IMPROVED SIFT-BASED IMAGE REGISTRATION USING BELIEF PROPAGATION

被引：10

作者：

Cheng, Samuel ^{[1
]}

Stankovic, Vladimir ^{[2
]}

Stankovic, Lina ^{[2
]}

机构：

[1] Univ Oklahoma, Dept Elect & Comp Engn, Tulsa, OK 74135 USA

[2] Univ Strathclyde, Dept Elect & Elect Engn, Glasgow G1 1XW, Lanark, Scotland

来源：

2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS | 2009年

关键词：

Image registration; belief propagation; SIFT;

D O I：

10.1109/ICASSP.2009.4960232

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

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.

引用

页码：2909 / +

页数：2

共 13 条

[1]

[Anonymous], 2007, 070012 UCLA CSD

[2]

Grabner M, 2006, LECT NOTES COMPUT SC, V3851, P918

[3]

HOPFIELD JJ, 2001, P NATL ACAD SCI USA, P67098

[4]

KE Y, 2004, P CVPR 2004, V2

[5]

KE Y, 1999, P CVPR 2004, V2

[6] Factor graphs and the sum-product algorithm [J].

Kschischang, FR ;

Frey, BJ ;

Loeliger, HA .

IEEE TRANSACTIONS ON INFORMATION THEORY, 2001, 47 (02) :498-519

[7] Distinctive image features from scale-invariant keypoints [J].

Lowe, DG .

INTERNATIONAL JOURNAL OF COMPUTER VISION, 2004, 60 (02) :91-110

[8]

MacKay David JC, 2003, Information Theory, Inference, and Learning Algorithms

[9] Good error-correcting codes based on very sparse matrices [J].

MacKay, DJC .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1999, 45 (02) :399-431

[10]

MACKAY DJC, 1997, IEEE J SELE IN PRESS

← 1 2 →