Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks

Since the COVID-19 pandemic, interest in non-contact diagnostic technologies has grown, leading to increased research into remote biosignal monitoring. The respiratory rate, widely used in previous studies, offers limited insight into pulmonary volume. To redress this, we propose a thermal imaging-b...

Full description

Saved in:
Bibliographic Details
Main Authors: Do-Kyeong Lee, Jae-Sung Shin, Jae-Sung Choi, Min-Hyung Choi, Min Hong
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Bioengineering
Subjects:
Online Access:https://www.mdpi.com/2306-5354/12/5/542
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850257384956493824
author Do-Kyeong Lee
Jae-Sung Shin
Jae-Sung Choi
Min-Hyung Choi
Min Hong
author_facet Do-Kyeong Lee
Jae-Sung Shin
Jae-Sung Choi
Min-Hyung Choi
Min Hong
author_sort Do-Kyeong Lee
collection DOAJ
description Since the COVID-19 pandemic, interest in non-contact diagnostic technologies has grown, leading to increased research into remote biosignal monitoring. The respiratory rate, widely used in previous studies, offers limited insight into pulmonary volume. To redress this, we propose a thermal imaging-based framework for respiratory segmentation aimed at estimating non-invasive pulmonary function. The proposed method uses an optical flow magnitude-based thresholding technique to automatically extract exhalation frames and segment them into frame sequences. A TransUNet-based network, combining a Convolutional Neural Network (CNN) encoder–decoder architecture with a Transformer module in the bottleneck, is trained on these sequences. The model’s Accuracy, Precision, Recall, IoU, Dice, and F1-score were 0.9832, 0.9833, 0.9830, 0.9651, 0.9822, and 0.9831, respectively, which results demonstrate high segmentation performance. The method enables the respiratory volume to be estimated by detecting exhalation behavior, suggesting its potential as a non-contact tool to monitor pulmonary function and estimate lung volume. Furthermore, research on thermal imaging-based respiratory volume analysis remains limited. This study expands upon conventional respiratory rate-based approaches to provide a new direction for respiratory analysis using vision-based techniques.
format Article
id doaj-art-220d2ef890c94a3f82cf75fe22f54fb2
institution OA Journals
issn 2306-5354
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Bioengineering
spelling doaj-art-220d2ef890c94a3f82cf75fe22f54fb22025-08-20T01:56:25ZengMDPI AGBioengineering2306-53542025-05-0112554210.3390/bioengineering12050542Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep NetworksDo-Kyeong Lee0Jae-Sung Shin1Jae-Sung Choi2Min-Hyung Choi3Min Hong4Department of Software Convergence, Soonchunhyang University, Asan 31538, Republic of KoreaDepartment of Software Convergence, Soonchunhyang University, Asan 31538, Republic of KoreaDepartment of Internal Medicine, Cheonan Hospital, College of Medicine, Soonchunhyang University, Cheonan 31151, Republic of KoreaDepartment of Computer Science, Saint Louis University, Louis, MO 63103, USADepartment of Computer Software Engineering, Soonchunhyang University, Asan 31538, Republic of KoreaSince the COVID-19 pandemic, interest in non-contact diagnostic technologies has grown, leading to increased research into remote biosignal monitoring. The respiratory rate, widely used in previous studies, offers limited insight into pulmonary volume. To redress this, we propose a thermal imaging-based framework for respiratory segmentation aimed at estimating non-invasive pulmonary function. The proposed method uses an optical flow magnitude-based thresholding technique to automatically extract exhalation frames and segment them into frame sequences. A TransUNet-based network, combining a Convolutional Neural Network (CNN) encoder–decoder architecture with a Transformer module in the bottleneck, is trained on these sequences. The model’s Accuracy, Precision, Recall, IoU, Dice, and F1-score were 0.9832, 0.9833, 0.9830, 0.9651, 0.9822, and 0.9831, respectively, which results demonstrate high segmentation performance. The method enables the respiratory volume to be estimated by detecting exhalation behavior, suggesting its potential as a non-contact tool to monitor pulmonary function and estimate lung volume. Furthermore, research on thermal imaging-based respiratory volume analysis remains limited. This study expands upon conventional respiratory rate-based approaches to provide a new direction for respiratory analysis using vision-based techniques.https://www.mdpi.com/2306-5354/12/5/542thermal imagingnon-contact pulmonary monitoringU-Net-based segmentationpulmonary diagnostics
spellingShingle Do-Kyeong Lee
Jae-Sung Shin
Jae-Sung Choi
Min-Hyung Choi
Min Hong
Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
Bioengineering
thermal imaging
non-contact pulmonary monitoring
U-Net-based segmentation
pulmonary diagnostics
title Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
title_full Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
title_fullStr Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
title_full_unstemmed Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
title_short Exhale-Focused Thermal Image Segmentation Using Optical Flow-Based Frame Filtering and Transformer-Aided Deep Networks
title_sort exhale focused thermal image segmentation using optical flow based frame filtering and transformer aided deep networks
topic thermal imaging
non-contact pulmonary monitoring
U-Net-based segmentation
pulmonary diagnostics
url https://www.mdpi.com/2306-5354/12/5/542
work_keys_str_mv AT dokyeonglee exhalefocusedthermalimagesegmentationusingopticalflowbasedframefilteringandtransformeraideddeepnetworks
AT jaesungshin exhalefocusedthermalimagesegmentationusingopticalflowbasedframefilteringandtransformeraideddeepnetworks
AT jaesungchoi exhalefocusedthermalimagesegmentationusingopticalflowbasedframefilteringandtransformeraideddeepnetworks
AT minhyungchoi exhalefocusedthermalimagesegmentationusingopticalflowbasedframefilteringandtransformeraideddeepnetworks
AT minhong exhalefocusedthermalimagesegmentationusingopticalflowbasedframefilteringandtransformeraideddeepnetworks