A parametric spectral model for texture-based salience

Kasim Terzić*, Sai Krishna, J. M.H. Du Buf

*Corresponding author for this work

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

Abstract

We present a novel saliency mechanism based on texture. Local texture at each pixel is characterised by the 2D spectrum obtained from oriented Gabor filters. We then apply a parametric model and describe the texture at each pixel by a combination of two 1D Gaussian approximations. This results in a simple model which consists of only four parameters. These four parameters are then used as feature channels and standard Difference-of-Gaussian blob detection is applied in order to detect salient areas in the image, similar to the Itti and Koch model. Finally, a diffusion process is used to sharpen the resulting regions. Evaluation on a large saliency dataset shows a significant improvement of our method over the baseline Itti and Koch model.

Original languageEnglish
Title of host publicationPattern Recognition
Subtitle of host publication37th German Conference, GCPR 2015, Aachen, Germany, October 7-10, 2015, Proceedings
EditorsJuergen Gall, Peter Gehler, Bastian Leibe
Place of PublicationCham
PublisherSpringer
Pages331-342
Number of pages12
ISBN (Electronic)9783319249476
ISBN (Print)9783319249469
DOIs
Publication statusPublished - 2015
Event37th German Conference on Pattern Recognition, GCPR 2015 - Aachen, Germany
Duration: 7 Oct 201510 Oct 2015
Conference number: 37
http://gcpr2015.rwth-aachen.de/

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9358
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference37th German Conference on Pattern Recognition, GCPR 2015
Abbreviated titleGCPR
Country/TerritoryGermany
CityAachen
Period7/10/1510/10/15
Internet address

Fingerprint

Dive into the research topics of 'A parametric spectral model for texture-based salience'. Together they form a unique fingerprint.

Cite this