High Accuracy Fundamental Matrix Computation and Its Performance Evaluation

Kenichi KANATANI; Yasuyuki SUGAYA

doi:10.1093/ietisy/e90-d.2.579

High Accuracy Fundamental Matrix Computation and Its Performance Evaluation

Kenichi KANATANI, Yasuyuki SUGAYA

Full Text Views

0

Share
Cite this

Summary :

We compare the convergence performance of different numerical schemes for computing the fundamental matrix from point correspondences over two images. First, we state the problem and the associated KCR lower bound. Then, we describe the algorithms of three well-known methods: FNS, HEIV, and renormalization. We also introduce Gauss-Newton iterations as a new method for fundamental matrix computation. For initial values, we test random choice, least squares, and Taubin's method. Experiments using simulated and real images reveal different characteristics of each method. Overall, FNS exhibits the best convergence properties.

Publication: IEICE TRANSACTIONS on Information Vol.E90-D No.2 pp.579-585

Publication Date: 2007/02/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1093/ietisy/e90-d.2.579

Type of Manuscript: PAPER

Category: Image Recognition, Computer Vision

Cite this

Copy

Kenichi KANATANI, Yasuyuki SUGAYA, "High Accuracy Fundamental Matrix Computation and Its Performance Evaluation" in IEICE TRANSACTIONS on Information, vol. E90-D, no. 2, pp. 579-585, February 2007, doi: 10.1093/ietisy/e90-d.2.579.
Abstract: We compare the convergence performance of different numerical schemes for computing the fundamental matrix from point correspondences over two images. First, we state the problem and the associated KCR lower bound. Then, we describe the algorithms of three well-known methods: FNS, HEIV, and renormalization. We also introduce Gauss-Newton iterations as a new method for fundamental matrix computation. For initial values, we test random choice, least squares, and Taubin's method. Experiments using simulated and real images reveal different characteristics of each method. Overall, FNS exhibits the best convergence properties.
URL: https://globals.ieice.org/en_transactions/information/10.1093/ietisy/e90-d.2.579/_p

Copy

@ARTICLE{e90-d_2_579,
author={Kenichi KANATANI, Yasuyuki SUGAYA, },
journal={IEICE TRANSACTIONS on Information},
title={High Accuracy Fundamental Matrix Computation and Its Performance Evaluation},
year={2007},
volume={E90-D},
number={2},
pages={579-585},
abstract={We compare the convergence performance of different numerical schemes for computing the fundamental matrix from point correspondences over two images. First, we state the problem and the associated KCR lower bound. Then, we describe the algorithms of three well-known methods: FNS, HEIV, and renormalization. We also introduce Gauss-Newton iterations as a new method for fundamental matrix computation. For initial values, we test random choice, least squares, and Taubin's method. Experiments using simulated and real images reveal different characteristics of each method. Overall, FNS exhibits the best convergence properties.},
keywords={},
doi={10.1093/ietisy/e90-d.2.579},
ISSN={1745-1361},
month={February},}

Copy

TY - JOUR
TI - High Accuracy Fundamental Matrix Computation and Its Performance Evaluation
T2 - IEICE TRANSACTIONS on Information
SP - 579
EP - 585
AU - Kenichi KANATANI
AU - Yasuyuki SUGAYA
PY - 2007
DO - 10.1093/ietisy/e90-d.2.579
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E90-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2007
AB - We compare the convergence performance of different numerical schemes for computing the fundamental matrix from point correspondences over two images. First, we state the problem and the associated KCR lower bound. Then, we describe the algorithms of three well-known methods: FNS, HEIV, and renormalization. We also introduce Gauss-Newton iterations as a new method for fundamental matrix computation. For initial values, we test random choice, least squares, and Taubin's method. Experiments using simulated and real images reveal different characteristics of each method. Overall, FNS exhibits the best convergence properties.
ER -