Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models
Introduction: Automatic systems based on Artificial intelligence (AI) algorithms have made significant advancements across various domains, most notably in the field of medicine. This study introduces a novel approach for identifying COVID-19-infected regions in lung computed tomography (CT) scan t...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Shahid Beheshti University of Medical Sciences
2024-12-01
|
| Series: | Archives of Academic Emergency Medicine |
| Subjects: | |
| Online Access: | https://journals.sbmu.ac.ir/aaem/index.php/AAEM/article/view/2515 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850252601155649536 |
|---|---|
| author | Zeinab Momeni pour Ali Asghar Beheshti Shirazi |
| author_facet | Zeinab Momeni pour Ali Asghar Beheshti Shirazi |
| author_sort | Zeinab Momeni pour |
| collection | DOAJ |
| description |
Introduction: Automatic systems based on Artificial intelligence (AI) algorithms have made significant advancements across various domains, most notably in the field of medicine. This study introduces a novel approach for identifying COVID-19-infected regions in lung computed tomography (CT) scan through the development of two innovative models.
Methods: In this study we used the Squeeze and Excitation based UNet TRansformers (SE-UNETR) and the Squeeze and Excitation based High-Quality Resolution Swin Transformer Network (SE-HQRSTNet), to develop two three-dimensional segmentation networks for identifying COVID-19-infected regions in lung CT scan. The SE-UNETR model is structured as a 3D UNet architecture with an encoder component built on Vision Transformers (ViTs). This model processes 3D patches directly as input and learns sequential representations of the volumetric data. The encoder connects to the decoder using skip connections, ultimately producing the final semantic segmentation output. Conversely, the SE-HQRSTNet model incorporates High-Resolution Networks (HRNet), Swin Transformer modules, and Squeeze and Excitation (SE) blocks. This architecture is designed to generate features at multiple resolutions, utilizing Multi-Resolution Feature Fusion (MRFF) blocks to effectively integrate semantic features across various scales. The proposed networks were evaluated using a 5-fold cross-validation methodology, along with data augmentation techniques, applied to the COVID-19-CT-Seg and MosMed datasets.
Results: Our experimental results demonstrate that the Dice value for the infection masks within the COVID-19-CT-Seg dataset improved by 3.81% and 4.84% with the SE-UNETR and SE-HQRSTNet models, respectively, compared to previously reported work. Furthermore, the Dice value for the MosMed dataset increased from 66.8% to 69.35% and 70.89% for the SE-UNETR and SE-HQRSTNet models, respectively.
Conclusion: These improvements indicate that the proposed models exhibit superior efficiency and performance relative to existing methodologies.
|
| format | Article |
| id | doaj-art-e478fa340c8946efbd3f8eb2fc288925 |
| institution | OA Journals |
| issn | 2645-4904 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | Shahid Beheshti University of Medical Sciences |
| record_format | Article |
| series | Archives of Academic Emergency Medicine |
| spelling | doaj-art-e478fa340c8946efbd3f8eb2fc2889252025-08-20T01:57:35ZengShahid Beheshti University of Medical SciencesArchives of Academic Emergency Medicine2645-49042024-12-0113110.22037/aaemj.v13i1.2515Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer ModelsZeinab Momeni pour0Ali Asghar Beheshti Shirazi1Department of Electrical Engineering, Iran University of Science and Technology, Tehran, IranDepartment of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran Introduction: Automatic systems based on Artificial intelligence (AI) algorithms have made significant advancements across various domains, most notably in the field of medicine. This study introduces a novel approach for identifying COVID-19-infected regions in lung computed tomography (CT) scan through the development of two innovative models. Methods: In this study we used the Squeeze and Excitation based UNet TRansformers (SE-UNETR) and the Squeeze and Excitation based High-Quality Resolution Swin Transformer Network (SE-HQRSTNet), to develop two three-dimensional segmentation networks for identifying COVID-19-infected regions in lung CT scan. The SE-UNETR model is structured as a 3D UNet architecture with an encoder component built on Vision Transformers (ViTs). This model processes 3D patches directly as input and learns sequential representations of the volumetric data. The encoder connects to the decoder using skip connections, ultimately producing the final semantic segmentation output. Conversely, the SE-HQRSTNet model incorporates High-Resolution Networks (HRNet), Swin Transformer modules, and Squeeze and Excitation (SE) blocks. This architecture is designed to generate features at multiple resolutions, utilizing Multi-Resolution Feature Fusion (MRFF) blocks to effectively integrate semantic features across various scales. The proposed networks were evaluated using a 5-fold cross-validation methodology, along with data augmentation techniques, applied to the COVID-19-CT-Seg and MosMed datasets. Results: Our experimental results demonstrate that the Dice value for the infection masks within the COVID-19-CT-Seg dataset improved by 3.81% and 4.84% with the SE-UNETR and SE-HQRSTNet models, respectively, compared to previously reported work. Furthermore, the Dice value for the MosMed dataset increased from 66.8% to 69.35% and 70.89% for the SE-UNETR and SE-HQRSTNet models, respectively. Conclusion: These improvements indicate that the proposed models exhibit superior efficiency and performance relative to existing methodologies. https://journals.sbmu.ac.ir/aaem/index.php/AAEM/article/view/2515COVID-19SegmentationSelf-AttentionSqueeze and ExcitationSwin TransformerVision Transformer |
| spellingShingle | Zeinab Momeni pour Ali Asghar Beheshti Shirazi Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models Archives of Academic Emergency Medicine COVID-19 Segmentation Self-Attention Squeeze and Excitation Swin Transformer Vision Transformer |
| title | Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models |
| title_full | Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models |
| title_fullStr | Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models |
| title_full_unstemmed | Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models |
| title_short | Identifying COVID-19-Infected Segments in Lung CT Scan Through Two Innovative Artificial Intelligence-Based Transformer Models |
| title_sort | identifying covid 19 infected segments in lung ct scan through two innovative artificial intelligence based transformer models |
| topic | COVID-19 Segmentation Self-Attention Squeeze and Excitation Swin Transformer Vision Transformer |
| url | https://journals.sbmu.ac.ir/aaem/index.php/AAEM/article/view/2515 |
| work_keys_str_mv | AT zeinabmomenipour identifyingcovid19infectedsegmentsinlungctscanthroughtwoinnovativeartificialintelligencebasedtransformermodels AT aliasgharbeheshtishirazi identifyingcovid19infectedsegmentsinlungctscanthroughtwoinnovativeartificialintelligencebasedtransformermodels |