COVID-19 detection on chest radiographs using feature fusion based deep learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Deep Learning

Background:

• The COVID-19 pandemic severely impacted global healthcare systems, highlighting the need for efficient diagnostic tools.
• Existing COVID-19 diagnostic tests were often expensive, slow, and yielded inconclusive results, hindering timely patient isolation and virus containment.
• Overwhelmed healthcare systems necessitated innovative solutions for rapid and accurate disease diagnosis.

Purpose of the Study:

• To propose a cost-effective method for diagnosing COVID-19 using X-ray imaging.
• To develop and evaluate a deep learning model for automated COVID-19 diagnosis based on chest X-ray images.

Main Methods:

• A multi-stream convolutional neural network (CNN) model was implemented for feature extraction and classification.
• The CNN model processed three types of input images: grayscale, Local Binary Patterns (LBP), and Histograms of Oriented Gradients (HOG).
• The model's performance was evaluated using fivefold cross-validation on a public dataset of 3886 X-ray images across three classes.

Main Results:

• The proposed deep learning model achieved a high diagnostic accuracy of 97.76%.
• The model demonstrated superior performance compared to other existing algorithms.
• The results indicate the model's potential for accurate and rapid COVID-19 diagnosis.

Conclusions:

• X-ray imaging, combined with deep learning, offers a viable and cost-effective approach for COVID-19 diagnosis.
• The developed artificial intelligence tool can significantly alleviate the workload in healthcare systems.
• This automated diagnosis system shows promise in improving pandemic response and patient management.