Publication Details

Wavelet Transform Based Feature Detection

BAŘINA, D. Wavelet Transform Based Feature Detection. Proceedings of the 16th Conference STUDENT EEICT 2010. Volume 5. Brno: Brno University of Technology, 2010. p. 208-212. ISBN: 978-80-214-4080-7.
Czech title
Detekce příznaků založená na vlnkové transformaci
Type
conference paper
Language
English
Authors
Bařina David, Ing., Ph.D. (DCGM)
URL
Keywords

wavelet transform, feature detection

Abstract

The subject of this contribution is to demonstrate the procedure of interest point detection using partial derivatives obtained in some level of stationary wavelet decomposition.

Annotation
This paper presents a new method of the local image feature detection. The method is based on a corner detection on top of a stationary wavelet transform. The point of this method is that the corner detector is applied on partial derivatives obtained from the wavelet transform. Thus, edges are obtained at different scales. A position of the detected corner can be estimated with sub-pixel accuracy. The proposed method can be used as part of a robust feature extractor.
Published
2010
Pages
208–212
Proceedings
Proceedings of the 16th Conference STUDENT EEICT 2010
Series
Volume 5
Conference
Student EEICT 2010, FEKT VUT v Brně, CZ
ISBN
978-80-214-4080-7
Publisher
Brno University of Technology
Place
Brno
BibTeX 
@inproceedings{BUT34735,
  author="David {Bařina}",
  title="Wavelet Transform Based Feature Detection",
  booktitle="Proceedings of the 16th Conference STUDENT EEICT 2010",
  year="2010",
  series="Volume 5",
  pages="208--212",
  publisher="Brno University of Technology",
  address="Brno",
  isbn="978-80-214-4080-7",
  url="https://www.fit.vut.cz/research/publication/9213/"
}
Back to top