Explainable DCNN based chest X-ray image analysis and classification for COVID-19 pneumonia detection.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Pneumonia Diagnosis

Background:

• Accurate and timely diagnosis of COVID-19 pneumonia is critical for patient management and public health.
• Chest X-rays are a common imaging modality for diagnosing pneumonia, but interpretation can be time-consuming and subjective.
• The need for efficient and accurate diagnostic tools is paramount, especially during pandemics.

Purpose of the Study:

• To develop and evaluate a deep convolutional neural network (DCNN) based diagnostic platform for COVID-19 pneumonia detection using chest X-ray images.
• To enhance the accuracy and speed of COVID-19 diagnosis, assisting radiologists in clinical decision-making.
• To incorporate explainable AI methods to improve model transparency and prediction accuracy.

Main Methods:

• Development of a DCNN model for classifying COVID-19 pneumonia from non-COVID-19 pneumonia based on chest X-rays.
• Utilizing an explainable AI approach to analyze and select relevant instances from the X-ray dataset for model training.
• Performance evaluation based on classification accuracy and comparison with manual interpretation.

Main Results:

• The DCNN platform demonstrated an average accuracy exceeding 96% in distinguishing COVID-19 pneumonia from other types.
• The explainable AI component contributed to higher prediction accuracy by refining model behavior.
• The developed tool showed potential to replace manual X-ray reading for large-scale screening.

Conclusions:

• The DCNN-based diagnostic platform offers a highly accurate and efficient solution for COVID-19 pneumonia detection from chest X-rays.
• The integration of explainable AI enhances model reliability and accuracy, supporting clinical adoption.
• This technology has the potential to significantly improve medical capacity for rapid screening and diagnosis of COVID-19.