Robust optimal granulometric bandpass filters

Edward R. Dougherty, Yidong Chen

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

A granulometric bandpass filter (GBF) passes certain size bands in a binary image. There exists an analytic formulation for the optimization of GBFs in terms of the granulometric size density of a random set. This size density plays the role of the power spectral density for optimization of linear filters. An optimal filter depends on the parameters governing the distribution of the random set. In practice, the filter will be applied in statistical conditions that do not exactly match those for which it has been designed. Hence the robustness question: to what extent does filter performance degrade as conditions vary from those under which it has been designed? This paper considers GBF robustness. It examines three issues: minimax robust filters, Bayesian robust filters, and global filters. A minimax robust filter is one whose worst performance over all states of nature is best among the optimal filters over all states. Minimax robustness does not take into account the probability mass of the states of nature. Bayesian robustness analysis takes state mass into account and is focused on mean robustness, which is the expected error increase owing to applying a filter designed for a specific state over all possible states. We would like to find maximally robust states. Finally, we consider global filters. A global filter is designed according to some mean condition of the states and is applied across all states. Here, we hope for a uniformly most robust global filter, which means that its expected error increase across all states is less than the mean robustness for any state-specific filter. At least we would like a global filter whose expected increase is close to that for a maximally robust state.

Original languageEnglish (US)
Pages (from-to)1357-1372
Number of pages16
JournalSignal Processing
Volume81
Issue number7
DOIs
StatePublished - Jul 2001
Externally publishedYes

Fingerprint

Bandpass filters
Binary images
Power spectral density

Keywords

  • Bayesian robust filter
  • Granulometry
  • Minimax robust filter
  • Optimal filter

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Robust optimal granulometric bandpass filters. / Dougherty, Edward R.; Chen, Yidong.

In: Signal Processing, Vol. 81, No. 7, 07.2001, p. 1357-1372.

Research output: Contribution to journalArticle

Dougherty, Edward R. ; Chen, Yidong. / Robust optimal granulometric bandpass filters. In: Signal Processing. 2001 ; Vol. 81, No. 7. pp. 1357-1372.
@article{ef495200459e45bf9ea2b601a07234cf,
title = "Robust optimal granulometric bandpass filters",
abstract = "A granulometric bandpass filter (GBF) passes certain size bands in a binary image. There exists an analytic formulation for the optimization of GBFs in terms of the granulometric size density of a random set. This size density plays the role of the power spectral density for optimization of linear filters. An optimal filter depends on the parameters governing the distribution of the random set. In practice, the filter will be applied in statistical conditions that do not exactly match those for which it has been designed. Hence the robustness question: to what extent does filter performance degrade as conditions vary from those under which it has been designed? This paper considers GBF robustness. It examines three issues: minimax robust filters, Bayesian robust filters, and global filters. A minimax robust filter is one whose worst performance over all states of nature is best among the optimal filters over all states. Minimax robustness does not take into account the probability mass of the states of nature. Bayesian robustness analysis takes state mass into account and is focused on mean robustness, which is the expected error increase owing to applying a filter designed for a specific state over all possible states. We would like to find maximally robust states. Finally, we consider global filters. A global filter is designed according to some mean condition of the states and is applied across all states. Here, we hope for a uniformly most robust global filter, which means that its expected error increase across all states is less than the mean robustness for any state-specific filter. At least we would like a global filter whose expected increase is close to that for a maximally robust state.",
keywords = "Bayesian robust filter, Granulometry, Minimax robust filter, Optimal filter",
author = "Dougherty, {Edward R.} and Yidong Chen",
year = "2001",
month = "7",
doi = "10.1016/S0165-1684(01)00017-2",
language = "English (US)",
volume = "81",
pages = "1357--1372",
journal = "Signal Processing",
issn = "0165-1684",
publisher = "Elsevier",
number = "7",

}

TY - JOUR

T1 - Robust optimal granulometric bandpass filters

AU - Dougherty, Edward R.

AU - Chen, Yidong

PY - 2001/7

Y1 - 2001/7

N2 - A granulometric bandpass filter (GBF) passes certain size bands in a binary image. There exists an analytic formulation for the optimization of GBFs in terms of the granulometric size density of a random set. This size density plays the role of the power spectral density for optimization of linear filters. An optimal filter depends on the parameters governing the distribution of the random set. In practice, the filter will be applied in statistical conditions that do not exactly match those for which it has been designed. Hence the robustness question: to what extent does filter performance degrade as conditions vary from those under which it has been designed? This paper considers GBF robustness. It examines three issues: minimax robust filters, Bayesian robust filters, and global filters. A minimax robust filter is one whose worst performance over all states of nature is best among the optimal filters over all states. Minimax robustness does not take into account the probability mass of the states of nature. Bayesian robustness analysis takes state mass into account and is focused on mean robustness, which is the expected error increase owing to applying a filter designed for a specific state over all possible states. We would like to find maximally robust states. Finally, we consider global filters. A global filter is designed according to some mean condition of the states and is applied across all states. Here, we hope for a uniformly most robust global filter, which means that its expected error increase across all states is less than the mean robustness for any state-specific filter. At least we would like a global filter whose expected increase is close to that for a maximally robust state.

AB - A granulometric bandpass filter (GBF) passes certain size bands in a binary image. There exists an analytic formulation for the optimization of GBFs in terms of the granulometric size density of a random set. This size density plays the role of the power spectral density for optimization of linear filters. An optimal filter depends on the parameters governing the distribution of the random set. In practice, the filter will be applied in statistical conditions that do not exactly match those for which it has been designed. Hence the robustness question: to what extent does filter performance degrade as conditions vary from those under which it has been designed? This paper considers GBF robustness. It examines three issues: minimax robust filters, Bayesian robust filters, and global filters. A minimax robust filter is one whose worst performance over all states of nature is best among the optimal filters over all states. Minimax robustness does not take into account the probability mass of the states of nature. Bayesian robustness analysis takes state mass into account and is focused on mean robustness, which is the expected error increase owing to applying a filter designed for a specific state over all possible states. We would like to find maximally robust states. Finally, we consider global filters. A global filter is designed according to some mean condition of the states and is applied across all states. Here, we hope for a uniformly most robust global filter, which means that its expected error increase across all states is less than the mean robustness for any state-specific filter. At least we would like a global filter whose expected increase is close to that for a maximally robust state.

KW - Bayesian robust filter

KW - Granulometry

KW - Minimax robust filter

KW - Optimal filter

UR - http://www.scopus.com/inward/record.url?scp=0035392480&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0035392480&partnerID=8YFLogxK

U2 - 10.1016/S0165-1684(01)00017-2

DO - 10.1016/S0165-1684(01)00017-2

M3 - Article

AN - SCOPUS:0035392480

VL - 81

SP - 1357

EP - 1372

JO - Signal Processing

JF - Signal Processing

SN - 0165-1684

IS - 7

ER -