Automated detection of Mycobacterium tuberculosis using transfer learning

Abstract

Introduction: Quantitative analysis of Mycobacterium tuberculosis using microscope is very critical for diagnosing tuberculosis diseases. Microbiologist encounter several challenges which can lead to misdiagnosis. However, there are 3 main challenges: (1) The size of Mycobacterium tuberculosis is very small and difficult to identify as a result of low contrast background, heterogenous shape, irregular appearance and faint boundaries (2) Mycobacterium tuberculosis overlapped with each other making it difficult to conduct accurate diagnosis (3) Large amount of slide can be time consuming and tedious to microbiologist and which can lead to misinterpretations.
 Methodology: To solve these challenges and limitations, we proposed an automated-based detection method using pretrained AlexNet to trained the model in 3 sets of experiments A, B and C and adjust the protocols accordingly. We compared the detection of tuberculosis using AlexNet Models with the ground truth result provided by microbiologist and analyzed inconsistencies between network models and human.
 Results: 98.15 % accuracy, 96.77% sensitivity and 100% specificity for experiment A, 98.09% accuracy, 98.59% sensitivity and 97.67% specificity for experiment B and 98.73% testing accuracy, 98.59 sensitivity, 98.84% specificity ofr experiment C which sound robust and promising.
 Conclusions: The results indicated that network performance was successful with high accuracies, sensitivities and specificities and it can be used to support microbiologist for diagnosis of tuberculosis.

Harnessing of Artificial Intelligence for the Diagnosis and Prevention of Hospital-Acquired Infections: A Systematic Review

“…Therefore, computer-aided identification systems represent a promising approach for timely and reproducible results. Candidate detection and classification using CNNs have been proposed for the effective and accurate detection and identification of M. tuberculosis in various studies [ 56 , 57 , 58 , 59 , 60 ]. Kuok et al have utilized a Refined Faster region-based CNN (Faster R-CCN) model for the automated detection of acid-fast bacilli (AFB) detection in smear sputum slides and reported an 86% detection rate of the Faster R-CCN model compared to support vector machine (SVM), which demonstrated a 70.93% overall detection rate [ 56 ].…”

The Impact of Artificial Intelligence on Microbial Diagnosis

“…In response to these challenges, there have been endeavors to automate AFB microscopic diagnosis using digital imaging and computer vision techniques. While cytology sputum smears were the primary focus in studies involving AI, a significant portion also explored CNNs on histological specimens like lung biopsies [118,119]. Image algorithms for AFB recognition have significantly improved, especially in sputum samples [120,121].…”

A Study of Mycobacterium tuberculosis Detection Using Different Neural Networks in Autopsy Specimens

“…They then performed transfer learning using AlexNet. As a result, they achieved accuracies of 98.15%, 98.09%, and 98.73% in experiments A, B, and C, respectively, demonstrating an analytical ability comparable to that of a pathologist [ 17 ]. Jeannette Chang et al used 390 fluorescence microscopy slide images (92 positive, 298 negative) and achieved accuracy of 89.2% using a support vector machine (SVM) as a classifier after Otsu binarization.…”