Global Similarity with Additive Smoothness for Spectral Segmentation

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

49 Downloads (Pure)

Abstract

Faithful representation of pairwise pixel affinities is crucial for the outcome of spectral segmentation methods. In conventional affinity models only close-range pixels interact, and a variety of subsequent techniques aims at faster propagation of local grouping cues across longrange connections. In this paper we propose a general framework for constructing a full-range affinity matrix. Our affinity matrix consists of a global similarity matrix and an additive proximity matrix. The similarity in appearance, including intensity and texture, is encoded for each pair of image pixels. Despite being full-range, our similarity matrix has a simple decomposition, which exploits an assignment of image pixels to dictionary elements. The additive proximity enforces smoothness to the segmentation by imposing interactions between near-by pixels. Our approach allows us to assess the advantages of using a full-range affinity for various spectral segmentation problems. Within our general framework we develop a few variants of full affinity for experimental validation. The performance we accomplish on composite textured images is excellent, and the results on natural images are promising.
Original languageEnglish
Title of host publicationScale Space and Variational Methods in Computer Vision
PublisherSpringer
Publication date2019
Pages357-368
ISBN (Print)978-3-030-22367-0
DOIs
Publication statusPublished - 2019
Event7th International Conference on Scale Space and Variational Methods in Computer Vision - Evangelische Tagungsstätte Hofgeismar, Hofgeismar, Germany
Duration: 30 Jun 20194 Jul 2019

Conference

Conference7th International Conference on Scale Space and Variational Methods in Computer Vision
LocationEvangelische Tagungsstätte Hofgeismar
CountryGermany
CityHofgeismar
Period30/06/201904/07/2019
SeriesLecture Notes in Computer Science
Volume11603
ISSN0302-9743

Keywords

  • Image segmentation
  • Spectral methods
  • Affinity matrix

Cite this

Dahl, V. A., & Dahl, A. B. (2019). Global Similarity with Additive Smoothness for Spectral Segmentation. In Scale Space and Variational Methods in Computer Vision (pp. 357-368). Springer. Lecture Notes in Computer Science, Vol.. 11603 https://doi.org/10.1007/978-3-030-22368-7_28
Dahl, Vedrana Andersen ; Dahl, Anders Bjorholm. / Global Similarity with Additive Smoothness for Spectral Segmentation. Scale Space and Variational Methods in Computer Vision. Springer, 2019. pp. 357-368 (Lecture Notes in Computer Science, Vol. 11603).
@inproceedings{5b97992e871248559427fd526d621ce6,
title = "Global Similarity with Additive Smoothness for Spectral Segmentation",
abstract = "Faithful representation of pairwise pixel affinities is crucial for the outcome of spectral segmentation methods. In conventional affinity models only close-range pixels interact, and a variety of subsequent techniques aims at faster propagation of local grouping cues across longrange connections. In this paper we propose a general framework for constructing a full-range affinity matrix. Our affinity matrix consists of a global similarity matrix and an additive proximity matrix. The similarity in appearance, including intensity and texture, is encoded for each pair of image pixels. Despite being full-range, our similarity matrix has a simple decomposition, which exploits an assignment of image pixels to dictionary elements. The additive proximity enforces smoothness to the segmentation by imposing interactions between near-by pixels. Our approach allows us to assess the advantages of using a full-range affinity for various spectral segmentation problems. Within our general framework we develop a few variants of full affinity for experimental validation. The performance we accomplish on composite textured images is excellent, and the results on natural images are promising.",
keywords = "Image segmentation, Spectral methods, Affinity matrix",
author = "Dahl, {Vedrana Andersen} and Dahl, {Anders Bjorholm}",
year = "2019",
doi = "10.1007/978-3-030-22368-7_28",
language = "English",
isbn = "978-3-030-22367-0",
pages = "357--368",
booktitle = "Scale Space and Variational Methods in Computer Vision",
publisher = "Springer",

}

Dahl, VA & Dahl, AB 2019, Global Similarity with Additive Smoothness for Spectral Segmentation. in Scale Space and Variational Methods in Computer Vision. Springer, Lecture Notes in Computer Science, vol. 11603, pp. 357-368, 7th International Conference on Scale Space and Variational Methods in Computer Vision, Hofgeismar, Germany, 30/06/2019. https://doi.org/10.1007/978-3-030-22368-7_28

Global Similarity with Additive Smoothness for Spectral Segmentation. / Dahl, Vedrana Andersen; Dahl, Anders Bjorholm.

Scale Space and Variational Methods in Computer Vision. Springer, 2019. p. 357-368 (Lecture Notes in Computer Science, Vol. 11603).

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - Global Similarity with Additive Smoothness for Spectral Segmentation

AU - Dahl, Vedrana Andersen

AU - Dahl, Anders Bjorholm

PY - 2019

Y1 - 2019

N2 - Faithful representation of pairwise pixel affinities is crucial for the outcome of spectral segmentation methods. In conventional affinity models only close-range pixels interact, and a variety of subsequent techniques aims at faster propagation of local grouping cues across longrange connections. In this paper we propose a general framework for constructing a full-range affinity matrix. Our affinity matrix consists of a global similarity matrix and an additive proximity matrix. The similarity in appearance, including intensity and texture, is encoded for each pair of image pixels. Despite being full-range, our similarity matrix has a simple decomposition, which exploits an assignment of image pixels to dictionary elements. The additive proximity enforces smoothness to the segmentation by imposing interactions between near-by pixels. Our approach allows us to assess the advantages of using a full-range affinity for various spectral segmentation problems. Within our general framework we develop a few variants of full affinity for experimental validation. The performance we accomplish on composite textured images is excellent, and the results on natural images are promising.

AB - Faithful representation of pairwise pixel affinities is crucial for the outcome of spectral segmentation methods. In conventional affinity models only close-range pixels interact, and a variety of subsequent techniques aims at faster propagation of local grouping cues across longrange connections. In this paper we propose a general framework for constructing a full-range affinity matrix. Our affinity matrix consists of a global similarity matrix and an additive proximity matrix. The similarity in appearance, including intensity and texture, is encoded for each pair of image pixels. Despite being full-range, our similarity matrix has a simple decomposition, which exploits an assignment of image pixels to dictionary elements. The additive proximity enforces smoothness to the segmentation by imposing interactions between near-by pixels. Our approach allows us to assess the advantages of using a full-range affinity for various spectral segmentation problems. Within our general framework we develop a few variants of full affinity for experimental validation. The performance we accomplish on composite textured images is excellent, and the results on natural images are promising.

KW - Image segmentation

KW - Spectral methods

KW - Affinity matrix

U2 - 10.1007/978-3-030-22368-7_28

DO - 10.1007/978-3-030-22368-7_28

M3 - Article in proceedings

SN - 978-3-030-22367-0

SP - 357

EP - 368

BT - Scale Space and Variational Methods in Computer Vision

PB - Springer

ER -

Dahl VA, Dahl AB. Global Similarity with Additive Smoothness for Spectral Segmentation. In Scale Space and Variational Methods in Computer Vision. Springer. 2019. p. 357-368. (Lecture Notes in Computer Science, Vol. 11603). https://doi.org/10.1007/978-3-030-22368-7_28