zooming

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Control and Systems Engineering
  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition

Cite this

Variational method for super-resolution optical flow. / Mochizuki, Yoshihiko; Kameda, Yusuke; Imiya, Atsushi; Sakai, Tomoya; Imaizumi, Takashi.

In: Signal Processing, Vol. 91, No. 7, 07.2011, p. 1535-1567.

Research output: Contribution to journalArticle

Mochizuki, Y, Kameda, Y, Imiya, A, Sakai, T & Imaizumi, T 2011, 'Variational method for super-resolution optical flow', Signal Processing, vol. 91, no. 7, pp. 1535-1567. https://doi.org/10.1016/j.sigpro.2010.11.010
Mochizuki, Yoshihiko ; Kameda, Yusuke ; Imiya, Atsushi ; Sakai, Tomoya ; Imaizumi, Takashi. / Variational method for super-resolution optical flow. In: Signal Processing. 2011 ; Vol. 91, No. 7. pp. 1535-1567.
@article{8713c40e73c54cff8bc460ea65e33b84,
title = "Variational method for super-resolution optical flow",
abstract = "The motion fields in an image sequence observed by a car-mounted imaging system depend on the positions in the imaging plane. Since the motion displacements in the regions close to the camera centre are small, for accurate optical flow computation in this region, we are required to use super-resolution of optical flow fields. We develop an algorithm for super-resolution optical flow computation. Super-resolution of images is a technique for recovering a high-resolution image from a low-resolution image and/or image sequence. Optical flow is the appearance motion of points on the image. Therefore, super-resolution optical flow computation yields the appearance motion of each point on the high-resolution image from a sequence of low-resolution images. We combine variational super-resolution and variational optical flow computation in super-resolution optical flow computation. Our method directly computes the gradient and spatial difference of high-resolution images from those of low-resolution images, without computing any high-resolution images used as intermediate data for the computation of optical flow vectors of the high-resolution image.",
keywords = "Gauss pyramid, Numerical analysis, Optical flow, Pyramid transform, Small motion displacement, Super-resolution, Variational method, Zooming",
author = "Yoshihiko Mochizuki and Yusuke Kameda and Atsushi Imiya and Tomoya Sakai and Takashi Imaizumi",
year = "2011",
month = "7",
doi = "10.1016/j.sigpro.2010.11.010",
language = "English",
volume = "91",
pages = "1535--1567",
journal = "Signal Processing",
issn = "0165-1684",
publisher = "Elsevier",
number = "7",

}

TY - JOUR

T1 - Variational method for super-resolution optical flow

AU - Mochizuki, Yoshihiko

AU - Kameda, Yusuke

AU - Imiya, Atsushi

AU - Sakai, Tomoya

AU - Imaizumi, Takashi

PY - 2011/7

Y1 - 2011/7

N2 - The motion fields in an image sequence observed by a car-mounted imaging system depend on the positions in the imaging plane. Since the motion displacements in the regions close to the camera centre are small, for accurate optical flow computation in this region, we are required to use super-resolution of optical flow fields. We develop an algorithm for super-resolution optical flow computation. Super-resolution of images is a technique for recovering a high-resolution image from a low-resolution image and/or image sequence. Optical flow is the appearance motion of points on the image. Therefore, super-resolution optical flow computation yields the appearance motion of each point on the high-resolution image from a sequence of low-resolution images. We combine variational super-resolution and variational optical flow computation in super-resolution optical flow computation. Our method directly computes the gradient and spatial difference of high-resolution images from those of low-resolution images, without computing any high-resolution images used as intermediate data for the computation of optical flow vectors of the high-resolution image.

AB - The motion fields in an image sequence observed by a car-mounted imaging system depend on the positions in the imaging plane. Since the motion displacements in the regions close to the camera centre are small, for accurate optical flow computation in this region, we are required to use super-resolution of optical flow fields. We develop an algorithm for super-resolution optical flow computation. Super-resolution of images is a technique for recovering a high-resolution image from a low-resolution image and/or image sequence. Optical flow is the appearance motion of points on the image. Therefore, super-resolution optical flow computation yields the appearance motion of each point on the high-resolution image from a sequence of low-resolution images. We combine variational super-resolution and variational optical flow computation in super-resolution optical flow computation. Our method directly computes the gradient and spatial difference of high-resolution images from those of low-resolution images, without computing any high-resolution images used as intermediate data for the computation of optical flow vectors of the high-resolution image.

KW - Gauss pyramid

KW - Numerical analysis

KW - Optical flow

KW - Pyramid transform

KW - Small motion displacement

KW - Super-resolution

KW - Variational method

KW - Zooming

UR - http://www.scopus.com/inward/record.url?scp=79952040109&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79952040109&partnerID=8YFLogxK

U2 - 10.1016/j.sigpro.2010.11.010

DO - 10.1016/j.sigpro.2010.11.010

M3 - Article

AN - SCOPUS:79952040109

VL - 91

SP - 1535

EP - 1567

JO - Signal Processing

JF - Signal Processing

SN - 0165-1684

IS - 7

ER -