Unsupervised orthogonal subspace projection approach to magnetic resonance image classification

Chuin Mu Wang, Clayton Chi Chang Chen, Sheng Chi Yang, Pau Choo Chung, Yi Nung Chung, Ching Wen Yang, Chein I. Chang

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

MR images and remotely sensed images share similar image structures and characteristics because they are acquired remotely as image sequences by spectral channels at different wavelengths. As a result, techniques developed for one may be also applicable to the other. In the past, we have witnessed that some techniques that were developed for magnetic resonance imaging (MRI) found great success in remote sensing image applications. Unfortunately, the opposite direction is yet to be investigated. In this paper, we present an application of one successful remote sensing image classification technique, called orthogonal subspace projection (OSP), to magnetic resonance image classification. Unlike classical image classification techniques, which are designed on a pure pixel basis, OSP is a mixed pixel classification technique that models an image pixel as a linear mixture of different material substances assumed to be present in the image data, then estimates the abundance fraction of each individual material substance within a pixel for classification. Technically, such mixed pixel classification is performed by estimating the abundance fractions of material substances resident in a pixel, rather than assigning a class label to it as usually done in pure-pixel-based classification techniques such as a minimum-distance or nearest-neighbor rule. The advantage of mixed pixel classification has been demonstrated in many applications in remote sensing image processing. The MRI experiments reported in this paper further show that mixed pixel classification may have advantages over the pure pixel classification.

Original languageEnglish
Pages (from-to)1546-1557
Number of pages12
JournalOptical Engineering
Volume41
Issue number7
DOIs
Publication statusPublished - 2002 Jul

Fingerprint

image classification
Image classification
Magnetic resonance
magnetic resonance
projection
Pixels
pixels
remote sensing
Remote sensing
Imaging techniques
image processing
Labels
Image processing
estimating

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics
  • Engineering(all)

Cite this

Wang, C. M., Chen, C. C. C., Yang, S. C., Chung, P. C., Chung, Y. N., Yang, C. W., & Chang, C. I. (2002). Unsupervised orthogonal subspace projection approach to magnetic resonance image classification. Optical Engineering, 41(7), 1546-1557. https://doi.org/10.1117/1.1479710
Wang, Chuin Mu ; Chen, Clayton Chi Chang ; Yang, Sheng Chi ; Chung, Pau Choo ; Chung, Yi Nung ; Yang, Ching Wen ; Chang, Chein I. / Unsupervised orthogonal subspace projection approach to magnetic resonance image classification. In: Optical Engineering. 2002 ; Vol. 41, No. 7. pp. 1546-1557.
@article{83114e8794f8496b9d18f05451e2c19b,
title = "Unsupervised orthogonal subspace projection approach to magnetic resonance image classification",
abstract = "MR images and remotely sensed images share similar image structures and characteristics because they are acquired remotely as image sequences by spectral channels at different wavelengths. As a result, techniques developed for one may be also applicable to the other. In the past, we have witnessed that some techniques that were developed for magnetic resonance imaging (MRI) found great success in remote sensing image applications. Unfortunately, the opposite direction is yet to be investigated. In this paper, we present an application of one successful remote sensing image classification technique, called orthogonal subspace projection (OSP), to magnetic resonance image classification. Unlike classical image classification techniques, which are designed on a pure pixel basis, OSP is a mixed pixel classification technique that models an image pixel as a linear mixture of different material substances assumed to be present in the image data, then estimates the abundance fraction of each individual material substance within a pixel for classification. Technically, such mixed pixel classification is performed by estimating the abundance fractions of material substances resident in a pixel, rather than assigning a class label to it as usually done in pure-pixel-based classification techniques such as a minimum-distance or nearest-neighbor rule. The advantage of mixed pixel classification has been demonstrated in many applications in remote sensing image processing. The MRI experiments reported in this paper further show that mixed pixel classification may have advantages over the pure pixel classification.",
author = "Wang, {Chuin Mu} and Chen, {Clayton Chi Chang} and Yang, {Sheng Chi} and Chung, {Pau Choo} and Chung, {Yi Nung} and Yang, {Ching Wen} and Chang, {Chein I.}",
year = "2002",
month = "7",
doi = "10.1117/1.1479710",
language = "English",
volume = "41",
pages = "1546--1557",
journal = "Optical Engineering",
issn = "0091-3286",
publisher = "SPIE",
number = "7",

}

Wang, CM, Chen, CCC, Yang, SC, Chung, PC, Chung, YN, Yang, CW & Chang, CI 2002, 'Unsupervised orthogonal subspace projection approach to magnetic resonance image classification', Optical Engineering, vol. 41, no. 7, pp. 1546-1557. https://doi.org/10.1117/1.1479710

Unsupervised orthogonal subspace projection approach to magnetic resonance image classification. / Wang, Chuin Mu; Chen, Clayton Chi Chang; Yang, Sheng Chi; Chung, Pau Choo; Chung, Yi Nung; Yang, Ching Wen; Chang, Chein I.

In: Optical Engineering, Vol. 41, No. 7, 07.2002, p. 1546-1557.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Unsupervised orthogonal subspace projection approach to magnetic resonance image classification

AU - Wang, Chuin Mu

AU - Chen, Clayton Chi Chang

AU - Yang, Sheng Chi

AU - Chung, Pau Choo

AU - Chung, Yi Nung

AU - Yang, Ching Wen

AU - Chang, Chein I.

PY - 2002/7

Y1 - 2002/7

N2 - MR images and remotely sensed images share similar image structures and characteristics because they are acquired remotely as image sequences by spectral channels at different wavelengths. As a result, techniques developed for one may be also applicable to the other. In the past, we have witnessed that some techniques that were developed for magnetic resonance imaging (MRI) found great success in remote sensing image applications. Unfortunately, the opposite direction is yet to be investigated. In this paper, we present an application of one successful remote sensing image classification technique, called orthogonal subspace projection (OSP), to magnetic resonance image classification. Unlike classical image classification techniques, which are designed on a pure pixel basis, OSP is a mixed pixel classification technique that models an image pixel as a linear mixture of different material substances assumed to be present in the image data, then estimates the abundance fraction of each individual material substance within a pixel for classification. Technically, such mixed pixel classification is performed by estimating the abundance fractions of material substances resident in a pixel, rather than assigning a class label to it as usually done in pure-pixel-based classification techniques such as a minimum-distance or nearest-neighbor rule. The advantage of mixed pixel classification has been demonstrated in many applications in remote sensing image processing. The MRI experiments reported in this paper further show that mixed pixel classification may have advantages over the pure pixel classification.

AB - MR images and remotely sensed images share similar image structures and characteristics because they are acquired remotely as image sequences by spectral channels at different wavelengths. As a result, techniques developed for one may be also applicable to the other. In the past, we have witnessed that some techniques that were developed for magnetic resonance imaging (MRI) found great success in remote sensing image applications. Unfortunately, the opposite direction is yet to be investigated. In this paper, we present an application of one successful remote sensing image classification technique, called orthogonal subspace projection (OSP), to magnetic resonance image classification. Unlike classical image classification techniques, which are designed on a pure pixel basis, OSP is a mixed pixel classification technique that models an image pixel as a linear mixture of different material substances assumed to be present in the image data, then estimates the abundance fraction of each individual material substance within a pixel for classification. Technically, such mixed pixel classification is performed by estimating the abundance fractions of material substances resident in a pixel, rather than assigning a class label to it as usually done in pure-pixel-based classification techniques such as a minimum-distance or nearest-neighbor rule. The advantage of mixed pixel classification has been demonstrated in many applications in remote sensing image processing. The MRI experiments reported in this paper further show that mixed pixel classification may have advantages over the pure pixel classification.

UR - http://www.scopus.com/inward/record.url?scp=0036648641&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036648641&partnerID=8YFLogxK

U2 - 10.1117/1.1479710

DO - 10.1117/1.1479710

M3 - Article

AN - SCOPUS:0036648641

VL - 41

SP - 1546

EP - 1557

JO - Optical Engineering

JF - Optical Engineering

SN - 0091-3286

IS - 7

ER -