Hyperspectral image enhancement and mixture deep-learning classification of corneal epithelium injury

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

In our preliminary study, the reflectance signatures obtained from hyperspectral imaging (HSI) of normal and abnormal corneal epithelium tissues of porcine show similar morphology with subtle differences. Here we present image enhancement algorithms that can be used to improve the interpretability of data into clinically relevant information to facilitate diagnostics. A total of 25 corneal epithelium images without the application of eye staining were used. Three image feature extraction approaches were applied for image classification: (i) image feature classification from histogram using a support vector machine with a Gaussian radial basis function (SVM-GRBF); (ii) physical image feature classification using deep-learning Convolutional Neural Networks (CNNs) only, and (iii) the combined classification of CNNs and SVM-Linear. The performance results indicate that our chosen image features from the histogram and length-scale parameter were able to classify with up to 100% accuracy; particularly, at CNNs and CNNs-SVM, by employing 80% of the data sample for training and 20% for testing. Thus, in the assessment of corneal epithelium injuries, HSI has high potential as a method that could surpass current technologies regarding speed, objectivity, and reliability.
LanguageEnglish
Article number2644
Number of pages24
JournalSensors
Volume17
Issue number11
Early online date16 Nov 2017
DOIs
Publication statusPublished - 24 Nov 2017

Fingerprint

Image Enhancement
Corneal Epithelium
image enhancement
epithelium
Image enhancement
learning
Learning
Neural networks
histograms
Image classification
image classification
Support vector machines
Feature extraction
staining
Swine
pattern recognition
Tissue
Staining and Labeling
Technology
education

Keywords

  • corneal epithelium
  • hyperspectral imaging
  • support vector machine
  • convolutional neural networks
  • image enhancement

Cite this

@article{b98d0f4fb8c54e018f55a33122e274ab,
title = "Hyperspectral image enhancement and mixture deep-learning classification of corneal epithelium injury",
abstract = "In our preliminary study, the reflectance signatures obtained from hyperspectral imaging (HSI) of normal and abnormal corneal epithelium tissues of porcine show similar morphology with subtle differences. Here we present image enhancement algorithms that can be used to improve the interpretability of data into clinically relevant information to facilitate diagnostics. A total of 25 corneal epithelium images without the application of eye staining were used. Three image feature extraction approaches were applied for image classification: (i) image feature classification from histogram using a support vector machine with a Gaussian radial basis function (SVM-GRBF); (ii) physical image feature classification using deep-learning Convolutional Neural Networks (CNNs) only, and (iii) the combined classification of CNNs and SVM-Linear. The performance results indicate that our chosen image features from the histogram and length-scale parameter were able to classify with up to 100{\%} accuracy; particularly, at CNNs and CNNs-SVM, by employing 80{\%} of the data sample for training and 20{\%} for testing. Thus, in the assessment of corneal epithelium injuries, HSI has high potential as a method that could surpass current technologies regarding speed, objectivity, and reliability.",
keywords = "corneal epithelium, hyperspectral imaging, support vector machine, convolutional neural networks, image enhancement",
author = "{Md Noor}, {Siti Salwa} and Kaleena Michael and Stephen Marshall and Jinchang Ren",
year = "2017",
month = "11",
day = "24",
doi = "10.3390/s17112644",
language = "English",
volume = "17",
journal = "Sensors",
issn = "1424-8220",
number = "11",

}

Hyperspectral image enhancement and mixture deep-learning classification of corneal epithelium injury. / Md Noor, Siti Salwa; Michael, Kaleena; Marshall, Stephen; Ren, Jinchang.

In: Sensors, Vol. 17, No. 11, 2644, 24.11.2017.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Hyperspectral image enhancement and mixture deep-learning classification of corneal epithelium injury

AU - Md Noor, Siti Salwa

AU - Michael, Kaleena

AU - Marshall, Stephen

AU - Ren, Jinchang

PY - 2017/11/24

Y1 - 2017/11/24

N2 - In our preliminary study, the reflectance signatures obtained from hyperspectral imaging (HSI) of normal and abnormal corneal epithelium tissues of porcine show similar morphology with subtle differences. Here we present image enhancement algorithms that can be used to improve the interpretability of data into clinically relevant information to facilitate diagnostics. A total of 25 corneal epithelium images without the application of eye staining were used. Three image feature extraction approaches were applied for image classification: (i) image feature classification from histogram using a support vector machine with a Gaussian radial basis function (SVM-GRBF); (ii) physical image feature classification using deep-learning Convolutional Neural Networks (CNNs) only, and (iii) the combined classification of CNNs and SVM-Linear. The performance results indicate that our chosen image features from the histogram and length-scale parameter were able to classify with up to 100% accuracy; particularly, at CNNs and CNNs-SVM, by employing 80% of the data sample for training and 20% for testing. Thus, in the assessment of corneal epithelium injuries, HSI has high potential as a method that could surpass current technologies regarding speed, objectivity, and reliability.

AB - In our preliminary study, the reflectance signatures obtained from hyperspectral imaging (HSI) of normal and abnormal corneal epithelium tissues of porcine show similar morphology with subtle differences. Here we present image enhancement algorithms that can be used to improve the interpretability of data into clinically relevant information to facilitate diagnostics. A total of 25 corneal epithelium images without the application of eye staining were used. Three image feature extraction approaches were applied for image classification: (i) image feature classification from histogram using a support vector machine with a Gaussian radial basis function (SVM-GRBF); (ii) physical image feature classification using deep-learning Convolutional Neural Networks (CNNs) only, and (iii) the combined classification of CNNs and SVM-Linear. The performance results indicate that our chosen image features from the histogram and length-scale parameter were able to classify with up to 100% accuracy; particularly, at CNNs and CNNs-SVM, by employing 80% of the data sample for training and 20% for testing. Thus, in the assessment of corneal epithelium injuries, HSI has high potential as a method that could surpass current technologies regarding speed, objectivity, and reliability.

KW - corneal epithelium

KW - hyperspectral imaging

KW - support vector machine

KW - convolutional neural networks

KW - image enhancement

UR - http://www.mdpi.com/1424-8220/17/11/2644

U2 - 10.3390/s17112644

DO - 10.3390/s17112644

M3 - Article

VL - 17

JO - Sensors

T2 - Sensors

JF - Sensors

SN - 1424-8220

IS - 11

M1 - 2644

ER -