Enabling CT-Scans for covid detection using transfer learning-based neural networks.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Diagnostics

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools to manage patient influx and prevent disease spread.
• Current diagnostic methods can be time-consuming and may require in-person visits, contributing to healthcare system strain and public gathering risks.
• The need for a robust, efficient system to identify COVID-19 cases quickly and remotely is paramount.

Purpose of the Study:

• To develop and evaluate an AI-driven system for detecting COVID-19 using chest CT images.
• To leverage transfer learning with a Convolutional Neural Network (CNN) model for improved diagnostic performance.
• To provide a tool that assists healthcare professionals and individuals in identifying potential COVID-19 infections without public exposure.

Main Methods:

• Utilized the VGG-19 model, a CNN architecture, for image classification tasks.
• Employed the open-source COVID-CT (CTSI) dataset, comprising 349 COVID-19 positive and 463 non-COVID-19 chest CT images.
• Applied transfer learning techniques to train the VGG-19 model on the provided dataset.

Main Results:

• The AI model achieved a high accuracy of 95% in distinguishing between COVID-19 and non-COVID-19 cases.
• Performance metrics included a precision of 96%, recall of 94%, and an F1-Score of 96%.
• Experimental results demonstrate the model's strong capability in classifying chest CT scans for COVID-19 detection.

Conclusions:

• The developed AI model, based on VGG-19 CNN, shows significant potential for accurate and rapid COVID-19 diagnosis from CT scans.
• This approach offers a scalable solution to aid in pandemic management by enabling remote screening and reducing healthcare facility crowds.
• The high performance metrics suggest the VGG-19 model is a viable tool for supporting clinical decisions in COVID-19 diagnostics.