A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology

Soumyajit Gupta, Shachi Mittal, Andre Kajdacsy-Balla, Rohit Bhargava, Chandrajit Bajaj

Research output: Contribution to journalArticle

Abstract

Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been extensively studied as a method for noise removal in HSI data. It too, however entails a manual speed-accuracy trade-off, namely the process of manually selecting the relevant bands in the MNF space. This process currently takes roughly around a month’s time for acquiring and pre-process-ing an entire TMA with acceptable signal to noise ratio. We present three approaches termed ‘Fast MNF’, ‘Approx MNF’ and ‘Rand MNF’ where the computational time of the algorithm is reduced, as well as the entire process of band selection is fully automated. This automated approach is shown to perform at the same level of accuracy as MNF with now large speedup factors, resulting in the same task to be accomplished in hours. The different approximations produced by the three algorithms, show the reconstruction accuracy vs storage (50×) and runtime speed (60×) trade-off. We apply the approach for automating the denoising of different tissue histology samples, in which the accuracy of classification (differentiating between the different histologic and pathologic classes) strongly depends on the SNR (signal to noise ratio) of recovered data. Therefore, we also compare the effect of the proposed denoising algorithms on classification accuracy. Since denoising HSI data is done unsupervised, we also use a metric that assesses the quality of denoising in the image domain between the noisy and denoised image in the absence of ground truth.

Original languageEnglish (US)
Article numbere0205219
JournalPloS one
Volume14
Issue number4
DOIs
StatePublished - Apr 1 2019

Fingerprint

histopathology
Noise
image analysis
Imaging techniques
Signal to noise ratio
Histology
Signal-To-Noise Ratio
Tissue
Hyperspectral imaging
histology

ASJC Scopus subject areas

  • Biochemistry, Genetics and Molecular Biology(all)
  • Agricultural and Biological Sciences(all)

Cite this

A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology. / Gupta, Soumyajit; Mittal, Shachi; Kajdacsy-Balla, Andre; Bhargava, Rohit; Bajaj, Chandrajit.

In: PloS one, Vol. 14, No. 4, e0205219, 01.04.2019.

Research output: Contribution to journalArticle

Gupta, Soumyajit ; Mittal, Shachi ; Kajdacsy-Balla, Andre ; Bhargava, Rohit ; Bajaj, Chandrajit. / A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology. In: PloS one. 2019 ; Vol. 14, No. 4.
@article{89990c1bd4df460c8589c5b56d986a71,
title = "A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology",
abstract = "Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been extensively studied as a method for noise removal in HSI data. It too, however entails a manual speed-accuracy trade-off, namely the process of manually selecting the relevant bands in the MNF space. This process currently takes roughly around a month’s time for acquiring and pre-process-ing an entire TMA with acceptable signal to noise ratio. We present three approaches termed ‘Fast MNF’, ‘Approx MNF’ and ‘Rand MNF’ where the computational time of the algorithm is reduced, as well as the entire process of band selection is fully automated. This automated approach is shown to perform at the same level of accuracy as MNF with now large speedup factors, resulting in the same task to be accomplished in hours. The different approximations produced by the three algorithms, show the reconstruction accuracy vs storage (50×) and runtime speed (60×) trade-off. We apply the approach for automating the denoising of different tissue histology samples, in which the accuracy of classification (differentiating between the different histologic and pathologic classes) strongly depends on the SNR (signal to noise ratio) of recovered data. Therefore, we also compare the effect of the proposed denoising algorithms on classification accuracy. Since denoising HSI data is done unsupervised, we also use a metric that assesses the quality of denoising in the image domain between the noisy and denoised image in the absence of ground truth.",
author = "Soumyajit Gupta and Shachi Mittal and Andre Kajdacsy-Balla and Rohit Bhargava and Chandrajit Bajaj",
year = "2019",
month = "4",
day = "1",
doi = "10.1371/journal.pone.0205219",
language = "English (US)",
volume = "14",
journal = "PLoS One",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "4",

}

TY - JOUR

T1 - A fully automated, faster noise rejection approach to increasing the analytical capability of chemical imaging for digital histopathology

AU - Gupta, Soumyajit

AU - Mittal, Shachi

AU - Kajdacsy-Balla, Andre

AU - Bhargava, Rohit

AU - Bajaj, Chandrajit

PY - 2019/4/1

Y1 - 2019/4/1

N2 - Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been extensively studied as a method for noise removal in HSI data. It too, however entails a manual speed-accuracy trade-off, namely the process of manually selecting the relevant bands in the MNF space. This process currently takes roughly around a month’s time for acquiring and pre-process-ing an entire TMA with acceptable signal to noise ratio. We present three approaches termed ‘Fast MNF’, ‘Approx MNF’ and ‘Rand MNF’ where the computational time of the algorithm is reduced, as well as the entire process of band selection is fully automated. This automated approach is shown to perform at the same level of accuracy as MNF with now large speedup factors, resulting in the same task to be accomplished in hours. The different approximations produced by the three algorithms, show the reconstruction accuracy vs storage (50×) and runtime speed (60×) trade-off. We apply the approach for automating the denoising of different tissue histology samples, in which the accuracy of classification (differentiating between the different histologic and pathologic classes) strongly depends on the SNR (signal to noise ratio) of recovered data. Therefore, we also compare the effect of the proposed denoising algorithms on classification accuracy. Since denoising HSI data is done unsupervised, we also use a metric that assesses the quality of denoising in the image domain between the noisy and denoised image in the absence of ground truth.

AB - Chemical hyperspectral imaging (HSI) data is naturally high dimensional and large. There are thus inherent manual trade-offs in acquisition time, and the quality of data. Minimum Noise Fraction (MNF) developed by Green et al. [1] has been extensively studied as a method for noise removal in HSI data. It too, however entails a manual speed-accuracy trade-off, namely the process of manually selecting the relevant bands in the MNF space. This process currently takes roughly around a month’s time for acquiring and pre-process-ing an entire TMA with acceptable signal to noise ratio. We present three approaches termed ‘Fast MNF’, ‘Approx MNF’ and ‘Rand MNF’ where the computational time of the algorithm is reduced, as well as the entire process of band selection is fully automated. This automated approach is shown to perform at the same level of accuracy as MNF with now large speedup factors, resulting in the same task to be accomplished in hours. The different approximations produced by the three algorithms, show the reconstruction accuracy vs storage (50×) and runtime speed (60×) trade-off. We apply the approach for automating the denoising of different tissue histology samples, in which the accuracy of classification (differentiating between the different histologic and pathologic classes) strongly depends on the SNR (signal to noise ratio) of recovered data. Therefore, we also compare the effect of the proposed denoising algorithms on classification accuracy. Since denoising HSI data is done unsupervised, we also use a metric that assesses the quality of denoising in the image domain between the noisy and denoised image in the absence of ground truth.

UR - http://www.scopus.com/inward/record.url?scp=85064606642&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064606642&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0205219

DO - 10.1371/journal.pone.0205219

M3 - Article

VL - 14

JO - PLoS One

JF - PLoS One

SN - 1932-6203

IS - 4

M1 - e0205219

ER -