Combing a porcupine via stereographic direction diffusion

Nir A. Sochen, Ron Kimmel

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

2 Scopus citations

Abstract

This paper addresses the problem of feature enhancement in noisy images when the feature is known to be constrained to a manifold. As an example, we study the problem of direction denoising. This problem was treated recently and several solutions were proposed. The various solutions share the same structure. They are composed of two terms: A diffusion term and a projection term. Analytically, the solutions differ in the diffusion part. The projection part is equivalent in all works. Yet, as it is often the case, the analytically equivalent projection terms differ from a numerical viewpoint. We present in this work a new parameterization of the problem that enables us to work always in a numerically stable way.

Original languageEnglish
Title of host publicationScale-Space and Morphology in Computer Vision - 3rd International Conference, Scale-Space 2001, Proceedings
EditorsMichael Kerckhove
PublisherSpringer Verlag
Pages308-316
Number of pages9
ISBN (Electronic)9783540423171
StatePublished - 2001
Event3rd International Conference on Scale-Space and Morphology in Computer Vision, Scale-Space 2001 - Vancouver, Canada
Duration: 7 Jul 20018 Jul 2001

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2106
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd International Conference on Scale-Space and Morphology in Computer Vision, Scale-Space 2001
Country/TerritoryCanada
CityVancouver
Period7/07/018/07/01

Fingerprint

Dive into the research topics of 'Combing a porcupine via stereographic direction diffusion'. Together they form a unique fingerprint.

Cite this