An effective and accurate identification system of Mycobacterium tuberculosis using convolution neural networks

Chan Pang Kuok, Ming Huwi Horng, Yu Ming Liao, Nan-Haw Chow, Yung-Nien Sun

Research output: Contribution to journalArticle

Abstract

Tuberculosis (TB) remains the leading cause of morbidity and mortality from infectious disease in developing countries. The sputum smear microscopy remains the primary diagnostic laboratory test. However, microscopic examination is always time-consuming and tedious. Therefore, an effective computer-aided image identification system is needed to provide timely assistance in diagnosis. The current identification system usually suffers from complex color variations of the images, resulting in plentiful of false object detection. To overcome the dilemma, we propose a two-stage Mycobacterium tuberculosis identification system, consisting of candidate detection and classification using convolution neural networks (CNNs). The refined Faster region-based CNN was used to distinguish candidates of M. tuberculosis and the actual ones were classified by utilizing CNN-based classifier. We first compared three different CNNs, including ensemble CNN, single-member CNN, and deep CNN. The experimental results showed that both ensemble and deep CNNs were on par with similar identification performance when analyzing more than 19,000 images. A much better recall value was achieved by using our proposed system in comparison with conventional pixel-based support vector machine method for M. tuberculosis bacilli detection.

Original languageEnglish
Pages (from-to)709-719
Number of pages11
JournalMicroscopy Research and Technique
Volume82
Issue number6
DOIs
Publication statusPublished - 2019 Jun 1

Fingerprint

tuberculosis
system identification
Convolution
Mycobacterium tuberculosis
convolution integrals
Identification (control systems)
Neural networks
Sputum
Routine Diagnostic Tests
Bacillus
Developing Countries
Communicable Diseases
Microscopy
Tuberculosis
Color
Morbidity
Microscopic examination
Mortality
smear
mortality

All Science Journal Classification (ASJC) codes

  • Anatomy
  • Histology
  • Instrumentation
  • Medical Laboratory Technology

Cite this

@article{8863fd353d6f41829eb6edadee2d377e,
title = "An effective and accurate identification system of Mycobacterium tuberculosis using convolution neural networks",
abstract = "Tuberculosis (TB) remains the leading cause of morbidity and mortality from infectious disease in developing countries. The sputum smear microscopy remains the primary diagnostic laboratory test. However, microscopic examination is always time-consuming and tedious. Therefore, an effective computer-aided image identification system is needed to provide timely assistance in diagnosis. The current identification system usually suffers from complex color variations of the images, resulting in plentiful of false object detection. To overcome the dilemma, we propose a two-stage Mycobacterium tuberculosis identification system, consisting of candidate detection and classification using convolution neural networks (CNNs). The refined Faster region-based CNN was used to distinguish candidates of M. tuberculosis and the actual ones were classified by utilizing CNN-based classifier. We first compared three different CNNs, including ensemble CNN, single-member CNN, and deep CNN. The experimental results showed that both ensemble and deep CNNs were on par with similar identification performance when analyzing more than 19,000 images. A much better recall value was achieved by using our proposed system in comparison with conventional pixel-based support vector machine method for M. tuberculosis bacilli detection.",
author = "Kuok, {Chan Pang} and Horng, {Ming Huwi} and Liao, {Yu Ming} and Nan-Haw Chow and Yung-Nien Sun",
year = "2019",
month = "6",
day = "1",
doi = "10.1002/jemt.23217",
language = "English",
volume = "82",
pages = "709--719",
journal = "Microscopy Research and Technique",
issn = "1059-910X",
publisher = "Wiley-Liss Inc.",
number = "6",

}

An effective and accurate identification system of Mycobacterium tuberculosis using convolution neural networks. / Kuok, Chan Pang; Horng, Ming Huwi; Liao, Yu Ming; Chow, Nan-Haw; Sun, Yung-Nien.

In: Microscopy Research and Technique, Vol. 82, No. 6, 01.06.2019, p. 709-719.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An effective and accurate identification system of Mycobacterium tuberculosis using convolution neural networks

AU - Kuok, Chan Pang

AU - Horng, Ming Huwi

AU - Liao, Yu Ming

AU - Chow, Nan-Haw

AU - Sun, Yung-Nien

PY - 2019/6/1

Y1 - 2019/6/1

N2 - Tuberculosis (TB) remains the leading cause of morbidity and mortality from infectious disease in developing countries. The sputum smear microscopy remains the primary diagnostic laboratory test. However, microscopic examination is always time-consuming and tedious. Therefore, an effective computer-aided image identification system is needed to provide timely assistance in diagnosis. The current identification system usually suffers from complex color variations of the images, resulting in plentiful of false object detection. To overcome the dilemma, we propose a two-stage Mycobacterium tuberculosis identification system, consisting of candidate detection and classification using convolution neural networks (CNNs). The refined Faster region-based CNN was used to distinguish candidates of M. tuberculosis and the actual ones were classified by utilizing CNN-based classifier. We first compared three different CNNs, including ensemble CNN, single-member CNN, and deep CNN. The experimental results showed that both ensemble and deep CNNs were on par with similar identification performance when analyzing more than 19,000 images. A much better recall value was achieved by using our proposed system in comparison with conventional pixel-based support vector machine method for M. tuberculosis bacilli detection.

AB - Tuberculosis (TB) remains the leading cause of morbidity and mortality from infectious disease in developing countries. The sputum smear microscopy remains the primary diagnostic laboratory test. However, microscopic examination is always time-consuming and tedious. Therefore, an effective computer-aided image identification system is needed to provide timely assistance in diagnosis. The current identification system usually suffers from complex color variations of the images, resulting in plentiful of false object detection. To overcome the dilemma, we propose a two-stage Mycobacterium tuberculosis identification system, consisting of candidate detection and classification using convolution neural networks (CNNs). The refined Faster region-based CNN was used to distinguish candidates of M. tuberculosis and the actual ones were classified by utilizing CNN-based classifier. We first compared three different CNNs, including ensemble CNN, single-member CNN, and deep CNN. The experimental results showed that both ensemble and deep CNNs were on par with similar identification performance when analyzing more than 19,000 images. A much better recall value was achieved by using our proposed system in comparison with conventional pixel-based support vector machine method for M. tuberculosis bacilli detection.

UR - http://www.scopus.com/inward/record.url?scp=85061434767&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061434767&partnerID=8YFLogxK

U2 - 10.1002/jemt.23217

DO - 10.1002/jemt.23217

M3 - Article

VL - 82

SP - 709

EP - 719

JO - Microscopy Research and Technique

JF - Microscopy Research and Technique

SN - 1059-910X

IS - 6

ER -