Estimation of Motion Vector Fields

Rasmus Larsen

Estimation of Motion Vector Fields

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

418 Downloads (Pure)

Abstract

This paper presents an approach to the estimation of 2-D motion vector fields from time varying image sequences. We use a piecewise smooth model based on coupled vector/binary Markov random fields. We find the maximum a posteriori solution by simulated annealing. The algorithm generate sample fields by means of stochastic relaxation implemented via the Gibbs sampler.

Original language	English
Title of host publication	Proceedings of the 2nd Danish Conference on Pattern Recognition and Image Analysis (DANKOMB, DSAGM yearly meeting), Aug. 25-26
Editors	Peter Johansen
Publisher	Department of Computer Science, University of Copenhagen (DIKU)
Publication date	1993
Pages	37-42
Publication status	Published - 1993
Event	2nd Danish Conference on Pattern Recognition and Image Analysis - , Denmark Duration: 25 Aug 2009 → 26 Aug 2009

Conference

Conference	2nd Danish Conference on Pattern Recognition and Image Analysis
Country/Territory	Denmark
Period	25/08/2009 → 26/08/2009

Keywords

Markov random fields
optical flow

Access to Document

Imm1111Final published version, 79.5 KB

OpenUrl availability

Full text

Cite this

@inproceedings{c5b07c80a3324f5caf907417808c0a82,

title = "Estimation of Motion Vector Fields",

abstract = "This paper presents an approach to the estimation of 2-D motion vector fields from time varying image sequences. We use a piecewise smooth model based on coupled vector/binary Markov random fields. We find the maximum a posteriori solution by simulated annealing. The algorithm generate sample fields by means of stochastic relaxation implemented via the Gibbs sampler.",

keywords = "Markov random fields, optical flow",

author = "Rasmus Larsen",

year = "1993",

language = "English",

pages = "37--42",

editor = "Peter Johansen",

booktitle = "Proceedings of the 2nd Danish Conference on Pattern Recognition and Image Analysis (DANKOMB, DSAGM yearly meeting), Aug. 25-26",

publisher = "Department of Computer Science, University of Copenhagen (DIKU)",

note = "2nd Danish Conference on Pattern Recognition and Image Analysis ; Conference date: 25-08-2009 Through 26-08-2009",

}

Estimation of Motion Vector Fields. / Larsen, Rasmus.
Proceedings of the 2nd Danish Conference on Pattern Recognition and Image Analysis (DANKOMB, DSAGM yearly meeting), Aug. 25-26. ed. / Peter Johansen. Department of Computer Science, University of Copenhagen (DIKU), 1993. p. 37-42.

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

TY - GEN

T1 - Estimation of Motion Vector Fields

AU - Larsen, Rasmus

PY - 1993

Y1 - 1993

N2 - This paper presents an approach to the estimation of 2-D motion vector fields from time varying image sequences. We use a piecewise smooth model based on coupled vector/binary Markov random fields. We find the maximum a posteriori solution by simulated annealing. The algorithm generate sample fields by means of stochastic relaxation implemented via the Gibbs sampler.

AB - This paper presents an approach to the estimation of 2-D motion vector fields from time varying image sequences. We use a piecewise smooth model based on coupled vector/binary Markov random fields. We find the maximum a posteriori solution by simulated annealing. The algorithm generate sample fields by means of stochastic relaxation implemented via the Gibbs sampler.

KW - Markov random fields

KW - optical flow

M3 - Article in proceedings

SP - 37

EP - 42

BT - Proceedings of the 2nd Danish Conference on Pattern Recognition and Image Analysis (DANKOMB, DSAGM yearly meeting), Aug. 25-26

A2 - Johansen, Peter

PB - Department of Computer Science, University of Copenhagen (DIKU)

T2 - 2nd Danish Conference on Pattern Recognition and Image Analysis

Y2 - 25 August 2009 through 26 August 2009

ER -

Estimation of Motion Vector Fields

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this