Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach

IntroductionAn automated computerized approach can aid radiologists in the early diagnosis of lung disease from video modalities. This study focuses on the difficulties associated with identifying and categorizing respiratory diseases, including COVID-19, influenza, and pneumonia.MethodsWe propose a...

Full description

Saved in:
Bibliographic Details
Main Authors: Arefin Ittesafun Abian, Mohaimenul Azam Khan Raiaan, Asif Karim, Sami Azam, Nur Mohammad Fahad, Niusha Shafiabady, Kheng Cher Yeo, Friso De Boer
Format: Article
Language:English
Published: Frontiers Media S.A. 2024-12-01
Series:Frontiers in Computer Science
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fcomp.2024.1438126/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850248272004775936
author Arefin Ittesafun Abian
Mohaimenul Azam Khan Raiaan
Asif Karim
Sami Azam
Nur Mohammad Fahad
Niusha Shafiabady
Kheng Cher Yeo
Friso De Boer
author_facet Arefin Ittesafun Abian
Mohaimenul Azam Khan Raiaan
Asif Karim
Sami Azam
Nur Mohammad Fahad
Niusha Shafiabady
Kheng Cher Yeo
Friso De Boer
author_sort Arefin Ittesafun Abian
collection DOAJ
description IntroductionAn automated computerized approach can aid radiologists in the early diagnosis of lung disease from video modalities. This study focuses on the difficulties associated with identifying and categorizing respiratory diseases, including COVID-19, influenza, and pneumonia.MethodsWe propose a novel method that combines three dimensional (3D) models, model explainability (XAI), and a Decision Support System (DSS) that utilizes lung ultrasound (LUS) videos. The objective of the study is to improve the quality of video frames, boost the diversity of the dataset, maintain the sequence of frames, and create a hybrid 3D model [Three-Dimensional Time Distributed Convolutional Neural Network-Long short-term memory (TD-CNNLSTM-LungNet)] for precise classification. The proposed methodology involves applying morphological opening and contour detection to improve frame quality, utilizing geometrical augmentation for dataset balance, introducing a graph-based approach for frame sequencing, and implementing a hybrid 3D model combining time-distributed CNN and LSTM networks utilizing vast ablation study. Model explainability is ensured through heatmap generation, region of interest segmentation, and Probability Density Function (PDF) graphs illustrating feature distribution.ResultsOur model TD-CNN-LSTM-LungNet attained a remarkable accuracy of 96.57% in classifying LUS videos into pneumonia, COVID-19, normal, and other lung disease classes, which is above compared to ten traditional transfer learning models experimented with in this study. The eleven-ablation case study reduced training costs and redundancy. K-fold cross-validation and accuracy-loss curves demonstrated model generalization. The DSS, incorporating Layer Class Activation Mapping (LayerCAM) and heatmaps, improved interpretability and reliability, and PDF graphs facilitated precise decision-making by identifying feature boundaries. The DSS facilitates clinical marker analysis, and the validation by using the proposed algorithms highlights its impact on a reliable diagnosis outcome.DiscussionOur proposed methodology could assist radiologists in accurately detecting and comprehending the patterns of respiratory disorders.
format Article
id doaj-art-a97a1449750e446b862b0640c4e145d5
institution OA Journals
issn 2624-9898
language English
publishDate 2024-12-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Computer Science
spelling doaj-art-a97a1449750e446b862b0640c4e145d52025-08-20T01:58:45ZengFrontiers Media S.A.Frontiers in Computer Science2624-98982024-12-01610.3389/fcomp.2024.14381261438126Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approachArefin Ittesafun Abian0Mohaimenul Azam Khan Raiaan1Asif Karim2Sami Azam3Nur Mohammad Fahad4Niusha Shafiabady5Kheng Cher Yeo6Friso De Boer7Department of Computer Science and Engineering, United International University, Dhaka, BangladeshDepartment of Computer Science and Engineering, United International University, Dhaka, BangladeshFaculty of Science and Technology, Charles Darwin University, Darwin, NT, AustraliaFaculty of Science and Technology, Charles Darwin University, Darwin, NT, AustraliaDepartment of Computer Science and Engineering, United International University, Dhaka, BangladeshDepartment of Information Technology, Australian Catholic University, North Sydney, NSW, AustraliaFaculty of Science and Technology, Charles Darwin University, Darwin, NT, AustraliaFaculty of Science and Technology, Charles Darwin University, Darwin, NT, AustraliaIntroductionAn automated computerized approach can aid radiologists in the early diagnosis of lung disease from video modalities. This study focuses on the difficulties associated with identifying and categorizing respiratory diseases, including COVID-19, influenza, and pneumonia.MethodsWe propose a novel method that combines three dimensional (3D) models, model explainability (XAI), and a Decision Support System (DSS) that utilizes lung ultrasound (LUS) videos. The objective of the study is to improve the quality of video frames, boost the diversity of the dataset, maintain the sequence of frames, and create a hybrid 3D model [Three-Dimensional Time Distributed Convolutional Neural Network-Long short-term memory (TD-CNNLSTM-LungNet)] for precise classification. The proposed methodology involves applying morphological opening and contour detection to improve frame quality, utilizing geometrical augmentation for dataset balance, introducing a graph-based approach for frame sequencing, and implementing a hybrid 3D model combining time-distributed CNN and LSTM networks utilizing vast ablation study. Model explainability is ensured through heatmap generation, region of interest segmentation, and Probability Density Function (PDF) graphs illustrating feature distribution.ResultsOur model TD-CNN-LSTM-LungNet attained a remarkable accuracy of 96.57% in classifying LUS videos into pneumonia, COVID-19, normal, and other lung disease classes, which is above compared to ten traditional transfer learning models experimented with in this study. The eleven-ablation case study reduced training costs and redundancy. K-fold cross-validation and accuracy-loss curves demonstrated model generalization. The DSS, incorporating Layer Class Activation Mapping (LayerCAM) and heatmaps, improved interpretability and reliability, and PDF graphs facilitated precise decision-making by identifying feature boundaries. The DSS facilitates clinical marker analysis, and the validation by using the proposed algorithms highlights its impact on a reliable diagnosis outcome.DiscussionOur proposed methodology could assist radiologists in accurately detecting and comprehending the patterns of respiratory disorders.https://www.frontiersin.org/articles/10.3389/fcomp.2024.1438126/fulllung ultrasoundCOVID-19LayerCAMdecision support systemLSTMCNN
spellingShingle Arefin Ittesafun Abian
Mohaimenul Azam Khan Raiaan
Asif Karim
Sami Azam
Nur Mohammad Fahad
Niusha Shafiabady
Kheng Cher Yeo
Friso De Boer
Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
Frontiers in Computer Science
lung ultrasound
COVID-19
LayerCAM
decision support system
LSTM
CNN
title Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
title_full Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
title_fullStr Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
title_full_unstemmed Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
title_short Automated diagnosis of respiratory diseases from lung ultrasound videos ensuring XAI: an innovative hybrid model approach
title_sort automated diagnosis of respiratory diseases from lung ultrasound videos ensuring xai an innovative hybrid model approach
topic lung ultrasound
COVID-19
LayerCAM
decision support system
LSTM
CNN
url https://www.frontiersin.org/articles/10.3389/fcomp.2024.1438126/full
work_keys_str_mv AT arefinittesafunabian automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT mohaimenulazamkhanraiaan automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT asifkarim automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT samiazam automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT nurmohammadfahad automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT niushashafiabady automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT khengcheryeo automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach
AT frisodeboer automateddiagnosisofrespiratorydiseasesfromlungultrasoundvideosensuringxaianinnovativehybridmodelapproach