A novel and efficient deep learning approach for COVID-19 detection using X-ray imaging modality.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Rising COVID-19 cases necessitate rapid, automated diagnostic tools.
• Current methods often have high computational demands, limiting accessibility.
• Accurate detection is crucial for controlling disease spread and enabling early intervention.

Purpose of the Study:

• To develop an accurate and computationally efficient deep learning model for detecting COVID-19 and pneumonia from chest X-ray images.
• To enhance the performance of Convolutional Neural Network (CNN) models through an ensemble approach.
• To provide a tool that assists medical practitioners in early disease diagnosis.

Main Methods:

• Utilized a dataset of 2161 COVID-19, 2022 pneumonia, and 5863 normal chest X-ray images.
• Applied contrast enhancement and image normalization for pre-processing.
• Employed data augmentation techniques to expand the training dataset.
• Trained an ensemble model using four efficient CNN architectures: Inceptionv3, DenseNet121, Xception, and InceptionResNetv2.

Main Results:

• Achieved 98.33% accuracy for binary classification (COVID-19 vs. normal).
• Attained 92.36% accuracy for multiclass classification (COVID-19, pneumonia, normal).
• Demonstrated the effectiveness of the proposed ensemble model in disease detection.

Conclusions:

• The developed deep learning ensemble model offers a highly accurate and efficient method for diagnosing COVID-19 and pneumonia from chest X-rays.
• The model's performance suggests its utility as a valuable tool for early disease detection in clinical settings.
• The approach addresses the need for reduced computational requirements in automated medical image analysis.