Improved SIFT-based image registration using belief propagation

S. Cheng; Vladimir M. Stankovic; L. Stankovic

doi:10.1109/ICASSP.2009.4960232

Improved SIFT-based image registration using belief propagation

S. Cheng, Vladimir M. Stankovic, L. Stankovic

Electronic And Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution book

12 Citations (Scopus)

Abstract

Scale Invariant Feature Transform (SIFT) is a very powerful technique for image registration. While SIFT descriptors accurately extract invariant image characteristics around keypoints, the commonly used matching approach for registration is overly simplified, because it completely ignores the geometric information among descriptors. In this paper, we formulate keypoint matching as a global optimization problem and provide a suboptimum solution using belief propagation. Experimental results show significant improvement over previous approaches.

Original language	English
Title of host publication	IEEE International Conference on Acoustics, Speech and Signal Processing, 2009
Publisher	IEEE
Pages	2909-2912
Number of pages	3
ISBN (Print)	978-1-4244-2353-8
DOIs	https://doi.org/10.1109/ICASSP.2009.4960232
Publication status	Published - 30 Apr 2009

Fingerprint

Keywords

belief networks
image matching
image registration
optimisation
transforms
belief propagation
descriptors
invariant image characteristics
scale invariant feature transform

Cite this

Cheng, S., Stankovic, V. M., & Stankovic, L. (2009). Improved SIFT-based image registration using belief propagation. In IEEE International Conference on Acoustics, Speech and Signal Processing, 2009 (pp. 2909-2912). IEEE. https://doi.org/10.1109/ICASSP.2009.4960232

@inproceedings{c85b9a1f59e440c0b69fdc3e40090425,

title = "Improved SIFT-based image registration using belief propagation",

abstract = "Scale Invariant Feature Transform (SIFT) is a very powerful technique for image registration. While SIFT descriptors accurately extract invariant image characteristics around keypoints, the commonly used matching approach for registration is overly simplified, because it completely ignores the geometric information among descriptors. In this paper, we formulate keypoint matching as a global optimization problem and provide a suboptimum solution using belief propagation. Experimental results show significant improvement over previous approaches.",

keywords = "belief networks, image matching, image registration, optimisation, transforms, belief propagation, descriptors, invariant image characteristics, scale invariant feature transform",

author = "S. Cheng and Stankovic, {Vladimir M.} and L. Stankovic",

year = "2009",

month = "4",

day = "30",

doi = "10.1109/ICASSP.2009.4960232",

language = "English",

isbn = "978-1-4244-2353-8",

pages = "2909--2912",

booktitle = "IEEE International Conference on Acoustics, Speech and Signal Processing, 2009",

publisher = "IEEE",

}

TY - GEN

T1 - Improved SIFT-based image registration using belief propagation

AU - Cheng, S.

AU - Stankovic, Vladimir M.

AU - Stankovic, L.

PY - 2009/4/30

Y1 - 2009/4/30

N2 - Scale Invariant Feature Transform (SIFT) is a very powerful technique for image registration. While SIFT descriptors accurately extract invariant image characteristics around keypoints, the commonly used matching approach for registration is overly simplified, because it completely ignores the geometric information among descriptors. In this paper, we formulate keypoint matching as a global optimization problem and provide a suboptimum solution using belief propagation. Experimental results show significant improvement over previous approaches.

AB - Scale Invariant Feature Transform (SIFT) is a very powerful technique for image registration. While SIFT descriptors accurately extract invariant image characteristics around keypoints, the commonly used matching approach for registration is overly simplified, because it completely ignores the geometric information among descriptors. In this paper, we formulate keypoint matching as a global optimization problem and provide a suboptimum solution using belief propagation. Experimental results show significant improvement over previous approaches.

KW - belief networks

KW - image matching

KW - image registration

KW - optimisation

KW - transforms

KW - belief propagation

KW - descriptors

KW - invariant image characteristics

KW - scale invariant feature transform

UR - http://dx.doi.org/10.1109/ICASSP.2009.4960232

U2 - 10.1109/ICASSP.2009.4960232

DO - 10.1109/ICASSP.2009.4960232

M3 - Conference contribution book

SN - 978-1-4244-2353-8

SP - 2909

EP - 2912

BT - IEEE International Conference on Acoustics, Speech and Signal Processing, 2009

PB - IEEE

ER -

Access to Document

10.1109/ICASSP.2009.4960232

http://dx.doi.org/10.1109/ICASSP.2009.4960232