HALF-SIFT

High-Accurate Localized Features for SIFT

verfasst von: Kai Cordes, Oliver Müller, Bodo Rosenhahn, Jörn Ostermann
Abstract: In this paper, the accuracy of feature points in images detected by the Scale Invariant Feature Transform (SIFT) is analyzed. It is shown that there is a systematic error in the feature point localization. The systematic error is caused by the improper subpel and subscale estimation, an interpolation with a parabolic function. To avoid the systematic error, the detection of High-Accurate Localized Features (HALF) is proposed. We present two models for the localization of a feature point in the scale-space, a Gaussian and a Difference of Gaussians based model function. For evaluation, ground truth image data is synthesized to experimentally prove the systematic error of SIFT and to show that the error is eliminated using HALF. Experiments with natural image data show that the proposed methods increase the accuracy of the feature point positions by 13.9% using the Gaussian and by 15.6% using the Difference of Gaussians model.
Organisationseinheit(en): Institut für Informationsverarbeitung
Typ: Aufsatz in Konferenzband
Seiten: 31-38
Anzahl der Seiten: 8
Publikationsdatum: 08.2009
Publikationsstatus: Veröffentlicht
Peer-reviewed: Ja
ASJC Scopus Sachgebiete: Maschinelles Sehen und Mustererkennung, Biomedizintechnik
Elektronische Version(en): https://doi.org/10.1109/CVPR.2009.5204283 (Zugang: Unbekannt)

BibTeX

@inproceedings{7252ee19c0514b7789ce35fb96d04b65,
title = "HALF-SIFT: High-Accurate Localized Features for SIFT",
abstract = "In this paper, the accuracy of feature points in images detected by the Scale Invariant Feature Transform (SIFT) is analyzed. It is shown that there is a systematic error in the feature point localization. The systematic error is caused by the improper subpel and subscale estimation, an interpolation with a parabolic function. To avoid the systematic error, the detection of High-Accurate Localized Features (HALF) is proposed. We present two models for the localization of a feature point in the scale-space, a Gaussian and a Difference of Gaussians based model function. For evaluation, ground truth image data is synthesized to experimentally prove the systematic error of SIFT and to show that the error is eliminated using HALF. Experiments with natural image data show that the proposed methods increase the accuracy of the feature point positions by 13.9% using the Gaussian and by 15.6% using the Difference of Gaussians model.",
author = "Kai Cordes and Oliver M{\"u}ller and Bodo Rosenhahn and J{\"o}rn Ostermann",
year = "2009",
month = aug,
doi = "10.1109/CVPR.2009.5204283",
language = "English",
isbn = "9781424439911",
series = "2009 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2009",
publisher = "IEEE",
pages = "31--38",
booktitle = "2009 IEEE Conference on Computer Vision and Pattern Recognition",
note = "2009 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2009, CVPR 2009 ; Conference date: 20-06-2009 Through 25-06-2009",
}