Large displacement optical flow from nearest neighbor fields

Zhuoyuan Chen, Hailin Jin, Zhe Lin, Scott Cohen, Ying Wu

Research output: Contribution to journalConference article

97 Citations (Scopus)

Abstract

We present an optical flow algorithm for large displacement motions. Most existing optical flow methods use the standard coarse-to-fine framework to deal with large displacement motions which has intrinsic limitations. Instead, we formulate the motion estimation problem as a motion segmentation problem. We use approximate nearest neighbor fields to compute an initial motion field and use a robust algorithm to compute a set of similarity transformations as the motion candidates for segmentation. To account for deviations from similarity transformations, we add local deformations in the segmentation process. We also observe that small objects can be better recovered using translations as the motion candidates. We fuse the motion results obtained under similarity transformations and under translations together before a final refinement. Experimental validation shows that our method can successfully handle large displacement motions. Although we particularly focus on large displacement motions in this work, we make no sacrifice in terms of overall performance. In particular, our method ranks at the top of the Middlebury benchmark.

Original languageEnglish (US)
Article number6619160
Pages (from-to)2443-2450
Number of pages8
JournalProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
DOIs
StatePublished - Nov 15 2013
Event26th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2013 - Portland, OR, United States
Duration: Jun 23 2013Jun 28 2013

Fingerprint

Optical flows
Electric fuses
Motion estimation
Nearest neighbor search

Keywords

  • Motion Segmentation
  • Optical Flow
  • PatchMatch
  • Randomized Algorithm

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition

Cite this

@article{e1452ca3559c4ec3ad62f8a613df2ff8,
title = "Large displacement optical flow from nearest neighbor fields",
abstract = "We present an optical flow algorithm for large displacement motions. Most existing optical flow methods use the standard coarse-to-fine framework to deal with large displacement motions which has intrinsic limitations. Instead, we formulate the motion estimation problem as a motion segmentation problem. We use approximate nearest neighbor fields to compute an initial motion field and use a robust algorithm to compute a set of similarity transformations as the motion candidates for segmentation. To account for deviations from similarity transformations, we add local deformations in the segmentation process. We also observe that small objects can be better recovered using translations as the motion candidates. We fuse the motion results obtained under similarity transformations and under translations together before a final refinement. Experimental validation shows that our method can successfully handle large displacement motions. Although we particularly focus on large displacement motions in this work, we make no sacrifice in terms of overall performance. In particular, our method ranks at the top of the Middlebury benchmark.",
keywords = "Motion Segmentation, Optical Flow, PatchMatch, Randomized Algorithm",
author = "Zhuoyuan Chen and Hailin Jin and Zhe Lin and Scott Cohen and Ying Wu",
year = "2013",
month = "11",
day = "15",
doi = "10.1109/CVPR.2013.316",
language = "English (US)",
pages = "2443--2450",
journal = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
issn = "1063-6919",
publisher = "IEEE Computer Society",

}

Large displacement optical flow from nearest neighbor fields. / Chen, Zhuoyuan; Jin, Hailin; Lin, Zhe; Cohen, Scott; Wu, Ying.

In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 15.11.2013, p. 2443-2450.

Research output: Contribution to journalConference article

TY - JOUR

T1 - Large displacement optical flow from nearest neighbor fields

AU - Chen, Zhuoyuan

AU - Jin, Hailin

AU - Lin, Zhe

AU - Cohen, Scott

AU - Wu, Ying

PY - 2013/11/15

Y1 - 2013/11/15

N2 - We present an optical flow algorithm for large displacement motions. Most existing optical flow methods use the standard coarse-to-fine framework to deal with large displacement motions which has intrinsic limitations. Instead, we formulate the motion estimation problem as a motion segmentation problem. We use approximate nearest neighbor fields to compute an initial motion field and use a robust algorithm to compute a set of similarity transformations as the motion candidates for segmentation. To account for deviations from similarity transformations, we add local deformations in the segmentation process. We also observe that small objects can be better recovered using translations as the motion candidates. We fuse the motion results obtained under similarity transformations and under translations together before a final refinement. Experimental validation shows that our method can successfully handle large displacement motions. Although we particularly focus on large displacement motions in this work, we make no sacrifice in terms of overall performance. In particular, our method ranks at the top of the Middlebury benchmark.

AB - We present an optical flow algorithm for large displacement motions. Most existing optical flow methods use the standard coarse-to-fine framework to deal with large displacement motions which has intrinsic limitations. Instead, we formulate the motion estimation problem as a motion segmentation problem. We use approximate nearest neighbor fields to compute an initial motion field and use a robust algorithm to compute a set of similarity transformations as the motion candidates for segmentation. To account for deviations from similarity transformations, we add local deformations in the segmentation process. We also observe that small objects can be better recovered using translations as the motion candidates. We fuse the motion results obtained under similarity transformations and under translations together before a final refinement. Experimental validation shows that our method can successfully handle large displacement motions. Although we particularly focus on large displacement motions in this work, we make no sacrifice in terms of overall performance. In particular, our method ranks at the top of the Middlebury benchmark.

KW - Motion Segmentation

KW - Optical Flow

KW - PatchMatch

KW - Randomized Algorithm

UR - http://www.scopus.com/inward/record.url?scp=84887336889&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84887336889&partnerID=8YFLogxK

U2 - 10.1109/CVPR.2013.316

DO - 10.1109/CVPR.2013.316

M3 - Conference article

AN - SCOPUS:84887336889

SP - 2443

EP - 2450

JO - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

JF - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

SN - 1063-6919

M1 - 6619160

ER -