Approximating Non-linear Diffusion

Erik B. Dam; Ole Fogh Olsen; Mads Nielsen

doi:10.1007/3-540-44935-3_9

Approximating Non-linear Diffusion

Erik B. Dam, Ole Fogh Olsen, Mads Nielsen

Abstract

We assess the feasibility of approximating non-linear diffusion processes with simple local Gaussian filters. The purpose of doing this is twofold. Firstly, the theoretical implications are by themselves interesting. Secondly, a successful method would reduce the need for computationally expensive implementations of non-linear diffusion schemes. We evaluate using isotropic and affine Gaussian filters for the task of approximating the local diffusion for a number of non-linear diffusion schemes. The approximations are firstly explored using an information theoretical approach and secondly evaluated based on their performance on a multi-scale segmentation application.

The results show that while the approximations do not perform quite as well as the original non-linear scheme, the decrease in performance is acceptable for the evaluated task. Furthermore, the affine approximations perform significantly better than the isotropic.

Original language	English
Title of host publication	Scale Space Methods in Computer Vision : 4th International Conference, Scale Space 2003 Isle of Skye, UK, June 10–12, 2003 Proceedings
Publisher	<Forlag uden navn>
Publication date	2003
Pages	117-131
ISBN (Print)	978-3-540-40368-5
DOIs	https://doi.org/10.1007/3-540-44935-3_9
Publication status	Published - 2003
Externally published	Yes
Event	Proceedings of the 4th International Conference in Scale Space - Isle of Skye, United Kingdom Duration: 29 Nov 2010 → … Conference number: 4

Conference

Conference	Proceedings of the 4th International Conference in Scale Space
Number	4
Country/Territory	United Kingdom
City	Isle of Skye
Period	29/11/2010 → …

Series	Lecture notes in computer science
Volume	2695/2003
ISSN	0302-9743

Access to Document

10.1007/3-540-44935-3_9

http://www.springerlink.de/content/b438x851p21615j0/fulltext.pdf

Cite this

Approximating Non-linear Diffusion. / Dam, Erik B.; Olsen, Ole Fogh; Nielsen, Mads.

Scale Space Methods in Computer Vision: 4th International Conference, Scale Space 2003 Isle of Skye, UK, June 10–12, 2003 Proceedings. <Forlag uden navn>, 2003. p. 117-131 (Lecture notes in computer science, Vol. 2695/2003).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Dam, EB, Olsen, OF & Nielsen, M 2003, Approximating Non-linear Diffusion. in Scale Space Methods in Computer Vision: 4th International Conference, Scale Space 2003 Isle of Skye, UK, June 10–12, 2003 Proceedings. <Forlag uden navn>, Lecture notes in computer science, vol. 2695/2003, pp. 117-131, Proceedings of the 4th International Conference in Scale Space, Isle of Skye, United Kingdom, 29/11/2010. https://doi.org/10.1007/3-540-44935-3_9

@inproceedings{88951ee06d2311dd8d9f000ea68e967b,

title = "Approximating Non-linear Diffusion",

abstract = "We assess the feasibility of approximating non-linear diffusion processes with simple local Gaussian filters. The purpose of doing this is twofold. Firstly, the theoretical implications are by themselves interesting. Secondly, a successful method would reduce the need for computationally expensive implementations of non-linear diffusion schemes. We evaluate using isotropic and affine Gaussian filters for the task of approximating the local diffusion for a number of non-linear diffusion schemes. The approximations are firstly explored using an information theoretical approach and secondly evaluated based on their performance on a multi-scale segmentation application. The results show that while the approximations do not perform quite as well as the original non-linear scheme, the decrease in performance is acceptable for the evaluated task. Furthermore, the affine approximations perform significantly better than the isotropic. ",

author = "Dam, {Erik B.} and Olsen, {Ole Fogh} and Mads Nielsen",

year = "2003",

doi = "10.1007/3-540-44935-3_9",

language = "English",

isbn = "978-3-540-40368-5",

series = "Lecture notes in computer science",

publisher = "<Forlag uden navn>",

pages = "117--131",

booktitle = "Scale Space Methods in Computer Vision",

note = "Proceedings of the 4th International Conference in Scale Space ; Conference date: 29-11-2010",

}

TY - GEN

T1 - Approximating Non-linear Diffusion

AU - Dam, Erik B.

AU - Olsen, Ole Fogh

AU - Nielsen, Mads

N1 - Conference code: 4

PY - 2003

Y1 - 2003

N2 - We assess the feasibility of approximating non-linear diffusion processes with simple local Gaussian filters. The purpose of doing this is twofold. Firstly, the theoretical implications are by themselves interesting. Secondly, a successful method would reduce the need for computationally expensive implementations of non-linear diffusion schemes. We evaluate using isotropic and affine Gaussian filters for the task of approximating the local diffusion for a number of non-linear diffusion schemes. The approximations are firstly explored using an information theoretical approach and secondly evaluated based on their performance on a multi-scale segmentation application. The results show that while the approximations do not perform quite as well as the original non-linear scheme, the decrease in performance is acceptable for the evaluated task. Furthermore, the affine approximations perform significantly better than the isotropic.

AB - We assess the feasibility of approximating non-linear diffusion processes with simple local Gaussian filters. The purpose of doing this is twofold. Firstly, the theoretical implications are by themselves interesting. Secondly, a successful method would reduce the need for computationally expensive implementations of non-linear diffusion schemes. We evaluate using isotropic and affine Gaussian filters for the task of approximating the local diffusion for a number of non-linear diffusion schemes. The approximations are firstly explored using an information theoretical approach and secondly evaluated based on their performance on a multi-scale segmentation application. The results show that while the approximations do not perform quite as well as the original non-linear scheme, the decrease in performance is acceptable for the evaluated task. Furthermore, the affine approximations perform significantly better than the isotropic.

U2 - 10.1007/3-540-44935-3_9

DO - 10.1007/3-540-44935-3_9

M3 - Article in proceedings

SN - 978-3-540-40368-5

T3 - Lecture notes in computer science

SP - 117

EP - 131

BT - Scale Space Methods in Computer Vision

PB - <Forlag uden navn>

T2 - Proceedings of the 4th International Conference in Scale Space

Y2 - 29 November 2010

ER -