Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility
X-ray technology has been recently employed for the detection of the lethal human coronavirus disease 2019 (COVID-19) as a timely, cheap, and helpful ancillary method for diagnosis. The scientific community evaluated deep learning methods to aid in the automatic detection of the disease, utilizing p...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Reports |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2571-841X/5/2/20 |
| _version_ | 1827656897217429504 |
|---|---|
| author | Ioannis D. Apostolopoulos Dimitris J. Apostolopoulos Nikolaos D. Papathanasiou |
| author_facet | Ioannis D. Apostolopoulos Dimitris J. Apostolopoulos Nikolaos D. Papathanasiou |
| author_sort | Ioannis D. Apostolopoulos |
| collection | DOAJ |
| description | X-ray technology has been recently employed for the detection of the lethal human coronavirus disease 2019 (COVID-19) as a timely, cheap, and helpful ancillary method for diagnosis. The scientific community evaluated deep learning methods to aid in the automatic detection of the disease, utilizing publicly available small samples of X-ray images. In the majority of cases, the results demonstrate the effectiveness of deep learning and suggest valid detection of the disease from X-ray scans. However, little has been investigated regarding the actual findings of deep learning through the image process. In the present study, a large-scale dataset of pulmonary diseases, including COVID-19, was utilized for experiments, aiming to shed light on this issue. For the detection task, MobileNet (v2) was employed, which has been proven very effective in our previous works. Through analytical experiments utilizing feature visualization techniques and altering the input dataset classes, it was suggested that MobileNet (v2) discovers important image findings and not only features. It was demonstrated that MobileNet (v2) is an effective, accurate, and low-computational-cost solution for distinguishing COVID-19 from 12 various other pulmonary abnormalities and normal subjects. This study offers an analysis of image features extracted from MobileNet (v2), aiming to investigate the validity of those features and their medical importance. The pipeline can detect abnormal X-rays with an accuracy of 95.45 ± 1.54% and can distinguish COVID-19 with an accuracy of 89.88 ± 3.66%. The visualized results of the Grad-CAM algorithm provide evidence that the methodology identifies meaningful areas on the images. Finally, the detected image features were reproducible in 98% of the times after repeating the experiment for three times. |
| first_indexed | 2024-03-09T22:35:45Z |
| format | Article |
| id | doaj.art-67b11ddbcf8348e0835e3cf3dfb6a063 |
| institution | Directory Open Access Journal |
| issn | 2571-841X |
| language | English |
| last_indexed | 2024-03-09T22:35:45Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Reports |
| spelling | doaj.art-67b11ddbcf8348e0835e3cf3dfb6a0632023-11-23T18:49:36ZengMDPI AGReports2571-841X2022-05-01522010.3390/reports5020020Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature ReproducibilityIoannis D. Apostolopoulos0Dimitris J. Apostolopoulos1Nikolaos D. Papathanasiou2Department of Medical Physics, School of Medicine, University of Patras, 265-00 Patras, GreeceLaboratory of Nuclear Medicine, University Hospital of Patras, 265-00 Patras, GreeceLaboratory of Nuclear Medicine, University Hospital of Patras, 265-00 Patras, GreeceX-ray technology has been recently employed for the detection of the lethal human coronavirus disease 2019 (COVID-19) as a timely, cheap, and helpful ancillary method for diagnosis. The scientific community evaluated deep learning methods to aid in the automatic detection of the disease, utilizing publicly available small samples of X-ray images. In the majority of cases, the results demonstrate the effectiveness of deep learning and suggest valid detection of the disease from X-ray scans. However, little has been investigated regarding the actual findings of deep learning through the image process. In the present study, a large-scale dataset of pulmonary diseases, including COVID-19, was utilized for experiments, aiming to shed light on this issue. For the detection task, MobileNet (v2) was employed, which has been proven very effective in our previous works. Through analytical experiments utilizing feature visualization techniques and altering the input dataset classes, it was suggested that MobileNet (v2) discovers important image findings and not only features. It was demonstrated that MobileNet (v2) is an effective, accurate, and low-computational-cost solution for distinguishing COVID-19 from 12 various other pulmonary abnormalities and normal subjects. This study offers an analysis of image features extracted from MobileNet (v2), aiming to investigate the validity of those features and their medical importance. The pipeline can detect abnormal X-rays with an accuracy of 95.45 ± 1.54% and can distinguish COVID-19 with an accuracy of 89.88 ± 3.66%. The visualized results of the Grad-CAM algorithm provide evidence that the methodology identifies meaningful areas on the images. Finally, the detected image features were reproducible in 98% of the times after repeating the experiment for three times.https://www.mdpi.com/2571-841X/5/2/20deep learningCOVID-19explainable artificial intelligence |
| spellingShingle | Ioannis D. Apostolopoulos Dimitris J. Apostolopoulos Nikolaos D. Papathanasiou Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility Reports deep learning COVID-19 explainable artificial intelligence |
| title | Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility |
| title_full | Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility |
| title_fullStr | Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility |
| title_full_unstemmed | Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility |
| title_short | Deep Learning Methods to Reveal Important X-ray Features in COVID-19 Detection: Investigation of Explainability and Feature Reproducibility |
| title_sort | deep learning methods to reveal important x ray features in covid 19 detection investigation of explainability and feature reproducibility |
| topic | deep learning COVID-19 explainable artificial intelligence |
| url | https://www.mdpi.com/2571-841X/5/2/20 |
| work_keys_str_mv | AT ioannisdapostolopoulos deeplearningmethodstorevealimportantxrayfeaturesincovid19detectioninvestigationofexplainabilityandfeaturereproducibility AT dimitrisjapostolopoulos deeplearningmethodstorevealimportantxrayfeaturesincovid19detectioninvestigationofexplainabilityandfeaturereproducibility AT nikolaosdpapathanasiou deeplearningmethodstorevealimportantxrayfeaturesincovid19detectioninvestigationofexplainabilityandfeaturereproducibility |