A deep learning approach for COVID-19 screening and localization on Chest X-Ray images
Carregando...
Citações na Scopus
0
Tipo de produção
conferenceObject
Data de publicação
2022
Título da Revista
ISSN da Revista
Título do Volume
Editora
SPIE-INT SOC OPTICAL ENGINEERING
Autores
MARCOMINI, Karem Daiane
TRAINA, Agma Juci Machado
Citação
MEDICAL IMAGING 2022: COMPUTER-AIDED DIAGNOSIS, v.12033, article ID 1203327, 9p, 2022
Resumo
Chest X-ray (CXR) images have a high potential in the monitoring and examination of various lung diseases, including COVID-19. However, the screening of a large number of patients with diagnostic hypothesis for COVID-19 poses a major challenge for physicians. In this paper, we propose a deep learning-based approach that can simultaneously suggest a diagnose and localize lung opacity areas in CXR images. We used a public dataset containing 5, 639 posteroanterior CXR images. Due to unbalanced classes (69.2% of the images are COVID-19 positive), data augmentation was applied only to images belonging to the normal category. We split the dataset into train and test sets with proportional rate at 90:10. To the classification task, we applied 5-fold cross-validation to the training set. The EfficientNetB4 architecture was used to perform this classification. We used a YOLOv5 pre-trained in COCO dataset to the detection task. Evaluations were based on accuracy and area under the ROC curve (AUROC) metrics to the classification task and mean average precision (mAP) to the detection task. The classification task achieved an average accuracy of 0.83 +/- 0.01 (95% CI [0.81, 0.84]) and AUC of 0.88 +/- 0.02 (95% CI [0.85, 0.89]) in 5-fold over the test dataset. The best result was reached in fold 3 (0.84 and 0.89 of accuracy and AUC, respectively). Positive results were evaluated by the opacity detector, which achieved a mAP of 59.51%. Thus, the good performance and rapid diagnostic prediction make the system a promising means to assist radiologists in decision making tasks.
Palavras-chave
COVID-19, classification, opacity detection, deep neural network
Referências
- Al-antari MA, 2021, APPL INTELL, V51, P2890, DOI 10.1007/s10489-020-02076-6
- de la Iglesia Vaya M., 2020, ARXIV
- Fang YC, 2020, RADIOLOGY, V296, pE115, DOI 10.1148/radiol.2020200432
- Floriano I, 2020, REV ASSOC MED BRAS, V66, P880, DOI 10.1590/1806-9282.66.7.880
- Iyer R. V., 2021, COMP YOLOV3 YOLOV5S
- Jain G, 2020, BIOCYBERN BIOMED ENG, V40, P1391, DOI 10.1016/j.bbe.2020.08.008
- Karthik R, 2021, APPL SOFT COMPUT, V99, DOI 10.1016/j.asoc.2020.106744
- Lakhani P., 2021, 2021 SIIM FISABIO RS
- Ohata EF, 2021, IEEE-CAA J AUTOMATIC, V8, P239, DOI 10.1109/JAS.2020.1003393
- Ozturk T, 2020, COMPUT BIOL MED, V121, DOI 10.1016/j.compbiomed.2020.103792
- Panwar H, 2020, CHAOS SOLITON FRACT, V138, DOI 10.1016/j.chaos.2020.109944
- Rahman T, 2021, COMPUT BIOL MED, V132, DOI 10.1016/j.compbiomed.2021.104319
- Sirazitdinov I, 2019, COMPUT ELECTR ENG, V78, P388, DOI 10.1016/j.compeleceng.2019.08.004
- Sitaula C, 2021, APPL INTELL, V51, P2850, DOI 10.1007/s10489-020-02055-x
- Tan MX, 2019, PR MACH LEARN RES, V97
- Tsai EB, 2021, RADIOLOGY, V299, pE204, DOI 10.1148/radiol.2021203957
- Zebin T, 2021, APPL INTELL, V51, P1010, DOI 10.1007/s10489-020-01867-1