GW- CNNDC: Gradient weighted CNN model for diagnosing COVID-19 using radiography X-ray images.

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Infectious Disease Diagnostics

Background:

• COVID-19 poses a significant global health threat, necessitating rapid and accurate diagnostic methods.
• Traditional COVID-19 tests like RTPCR are time-consuming, while CT scans involve radiation exposure.
• Limited availability and speed of current testing units create an urgent need for alternative detection measures.

Purpose of the Study:

• To develop and evaluate a novel deep learning model for efficient COVID-19 detection using chest X-ray (CXR) images.
• To overcome the limitations of existing COVID-19 diagnostic techniques, such as RTPCR and CT scans.
• To enhance diagnostic accuracy and speed for COVID-19 detection in medical imaging.

Main Methods:

• Utilized chest X-ray (CXR) images for COVID-19 detection.
• Employed a deep learning approach, specifically a Convolutional Neural Network (CNN) model enhanced with RESNET-50 architecture.
• Implemented a Gradient Weighted model (GW-CNNDC) for precise identification of COVID-19 affected areas in lung radiography.
• Processed images at a resolution of 255*255 pixels.

Main Results:

• The GW-CNNDC model demonstrated proficiency in binary classification tasks for COVID-19 detection.
• Achieved high performance metrics including accuracy, precision, recall, and F1-score.
• The model proved efficient for large datasets, processing them in a reduced amount of time.
• Successfully identified specific COVID-19 affected regions within lung radiography images.

Conclusions:

• The GW-CNNDC model presents a viable and efficient alternative for COVID-19 detection using CXR images.
• This deep learning framework offers a safer alternative to radiation-exposed CT scans and a faster alternative to RTPCR.
• The model's ability to accurately analyze CXR images contributes to improved early detection and management of COVID-19.