Image specific feature similarities

Ido Omer*, Michael Werman

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

6 Scopus citations


Calculating a reliable similarity measure between pixel features is essential for many computer vision and image processing applications. We propose a similarity measure (affinity) between pixel features, which depends on the feature space histogram of the image. We use the observation that clusters in the feature space histogram are typically smooth and roughly convex. Given two feature points we adjust their similarity according to the bottleneck in the histogram values on the straight line between them. We call our new similarities Bottleneck Affinities. These measures are computed efficiently, we demonstrate superior segmentation results compared to the use of the Euclidean metric.

Original languageAmerican English
Title of host publicationComputer Vision - ECCV 2006, 9th European Conference on Computer Vision, Proceedings
Number of pages13
StatePublished - 2006
Event9th European Conference on Computer Vision, ECCV 2006 - Graz, Austria
Duration: 7 May 200613 May 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3952 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference9th European Conference on Computer Vision, ECCV 2006


Dive into the research topics of 'Image specific feature similarities'. Together they form a unique fingerprint.

Cite this