Deep learning representations to support COVID-19 diagnosis on CT slices
Introduction: The coronavirus disease 2019 (COVID-19) has become a significant public health problem worldwide. In this context, CT-scan automatic analysis has emerged as a COVID-19 complementary diagnosis tool allowing for radiological finding characterization, patient categorization, and disease f...
| Main Authors: | , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Instituto Nacional de Salud
2022-03-01
|
| Series: | Biomédica: revista del Instituto Nacional de Salud |
| Subjects: | |
| Online Access: | https://revistabiomedica.org/index.php/biomedica/article/view/5927 |
| _version_ | 1828522503792754688 |
|---|---|
| author | Josué Ruano John Arcila David Romo-Bucheli Carlos Vargas Jefferson Rodríguez Óscar Mendoza Miguel Plazas Lola Bautista Jorge Villamizar Gabriel Pedraza Alejandra Moreno Diana Valenzuela Lina Vázquez Carolina Valenzuela-Santos Paul Camacho Daniel Mantilla Fabio Martínez Carrillo |
| author_facet | Josué Ruano John Arcila David Romo-Bucheli Carlos Vargas Jefferson Rodríguez Óscar Mendoza Miguel Plazas Lola Bautista Jorge Villamizar Gabriel Pedraza Alejandra Moreno Diana Valenzuela Lina Vázquez Carolina Valenzuela-Santos Paul Camacho Daniel Mantilla Fabio Martínez Carrillo |
| author_sort | Josué Ruano |
| collection | DOAJ |
| description | Introduction: The coronavirus disease 2019 (COVID-19) has become a significant public health problem worldwide. In this context, CT-scan automatic analysis has emerged as a COVID-19 complementary diagnosis tool allowing for radiological finding characterization, patient categorization, and disease follow-up. However, this analysis depends on the radiologist’s expertise, which may result in subjective evaluations.
Objective: To explore deep learning representations, trained from thoracic CT-slices, to automatically distinguish COVID-19 disease from control samples.
Materials and methods: Two datasets were used: SARS-CoV-2 CT Scan (Set-1) and FOSCAL clinic’s dataset (Set-2). The deep representations took advantage of supervised learning models previously trained on the natural image domain, which were adjusted following a transfer learning scheme. The deep classification was carried out: (a) via an end-to-end deep learning approach and (b) via random forest and support vector machine classifiers by feeding the deep representation embedding vectors into these classifiers.
Results: The end-to-end classification achieved an average accuracy of 92.33% (89.70% precision) for Set-1 and 96.99% (96.62% precision) for Set-2. The deep feature embedding with a support vector machine achieved an average accuracy of 91.40% (95.77% precision) and 96.00% (94.74% precision) for Set-1 and Set-2, respectively.
Conclusion: Deep representations have achieved outstanding performance in the identification of COVID-19 cases on CT scans demonstrating good characterization of the COVID-19 radiological patterns. These representations could potentially support the COVID-19 diagnosis in clinical settings. |
| first_indexed | 2024-12-11T20:09:01Z |
| format | Article |
| id | doaj.art-21b384a2a1f141549dc6e006e65e12a2 |
| institution | Directory Open Access Journal |
| issn | 0120-4157 |
| language | English |
| last_indexed | 2024-12-11T20:09:01Z |
| publishDate | 2022-03-01 |
| publisher | Instituto Nacional de Salud |
| record_format | Article |
| series | Biomédica: revista del Instituto Nacional de Salud |
| spelling | doaj.art-21b384a2a1f141549dc6e006e65e12a22022-12-22T00:52:19ZengInstituto Nacional de SaludBiomédica: revista del Instituto Nacional de Salud0120-41572022-03-0142117018310.7705/biomedica.59275927Deep learning representations to support COVID-19 diagnosis on CT slicesJosué Ruano0John Arcila1David Romo-Bucheli2Carlos Vargas3Jefferson Rodríguez4Óscar Mendoza5Miguel Plazas6Lola Bautista7Jorge Villamizar8Gabriel Pedraza9Alejandra Moreno10Diana Valenzuela11Lina Vázquez12Carolina Valenzuela-Santos13Paul Camacho14Daniel Mantilla15Fabio Martínez Carrillo16BIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, Colombia; Facultad de Ingeniería, Universidad de Los Andes, Mérida, VenezuelaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaClínica FOSCAL, Fundación Oftalmológica de Santander, Bucaramanga, ColombiaClínica FOSCAL, Fundación Oftalmológica de Santander, Bucaramanga, ColombiaClínica FOSCAL, Fundación Oftalmológica de Santander, Bucaramanga, ColombiaClínica FOSCAL, Fundación Oftalmológica de Santander, Bucaramanga, ColombiaClínica FOSCAL, Fundación Oftalmológica de Santander, Bucaramanga, ColombiaBIVL2ab Biomedical Imaging, Vision and Learning Laboratory, Escuela de Ingeniería de Sistemas e Informática, Universidad Industrial de Santander, Bucaramanga, ColombiaIntroduction: The coronavirus disease 2019 (COVID-19) has become a significant public health problem worldwide. In this context, CT-scan automatic analysis has emerged as a COVID-19 complementary diagnosis tool allowing for radiological finding characterization, patient categorization, and disease follow-up. However, this analysis depends on the radiologist’s expertise, which may result in subjective evaluations. Objective: To explore deep learning representations, trained from thoracic CT-slices, to automatically distinguish COVID-19 disease from control samples. Materials and methods: Two datasets were used: SARS-CoV-2 CT Scan (Set-1) and FOSCAL clinic’s dataset (Set-2). The deep representations took advantage of supervised learning models previously trained on the natural image domain, which were adjusted following a transfer learning scheme. The deep classification was carried out: (a) via an end-to-end deep learning approach and (b) via random forest and support vector machine classifiers by feeding the deep representation embedding vectors into these classifiers. Results: The end-to-end classification achieved an average accuracy of 92.33% (89.70% precision) for Set-1 and 96.99% (96.62% precision) for Set-2. The deep feature embedding with a support vector machine achieved an average accuracy of 91.40% (95.77% precision) and 96.00% (94.74% precision) for Set-1 and Set-2, respectively. Conclusion: Deep representations have achieved outstanding performance in the identification of COVID-19 cases on CT scans demonstrating good characterization of the COVID-19 radiological patterns. These representations could potentially support the COVID-19 diagnosis in clinical settings.https://revistabiomedica.org/index.php/biomedica/article/view/5927coronavirus infections/diagnosistomography, x-ray computeddeep learning |
| spellingShingle | Josué Ruano John Arcila David Romo-Bucheli Carlos Vargas Jefferson Rodríguez Óscar Mendoza Miguel Plazas Lola Bautista Jorge Villamizar Gabriel Pedraza Alejandra Moreno Diana Valenzuela Lina Vázquez Carolina Valenzuela-Santos Paul Camacho Daniel Mantilla Fabio Martínez Carrillo Deep learning representations to support COVID-19 diagnosis on CT slices Biomédica: revista del Instituto Nacional de Salud coronavirus infections/diagnosis tomography, x-ray computed deep learning |
| title | Deep learning representations to support COVID-19 diagnosis on CT slices |
| title_full | Deep learning representations to support COVID-19 diagnosis on CT slices |
| title_fullStr | Deep learning representations to support COVID-19 diagnosis on CT slices |
| title_full_unstemmed | Deep learning representations to support COVID-19 diagnosis on CT slices |
| title_short | Deep learning representations to support COVID-19 diagnosis on CT slices |
| title_sort | deep learning representations to support covid 19 diagnosis on ct slices |
| topic | coronavirus infections/diagnosis tomography, x-ray computed deep learning |
| url | https://revistabiomedica.org/index.php/biomedica/article/view/5927 |
| work_keys_str_mv | AT josueruano deeplearningrepresentationstosupportcovid19diagnosisonctslices AT johnarcila deeplearningrepresentationstosupportcovid19diagnosisonctslices AT davidromobucheli deeplearningrepresentationstosupportcovid19diagnosisonctslices AT carlosvargas deeplearningrepresentationstosupportcovid19diagnosisonctslices AT jeffersonrodriguez deeplearningrepresentationstosupportcovid19diagnosisonctslices AT oscarmendoza deeplearningrepresentationstosupportcovid19diagnosisonctslices AT miguelplazas deeplearningrepresentationstosupportcovid19diagnosisonctslices AT lolabautista deeplearningrepresentationstosupportcovid19diagnosisonctslices AT jorgevillamizar deeplearningrepresentationstosupportcovid19diagnosisonctslices AT gabrielpedraza deeplearningrepresentationstosupportcovid19diagnosisonctslices AT alejandramoreno deeplearningrepresentationstosupportcovid19diagnosisonctslices AT dianavalenzuela deeplearningrepresentationstosupportcovid19diagnosisonctslices AT linavazquez deeplearningrepresentationstosupportcovid19diagnosisonctslices AT carolinavalenzuelasantos deeplearningrepresentationstosupportcovid19diagnosisonctslices AT paulcamacho deeplearningrepresentationstosupportcovid19diagnosisonctslices AT danielmantilla deeplearningrepresentationstosupportcovid19diagnosisonctslices AT fabiomartinezcarrillo deeplearningrepresentationstosupportcovid19diagnosisonctslices |