Detecting COVID-19 Opacity in X-ray Images Using YOLO and RetinaNet Ensemble
2022 IEEE Delhi Section Conference, DELCON 2022
; 2022.
Article
in English
| Scopus | ID: covidwho-1846076
ABSTRACT
Coronavirus disease(COVID-19) is caused by SARS-COV-2 virus and has been declared as a pandemic. After almost two years of this pandemic, over five million people have lost their lives worldwide due to complications like pneumonia and acute respiratory distress syndrome. Many countries have already witnessed the second wave of pandemic and a huge loss of lives. One way to curb the disease spread is by timely and accurate diagnosis. X-rays and CT-scans can help a radiologist to detect the disease, but detecting COVID-19 on chest radiographs can become challenging as it has similarities with other bacterial and viral pneumonias. Hence, there is a need to develop an algorithm for accurate and fast detection of COVID-19 in a patient. This work showcases the use of object detection deep learning models-You Only Look Once (YOLO) and RetinaNet for accurate localization of regions associated with COVID-19. Proposed method using ensemble of both the models achieves a mean average precision (mAP) score of 0.552, offering an improvement over their individual predictions. © 2022 IEEE.
Convolution Neural Network (CNN); Fully Convolutional Network (FCN); RetinaNet; Weighted Boxes Fusion (WBF); You Only Look Once (YOLO); Computerized tomography; Convolution; Convolutional neural networks; Deep learning; Diagnosis; Object detection; SARS; Acute respiratory distress syndrome; Convolution neural network; Convolutional networks; Coronaviruses; Fully convolutional network; Weighted box fusion; X-ray image; You only look once; Coronavirus
Full text:
Available
Collection:
Databases of international organizations
Database:
Scopus
Language:
English
Journal:
2022 IEEE Delhi Section Conference, DELCON 2022
Year:
2022
Document Type:
Article
Similar
MEDLINE
...
LILACS
LIS