Component-Tree Simplification through Fast Alpha Cuts

M.H.F. Wilkinson

doi:10.1007/978-3-031-19897-7_19

Component-Tree Simplification through Fast Alpha Cuts

M.H.F. Wilkinson^*

^*Corresponding author voor dit werk

Intelligente Systemen

Onderzoeksoutput › Academic › peer review

221 Downloads (Pure)

Samenvatting

Tree-based hierarchical image representations are commonly used in connected morphological image filtering, segmentation and multi-scale analysis. In the case of component trees, filtering is generally based on thresholding single attributes computed for all the nodes in the tree. Alternatively, so-called shapings are used, which rely on building a component tree of a component tree to filter the image. Neither method is practical when using vector attributes. In this case, more complicated machine learning methods are required, including clustering methods.
In this paper I present a simple, fast hierarchical clustering algorithm based on cuts of α-trees to simplify and filter component trees.

Originele taal-2	English
Titel	Discrete Geometry and Mathematical Morphology 2022
Redacteuren	É. Baudrier, B Naegel, A. Krähenbühl, M. Tajine
Plaats van productie	Cham
Uitgeverij	Springer International Publishing AG
Pagina's	236-247
Aantal pagina's	12
ISBN van elektronische versie	978-3-031-19897-7
ISBN van geprinte versie	978-3-031-19896-0
DOI's	https://doi.org/10.1007/978-3-031-19897-7_19
Status	Published - 20-okt.-2022
Evenement	Discrete Geometry and Mathematical Morphology 2022 - Starsbourg, France Duur: 24-okt.-2022 → 27-okt.-2022 https://dgmm2022.sciencesconf.org/

Publicatie series

Naam	Lecture Notes in Computer Science
Uitgeverij	Springer
Volume	13493
ISSN van geprinte versie	0302-9743
ISSN van elektronische versie	1611-3349

Conference

Conference	Discrete Geometry and Mathematical Morphology 2022
Verkorte titel	DGMM 2022
Land/Regio	France
Stad	Starsbourg
Periode	24/10/2022 → 27/10/2022
Internet adres	https://dgmm2022.sciencesconf.org/

Toegang tot document

10.1007/978-3-031-19897-7_19

Component-Tree Simplification Through Fast Alpha Cuts Final publisher's version, 45,1 MBLicentie: Taverne

Handle.net

Permanente koppeling

Citeer dit

@inproceedings{e36f7405f60844a993a80fa5012504b0,

title = "Component-Tree Simplification through Fast Alpha Cuts",

abstract = "Tree-based hierarchical image representations are commonly used in connected morphological image filtering, segmentation and multi-scale analysis. In the case of component trees, filtering is generally based on thresholding single attributes computed for all the nodes in the tree. Alternatively, so-called shapings are used, which rely on building a component tree of a component tree to filter the image. Neither method is practical when using vector attributes. In this case, more complicated machine learning methods are required, including clustering methods. In this paper I present a simple, fast hierarchical clustering algorithm based on cuts of α-trees to simplify and filter component trees. ",

author = "M.H.F. Wilkinson",

year = "2022",

month = oct,

day = "20",

doi = "10.1007/978-3-031-19897-7_19",

language = "English",

isbn = "978-3-031-19896-0",

series = "Lecture Notes in Computer Science",

publisher = "Springer International Publishing AG",

pages = "236--247",

editor = "{\'E}. Baudrier and B Naegel and Kr{\"a}henb{\"u}hl, { A.} and Tajine, {M. }",

booktitle = "Discrete Geometry and Mathematical Morphology 2022",

note = "Discrete Geometry and Mathematical Morphology 2022, DGMM 2022 ; Conference date: 24-10-2022 Through 27-10-2022",

url = "https://dgmm2022.sciencesconf.org/",

}

Wilkinson, MHF 2022, Component-Tree Simplification through Fast Alpha Cuts. in É Baudrier, B Naegel, A Krähenbühl & M Tajine (uitgave), Discrete Geometry and Mathematical Morphology 2022. Lecture Notes in Computer Science, vol. 13493, Springer International Publishing AG, Cham, blz. 236-247, Discrete Geometry and Mathematical Morphology 2022, Starsbourg, France, 24/10/2022. https://doi.org/10.1007/978-3-031-19897-7_19

TY - GEN

T1 - Component-Tree Simplification through Fast Alpha Cuts

AU - Wilkinson, M.H.F.

PY - 2022/10/20

Y1 - 2022/10/20

N2 - Tree-based hierarchical image representations are commonly used in connected morphological image filtering, segmentation and multi-scale analysis. In the case of component trees, filtering is generally based on thresholding single attributes computed for all the nodes in the tree. Alternatively, so-called shapings are used, which rely on building a component tree of a component tree to filter the image. Neither method is practical when using vector attributes. In this case, more complicated machine learning methods are required, including clustering methods. In this paper I present a simple, fast hierarchical clustering algorithm based on cuts of α-trees to simplify and filter component trees.

AB - Tree-based hierarchical image representations are commonly used in connected morphological image filtering, segmentation and multi-scale analysis. In the case of component trees, filtering is generally based on thresholding single attributes computed for all the nodes in the tree. Alternatively, so-called shapings are used, which rely on building a component tree of a component tree to filter the image. Neither method is practical when using vector attributes. In this case, more complicated machine learning methods are required, including clustering methods. In this paper I present a simple, fast hierarchical clustering algorithm based on cuts of α-trees to simplify and filter component trees.

U2 - 10.1007/978-3-031-19897-7_19

DO - 10.1007/978-3-031-19897-7_19

M3 - Conference contribution

SN - 978-3-031-19896-0

T3 - Lecture Notes in Computer Science

SP - 236

EP - 247

BT - Discrete Geometry and Mathematical Morphology 2022

A2 - Baudrier, É.

A2 - Naegel, B

A2 - Krähenbühl, A.

A2 - Tajine, M.

PB - Springer International Publishing AG

CY - Cham

T2 - Discrete Geometry and Mathematical Morphology 2022

Y2 - 24 October 2022 through 27 October 2022

ER -

Component-Tree Simplification through Fast Alpha Cuts

Samenvatting

Publicatie series

Conference

Toegang tot document

Handle.net

Vingerafdruk

Citeer dit