Efficient MRF Deformation Model for Image Matching
Shekhovtsov, Alexander, Kovtun, Ivan, Hlaváč, Václav
CTU-CMP-2006-08
October, 2006
Abstract
We propose a novel MRF-based model for image matching. Given two
images, the task is to estimate a mapping from one image to another,
in order to maximize the matching quality. We consider mappings
defined by discrete deformation field constrained to preserve
2-dimensional continuity. We approach the corresponding optimization
problem by the TRW-S (sequential Tree-reweighted message passing)
algorithm [Wainwright-03, Kolmogorov-05]. Our model design allows for
a considerably wider class of natural transformation and yields a
compact representation of the optimization task. For this model TRW-S
algorithm demonstrated nice practical performance on our
experiments. We also propose a concise derivation of the TRW-S
algorithm as a sequential maximization of the lower bound on the
energy function.
Keywords
optical flow, registration, Energy minimization, MRF, message passing, early vision, graphical models
Full Paper
Portable document format file PDF
Gzipped postscript file ps.gz ()
Bibtex entry
@TechReport{Shekhovtsov-TR-2006-08,
author = {Shekhovtsov, Alexander and Kovtun, Ivan and
Hlav{\'a}{\vc}, V{\'a}clav},
title = {Efficient {MRF} Deformation Model for Image Matching},
institution = {Center for Machine Perception, K13133 FEE Czech Technical
University},
address = {Prague, Czech Republic},
year = {2006},
month = {October},
type = {Research Report},
number = {CTU--CMP--2006--08},
issn = {1213-2365},
pages = {12},
figures = {4},
psurl = {[Shekhovtsov-TR-2006-08.pdf]},
annote = {We propose a novel MRF-based model for image
matching. Given two images, the task is to estimate a mapping from
one image to another, in order to maximize the matching
quality. We consider mappings defined by discrete deformation
field constrained to preserve 2-dimensional continuity. We
approach the corresponding optimization problem by the TRW-S
(sequential Tree-reweighted message passing) algorithm
[Wainwright-03, Kolmogorov-05]. Our model design allows for a
considerably wider class of natural transformation and yields a
compact representation of the optimization task. For this model
TRW-S algorithm demonstrated nice practical performance on our
experiments. We also propose a concise derivation of the TRW-S
algorithm as a sequential maximization of the lower bound on the
energy function.},
keywords = {optical flow, registration, Energy minimization, MRF,
message passing, early vision, graphical models},
}