A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection

Tuberculosis (TB) is the most serious worldwide infectious disease and the leading cause of death among people with HIV. Early diagnosis and prompt treatment can cut off the rising number of TB deaths, and analysis of chest X-rays is a cost-effective method. We describe a deep learning-based cascade...

Full description

Saved in:
Bibliographic Details
Main Authors: Noppadol Maneerat, Athasart Narkthewan, Kazuhiko Hamamoto
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/13/7300
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849704116685635584
author Noppadol Maneerat
Athasart Narkthewan
Kazuhiko Hamamoto
author_facet Noppadol Maneerat
Athasart Narkthewan
Kazuhiko Hamamoto
author_sort Noppadol Maneerat
collection DOAJ
description Tuberculosis (TB) is the most serious worldwide infectious disease and the leading cause of death among people with HIV. Early diagnosis and prompt treatment can cut off the rising number of TB deaths, and analysis of chest X-rays is a cost-effective method. We describe a deep learning-based cascade algorithm for detecting TB in chest X-rays. Firstly, the lung regions were segregated from other anatomical structures by an encoder–decoder with an atrous separable convolution network—DeepLabv3+ with an XceptionNet backbone, DLabv3+X, and then cropped by a bounding box. Using the cropped lung images, we trained several pre-trained Deep Convolutional Neural Networks (DCNNs) on the images with hyperparameters optimized by a Bayesian algorithm. Different combinations of trained DCNNs were compared, and the combination with the maximum accuracy was retained as the winning combination. The ensemble classifier was designed to predict the presence of TB by fusing DCNNs from the winning combination via weighted averaging. Our lung segmentation was evaluated on three publicly available datasets: it provided better Intercept over Union (IoU) values: 95.1% for Montgomery County (MC), 92.8% for Shenzhen (SZ), and 96.1% for JSRT datasets. For TB prediction, our ensemble classifier produced a better accuracy of 92.7% for the MC dataset and obtained a comparable accuracy of 95.5% for the SZ dataset. Finally, occlusion sensitivity and gradient-weighted class activation maps (Grad-CAM) were generated to indicate the most influential regions for the prediction of TB and to localize TB manifestations.
format Article
id doaj-art-6193c9b2392e4794bc4596afaddeb06e
institution DOAJ
issn 2076-3417
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-6193c9b2392e4794bc4596afaddeb06e2025-08-20T03:16:52ZengMDPI AGApplied Sciences2076-34172025-06-011513730010.3390/app15137300A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis DetectionNoppadol Maneerat0Athasart Narkthewan1Kazuhiko Hamamoto2School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, ThailandSchool of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, ThailandSchool of Information and Telecommunication Engineering, Tokai University, Tokyo 108-8619, JapanTuberculosis (TB) is the most serious worldwide infectious disease and the leading cause of death among people with HIV. Early diagnosis and prompt treatment can cut off the rising number of TB deaths, and analysis of chest X-rays is a cost-effective method. We describe a deep learning-based cascade algorithm for detecting TB in chest X-rays. Firstly, the lung regions were segregated from other anatomical structures by an encoder–decoder with an atrous separable convolution network—DeepLabv3+ with an XceptionNet backbone, DLabv3+X, and then cropped by a bounding box. Using the cropped lung images, we trained several pre-trained Deep Convolutional Neural Networks (DCNNs) on the images with hyperparameters optimized by a Bayesian algorithm. Different combinations of trained DCNNs were compared, and the combination with the maximum accuracy was retained as the winning combination. The ensemble classifier was designed to predict the presence of TB by fusing DCNNs from the winning combination via weighted averaging. Our lung segmentation was evaluated on three publicly available datasets: it provided better Intercept over Union (IoU) values: 95.1% for Montgomery County (MC), 92.8% for Shenzhen (SZ), and 96.1% for JSRT datasets. For TB prediction, our ensemble classifier produced a better accuracy of 92.7% for the MC dataset and obtained a comparable accuracy of 95.5% for the SZ dataset. Finally, occlusion sensitivity and gradient-weighted class activation maps (Grad-CAM) were generated to indicate the most influential regions for the prediction of TB and to localize TB manifestations.https://www.mdpi.com/2076-3417/15/13/7300atrous convolutionBayesian Optimizationchest radiographstuberculosis detectiondeep convolutional neural networksDeepLabv3+
spellingShingle Noppadol Maneerat
Athasart Narkthewan
Kazuhiko Hamamoto
A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
Applied Sciences
atrous convolution
Bayesian Optimization
chest radiographs
tuberculosis detection
deep convolutional neural networks
DeepLabv3+
title A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
title_full A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
title_fullStr A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
title_full_unstemmed A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
title_short A Cascade of Encoder–Decoder with Atrous Convolution and Ensemble Deep Convolutional Neural Networks for Tuberculosis Detection
title_sort cascade of encoder decoder with atrous convolution and ensemble deep convolutional neural networks for tuberculosis detection
topic atrous convolution
Bayesian Optimization
chest radiographs
tuberculosis detection
deep convolutional neural networks
DeepLabv3+
url https://www.mdpi.com/2076-3417/15/13/7300
work_keys_str_mv AT noppadolmaneerat acascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection
AT athasartnarkthewan acascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection
AT kazuhikohamamoto acascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection
AT noppadolmaneerat cascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection
AT athasartnarkthewan cascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection
AT kazuhikohamamoto cascadeofencoderdecoderwithatrousconvolutionandensembledeepconvolutionalneuralnetworksfortuberculosisdetection