Robust anisotropic diffusion

Michael J. Black, Guillermo Sapiro, David H. Marimont, David Heeger

Research output: Contribution to journalArticle

Abstract

Relations between anisotropic diffusion and robust statistics are described in this paper. Specifically, we show that anisotropic diffusion can be seen as a robust estimation procedure that estimates a piecewise smooth image from a noisy input image. The "edge-stopping" function in the anisotropic diffusion equation is closely related to the error norm and influence function in the robust estimation framework. This connection leads to a new "edge-stopping" function based on Tukey's biweight robust estimator that preserves sharper boundaries than previous formulations and improves the automatic stopping of the diffusion. The robust statistical interpretation also provides a means for detecting the boundaries (edges) between the piecewise smooth regions in an image that has been smoothed with anisotropic diffusion. Additionally, we derive a relationship between anisotropic diffusion and regularization with line processes. Adding constraints on the spatial organization of the line processes allows us to develop new anisotropic diffusion equations that result in a qualitative improvement in the continuity of edges.

Original languageEnglish (US)
Pages (from-to)421-432
Number of pages12
JournalIEEE Transactions on Image Processing
Volume7
Issue number3
DOIs
StatePublished - 1998

Fingerprint

Anisotropic Diffusion
Robust Estimation
Diffusion equation
Robust Statistics
Influence Function
Robust Estimators
Line
Regularization
Norm
Statistics
Formulation
Estimate

Keywords

  • Anisotropic diffusion
  • Line processes
  • Robust statistics

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Graphics and Computer-Aided Design
  • Software
  • Theoretical Computer Science
  • Computational Theory and Mathematics
  • Computer Vision and Pattern Recognition

Cite this

Robust anisotropic diffusion. / Black, Michael J.; Sapiro, Guillermo; Marimont, David H.; Heeger, David.

In: IEEE Transactions on Image Processing, Vol. 7, No. 3, 1998, p. 421-432.

Research output: Contribution to journalArticle

Black, MJ, Sapiro, G, Marimont, DH & Heeger, D 1998, 'Robust anisotropic diffusion', IEEE Transactions on Image Processing, vol. 7, no. 3, pp. 421-432. https://doi.org/10.1109/83.661192
Black, Michael J. ; Sapiro, Guillermo ; Marimont, David H. ; Heeger, David. / Robust anisotropic diffusion. In: IEEE Transactions on Image Processing. 1998 ; Vol. 7, No. 3. pp. 421-432.
@article{486e36a52d2c4c28a21f011bdcba3649,
title = "Robust anisotropic diffusion",
abstract = "Relations between anisotropic diffusion and robust statistics are described in this paper. Specifically, we show that anisotropic diffusion can be seen as a robust estimation procedure that estimates a piecewise smooth image from a noisy input image. The {"}edge-stopping{"} function in the anisotropic diffusion equation is closely related to the error norm and influence function in the robust estimation framework. This connection leads to a new {"}edge-stopping{"} function based on Tukey's biweight robust estimator that preserves sharper boundaries than previous formulations and improves the automatic stopping of the diffusion. The robust statistical interpretation also provides a means for detecting the boundaries (edges) between the piecewise smooth regions in an image that has been smoothed with anisotropic diffusion. Additionally, we derive a relationship between anisotropic diffusion and regularization with line processes. Adding constraints on the spatial organization of the line processes allows us to develop new anisotropic diffusion equations that result in a qualitative improvement in the continuity of edges.",
keywords = "Anisotropic diffusion, Line processes, Robust statistics",
author = "Black, {Michael J.} and Guillermo Sapiro and Marimont, {David H.} and David Heeger",
year = "1998",
doi = "10.1109/83.661192",
language = "English (US)",
volume = "7",
pages = "421--432",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "3",

}

TY - JOUR

T1 - Robust anisotropic diffusion

AU - Black, Michael J.

AU - Sapiro, Guillermo

AU - Marimont, David H.

AU - Heeger, David

PY - 1998

Y1 - 1998

N2 - Relations between anisotropic diffusion and robust statistics are described in this paper. Specifically, we show that anisotropic diffusion can be seen as a robust estimation procedure that estimates a piecewise smooth image from a noisy input image. The "edge-stopping" function in the anisotropic diffusion equation is closely related to the error norm and influence function in the robust estimation framework. This connection leads to a new "edge-stopping" function based on Tukey's biweight robust estimator that preserves sharper boundaries than previous formulations and improves the automatic stopping of the diffusion. The robust statistical interpretation also provides a means for detecting the boundaries (edges) between the piecewise smooth regions in an image that has been smoothed with anisotropic diffusion. Additionally, we derive a relationship between anisotropic diffusion and regularization with line processes. Adding constraints on the spatial organization of the line processes allows us to develop new anisotropic diffusion equations that result in a qualitative improvement in the continuity of edges.

AB - Relations between anisotropic diffusion and robust statistics are described in this paper. Specifically, we show that anisotropic diffusion can be seen as a robust estimation procedure that estimates a piecewise smooth image from a noisy input image. The "edge-stopping" function in the anisotropic diffusion equation is closely related to the error norm and influence function in the robust estimation framework. This connection leads to a new "edge-stopping" function based on Tukey's biweight robust estimator that preserves sharper boundaries than previous formulations and improves the automatic stopping of the diffusion. The robust statistical interpretation also provides a means for detecting the boundaries (edges) between the piecewise smooth regions in an image that has been smoothed with anisotropic diffusion. Additionally, we derive a relationship between anisotropic diffusion and regularization with line processes. Adding constraints on the spatial organization of the line processes allows us to develop new anisotropic diffusion equations that result in a qualitative improvement in the continuity of edges.

KW - Anisotropic diffusion

KW - Line processes

KW - Robust statistics

UR - http://www.scopus.com/inward/record.url?scp=0032023715&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032023715&partnerID=8YFLogxK

U2 - 10.1109/83.661192

DO - 10.1109/83.661192

M3 - Article

VL - 7

SP - 421

EP - 432

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 3

ER -