Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19
Digital twins is a virtual representation of a device and process that captures the physical properties of the environment and operational algorithms/techniques in the context of medical devices and technology. It may allow and facilitate healthcare organizations to determine ways to improve medical...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
KeAi Communications Co., Ltd.
2022-08-01
|
| Series: | Virtual Reality & Intelligent Hardware |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2096579622000183 |
| _version_ | 1811210646726901760 |
|---|---|
| author | Imran Ahmed Misbah Ahmad Gwanggil Jeon |
| author_facet | Imran Ahmed Misbah Ahmad Gwanggil Jeon |
| author_sort | Imran Ahmed |
| collection | DOAJ |
| description | Digital twins is a virtual representation of a device and process that captures the physical properties of the environment and operational algorithms/techniques in the context of medical devices and technology. It may allow and facilitate healthcare organizations to determine ways to improve medical processes, enhance the patient experience, lower operating expenses, and extend the value of care. Considering the current pandemic situation of COVID-19, various medical devices, e.g., X-rays and CT scan machines and processes, are constantly being used to collect and analyze medical images. In this situation, while collecting and processing an extensive volume of data in the form of images, machines and processes sometimes suffer from system failures that can create critical issues for hospitals and patients. Thus, in this regard, we introduced a digital twin based smart healthcare system integrated with medical devices so that it can be utilized to collect information about the current health condition, configuration, and maintenance history of the device/machine/system. Furthermore, the medical images, i.e., X-rays, are further analyzed by a deep learning model to detect the infection of COVID-19. The designed system is based on Cascade RCNN architecture. In this architecture, detector stages are deeper and are more sequentially selective against close and small false positives. It is a multi stage extension of the Recurrent Convolution Neural Network (RCNN) model and sequentially trained using the output of one stage for the training of the other one. At each stage, the bounding boxes are adjusted in order to locate a suitable value of nearest false positives during training of the different stages. In this way, an arrangement of detectors is adjusted to increase Intersection over Union (IoU) that overcome the problem of overfitting. We trained the model for X-ray images as the model was previously trained on another data set. The developed system achieves good accuracy during the detection phase of the COVID-19. Experimental outcomes reveal the efficiency of the detection architecture, which gains a mean Average Precision (mAP) rate of 0.94. |
| first_indexed | 2024-04-12T04:59:23Z |
| format | Article |
| id | doaj.art-92f5d1c21ba049a1ad5614c2a9e3bd6e |
| institution | Directory Open Access Journal |
| issn | 2096-5796 |
| language | English |
| last_indexed | 2024-04-12T04:59:23Z |
| publishDate | 2022-08-01 |
| publisher | KeAi Communications Co., Ltd. |
| record_format | Article |
| series | Virtual Reality & Intelligent Hardware |
| spelling | doaj.art-92f5d1c21ba049a1ad5614c2a9e3bd6e2022-12-22T03:47:02ZengKeAi Communications Co., Ltd.Virtual Reality & Intelligent Hardware2096-57962022-08-0144292305Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19Imran Ahmed0Misbah Ahmad1Gwanggil Jeon2Center of Excellence in IT, Institute of Management Sciences, Hayatabad, Peshawar, PakistanCenter of Excellence in IT, Institute of Management Sciences, Hayatabad, Peshawar, PakistanDepartment of Embedded Systems Engineering, Incheon National University, Incheon, Korea; Corresponding author.Digital twins is a virtual representation of a device and process that captures the physical properties of the environment and operational algorithms/techniques in the context of medical devices and technology. It may allow and facilitate healthcare organizations to determine ways to improve medical processes, enhance the patient experience, lower operating expenses, and extend the value of care. Considering the current pandemic situation of COVID-19, various medical devices, e.g., X-rays and CT scan machines and processes, are constantly being used to collect and analyze medical images. In this situation, while collecting and processing an extensive volume of data in the form of images, machines and processes sometimes suffer from system failures that can create critical issues for hospitals and patients. Thus, in this regard, we introduced a digital twin based smart healthcare system integrated with medical devices so that it can be utilized to collect information about the current health condition, configuration, and maintenance history of the device/machine/system. Furthermore, the medical images, i.e., X-rays, are further analyzed by a deep learning model to detect the infection of COVID-19. The designed system is based on Cascade RCNN architecture. In this architecture, detector stages are deeper and are more sequentially selective against close and small false positives. It is a multi stage extension of the Recurrent Convolution Neural Network (RCNN) model and sequentially trained using the output of one stage for the training of the other one. At each stage, the bounding boxes are adjusted in order to locate a suitable value of nearest false positives during training of the different stages. In this way, an arrangement of detectors is adjusted to increase Intersection over Union (IoU) that overcome the problem of overfitting. We trained the model for X-ray images as the model was previously trained on another data set. The developed system achieves good accuracy during the detection phase of the COVID-19. Experimental outcomes reveal the efficiency of the detection architecture, which gains a mean Average Precision (mAP) rate of 0.94.http://www.sciencedirect.com/science/article/pii/S2096579622000183Digital TwinsDeep LearningHealthcareCOVID-19Chest X-raysArtificial Intelligence |
| spellingShingle | Imran Ahmed Misbah Ahmad Gwanggil Jeon Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 Virtual Reality & Intelligent Hardware Digital Twins Deep Learning Healthcare COVID-19 Chest X-rays Artificial Intelligence |
| title | Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 |
| title_full | Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 |
| title_fullStr | Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 |
| title_full_unstemmed | Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 |
| title_short | Integrating Digital Twins and Deep Learning for Medical Image Analysis in the era of COVID-19 |
| title_sort | integrating digital twins and deep learning for medical image analysis in the era of covid 19 |
| topic | Digital Twins Deep Learning Healthcare COVID-19 Chest X-rays Artificial Intelligence |
| url | http://www.sciencedirect.com/science/article/pii/S2096579622000183 |
| work_keys_str_mv | AT imranahmed integratingdigitaltwinsanddeeplearningformedicalimageanalysisintheeraofcovid19 AT misbahahmad integratingdigitaltwinsanddeeplearningformedicalimageanalysisintheeraofcovid19 AT gwanggiljeon integratingdigitaltwinsanddeeplearningformedicalimageanalysisintheeraofcovid19 |