Publication detail

Information Content of Blobs in Images

ZUKAL, M. ČÍKA, P. FU, D.

Original Title

Information Content of Blobs in Images

English Title

Information Content of Blobs in Images

Type

conference paper

Language

en

Original Abstract

This paper deals with the comparison of information content of interest points in images detected with different blob detectors. Specifically, the Hessian, Hessian-Laplace, Fast Hessian and Difference of Gaussian detectors are compared. We have performed the measurement on five image collections each containing images of one category. The tested categories are namely: houses, cars, nature, portraits and interior images. Each collection contains one hundred greyscale images. We have compared the information content of the interest points and randomly selected points and observed that the interest points convey more information than the randomly distributed points. The Fast Hessian detector and Difference of Gaussians detector perform best and entropy of points detected by these algorithms reaches 5 Shannons.

English abstract

This paper deals with the comparison of information content of interest points in images detected with different blob detectors. Specifically, the Hessian, Hessian-Laplace, Fast Hessian and Difference of Gaussian detectors are compared. We have performed the measurement on five image collections each containing images of one category. The tested categories are namely: houses, cars, nature, portraits and interior images. Each collection contains one hundred greyscale images. We have compared the information content of the interest points and randomly selected points and observed that the interest points convey more information than the randomly distributed points. The Fast Hessian detector and Difference of Gaussians detector perform best and entropy of points detected by these algorithms reaches 5 Shannons.

Keywords

blob detector, Difference of Gaussians, Fast Hessian, information content, interest point detector

RIV year

2012

Released

12.09.2012

ISBN

978-80-554-0569-8

Book

The 14th International Conference on Research in Telecommunication Technologies RTT - 2012

Edition number

1

Pages from

221

Pages to

224

Pages count

4

BibTex


@inproceedings{BUT94627,
  author="Martin {Zukal} and Petr {Číka} and Dongmei {Fu}",
  title="Information Content of Blobs in Images",
  annote="This paper deals with the comparison of information content of interest points in images detected with different blob detectors. Specifically, the Hessian, Hessian-Laplace, Fast Hessian and Difference of Gaussian detectors are compared. We have performed the measurement on five image collections each containing images of one category. The tested categories are namely: houses, cars, nature, portraits and interior images. Each collection contains one hundred greyscale images. We have compared the information content of the interest points and randomly selected points and observed that the interest points convey more information than the randomly distributed points. The Fast Hessian detector and Difference of Gaussians detector perform best and entropy of points detected by these algorithms reaches 5 Shannons.",
  booktitle="The 14th International Conference on Research in Telecommunication Technologies RTT - 2012",
  chapter="94627",
  howpublished="electronic, physical medium",
  year="2012",
  month="september",
  pages="221--224",
  type="conference paper"
}