Efficient keypoint detection and description using filter kernel decomposition in scale space

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Keypoint detection and description in a continuous scale space achieves better robustness to scale change than those in a discretized scale space. State-of-the-art methods first decompose a continuous scale space into M + 1 component images weighted by M-order polynomials of scale σ, and then reconstruct the value at an arbitrary point in the scale space by a linear combination of the component images. However, these methods create the mismatch that, if σ is large, common filter kernels such as Gaussian converge to zero; but the polynomials of σ diverge. This paper presents a more efficient approximation that suppresses this mismatch by normalizing the weighting functions. Experiments show that the proposed method achieves higher performance tradeoff than state-of-the-art methods: it reduces the number of component images and total running time by 20-40% without a sacrifice of stability in keypoints detection.

Original languageEnglish
Title of host publication2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings
PublisherIEEE Computer Society
Pages31-35
Number of pages5
Volume2016-August
ISBN (Electronic)9781467399616
DOIs
Publication statusPublished - 2016 Aug 3
Event23rd IEEE International Conference on Image Processing, ICIP 2016 - Phoenix, United States
Duration: 2016 Sep 252016 Sep 28

Other

Other23rd IEEE International Conference on Image Processing, ICIP 2016
CountryUnited States
CityPhoenix
Period16/9/2516/9/28

Fingerprint

Polynomials
Decomposition
Experiments

Keywords

  • Feature description
  • Keypoint detection
  • Scale space
  • SIFT
  • Spectral SIFT

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Okutani, R., Sugimoto, K., & Kamata, S. (2016). Efficient keypoint detection and description using filter kernel decomposition in scale space. In 2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings (Vol. 2016-August, pp. 31-35). [7532313] IEEE Computer Society. https://doi.org/10.1109/ICIP.2016.7532313

Efficient keypoint detection and description using filter kernel decomposition in scale space. / Okutani, Ryo; Sugimoto, Kenjiro; Kamata, Seiichiro.

2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings. Vol. 2016-August IEEE Computer Society, 2016. p. 31-35 7532313.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Okutani, R, Sugimoto, K & Kamata, S 2016, Efficient keypoint detection and description using filter kernel decomposition in scale space. in 2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings. vol. 2016-August, 7532313, IEEE Computer Society, pp. 31-35, 23rd IEEE International Conference on Image Processing, ICIP 2016, Phoenix, United States, 16/9/25. https://doi.org/10.1109/ICIP.2016.7532313
Okutani R, Sugimoto K, Kamata S. Efficient keypoint detection and description using filter kernel decomposition in scale space. In 2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings. Vol. 2016-August. IEEE Computer Society. 2016. p. 31-35. 7532313 https://doi.org/10.1109/ICIP.2016.7532313
Okutani, Ryo ; Sugimoto, Kenjiro ; Kamata, Seiichiro. / Efficient keypoint detection and description using filter kernel decomposition in scale space. 2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings. Vol. 2016-August IEEE Computer Society, 2016. pp. 31-35
@inproceedings{45325fd42c98411da9ee4e4f36c232b0,
title = "Efficient keypoint detection and description using filter kernel decomposition in scale space",
abstract = "Keypoint detection and description in a continuous scale space achieves better robustness to scale change than those in a discretized scale space. State-of-the-art methods first decompose a continuous scale space into M + 1 component images weighted by M-order polynomials of scale σ, and then reconstruct the value at an arbitrary point in the scale space by a linear combination of the component images. However, these methods create the mismatch that, if σ is large, common filter kernels such as Gaussian converge to zero; but the polynomials of σ diverge. This paper presents a more efficient approximation that suppresses this mismatch by normalizing the weighting functions. Experiments show that the proposed method achieves higher performance tradeoff than state-of-the-art methods: it reduces the number of component images and total running time by 20-40{\%} without a sacrifice of stability in keypoints detection.",
keywords = "Feature description, Keypoint detection, Scale space, SIFT, Spectral SIFT",
author = "Ryo Okutani and Kenjiro Sugimoto and Seiichiro Kamata",
year = "2016",
month = "8",
day = "3",
doi = "10.1109/ICIP.2016.7532313",
language = "English",
volume = "2016-August",
pages = "31--35",
booktitle = "2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings",
publisher = "IEEE Computer Society",
address = "United States",

}

TY - GEN

T1 - Efficient keypoint detection and description using filter kernel decomposition in scale space

AU - Okutani, Ryo

AU - Sugimoto, Kenjiro

AU - Kamata, Seiichiro

PY - 2016/8/3

Y1 - 2016/8/3

N2 - Keypoint detection and description in a continuous scale space achieves better robustness to scale change than those in a discretized scale space. State-of-the-art methods first decompose a continuous scale space into M + 1 component images weighted by M-order polynomials of scale σ, and then reconstruct the value at an arbitrary point in the scale space by a linear combination of the component images. However, these methods create the mismatch that, if σ is large, common filter kernels such as Gaussian converge to zero; but the polynomials of σ diverge. This paper presents a more efficient approximation that suppresses this mismatch by normalizing the weighting functions. Experiments show that the proposed method achieves higher performance tradeoff than state-of-the-art methods: it reduces the number of component images and total running time by 20-40% without a sacrifice of stability in keypoints detection.

AB - Keypoint detection and description in a continuous scale space achieves better robustness to scale change than those in a discretized scale space. State-of-the-art methods first decompose a continuous scale space into M + 1 component images weighted by M-order polynomials of scale σ, and then reconstruct the value at an arbitrary point in the scale space by a linear combination of the component images. However, these methods create the mismatch that, if σ is large, common filter kernels such as Gaussian converge to zero; but the polynomials of σ diverge. This paper presents a more efficient approximation that suppresses this mismatch by normalizing the weighting functions. Experiments show that the proposed method achieves higher performance tradeoff than state-of-the-art methods: it reduces the number of component images and total running time by 20-40% without a sacrifice of stability in keypoints detection.

KW - Feature description

KW - Keypoint detection

KW - Scale space

KW - SIFT

KW - Spectral SIFT

UR - http://www.scopus.com/inward/record.url?scp=85006826723&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006826723&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2016.7532313

DO - 10.1109/ICIP.2016.7532313

M3 - Conference contribution

AN - SCOPUS:85006826723

VL - 2016-August

SP - 31

EP - 35

BT - 2016 IEEE International Conference on Image Processing, ICIP 2016 - Proceedings

PB - IEEE Computer Society

ER -