ECOVNet: a highly effective ensemble based deep learning model for detecting COVID-19.

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Deep Learning

Background:

• Accurate and timely detection of COVID-19 is crucial for effective patient management and disease control.
• Chest X-rays are a widely accessible imaging modality for diagnosing respiratory illnesses, including COVID-19.

Purpose of the Study:

• To develop and implement a highly effective deep learning model for COVID-19 detection using chest X-rays.
• To enhance the robustness and classification performance of the deep learning model through ensemble strategies.

Main Methods:

• Proposed an ensemble of Convolutional Neural Network (CNN) based on EfficientNet, named ECOVNet.
• Utilized the large COVIDx chest X-ray dataset, including COVID-19, normal, and pneumonia cases.
• Employed EfficientNet with ImageNet pre-training for feature extraction and applied hard and soft ensemble strategies for prediction consolidation.
• Incorporated a visualization technique to identify key areas in X-rays indicative of COVID-19.

Main Results:

• The ECOVNet model demonstrated superior performance compared to state-of-the-art approaches.
• Achieved 100% recall for COVID-19 detection and an overall accuracy of 96.07%.
• Visualization techniques highlighted discriminative regions related to COVID-19, improving model interpretability.

Conclusions:

• ECOVNet offers a robust and highly effective solution for automated COVID-19 detection from chest X-rays.
• The proposed model has the potential to significantly enhance diagnostic capabilities and support public health efforts.
• Further development could lead to a fully automated and efficacious COVID-19 detection system.