A BiLSTM model enhanced with multi-objective arithmetic optimization for COVID-19 diagnosis from CT images.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Current artificial intelligence (AI) algorithms for COVID-19 diagnosis using CT imaging struggle with accuracy and efficiency due to viral mutations.
• The need for improved automated diagnostic tools is critical for timely patient management.

Purpose of the Study:

• To enhance the accuracy and efficiency of automated COVID-19 diagnosis from CT images.
• To propose and validate a novel multi-objective optimization algorithm (MOAOA) integrated with a Bidirectional Long Short-Term Memory (BiLSTM) model.

Main Methods:

• Development of a multi-objective optimization algorithm (MOAOA) to optimize hyperparameters of the Bidirectional Long Short-Term Memory (BiLSTM) model.
• Training and validation of the enhanced BiLSTM model on publicly accessible medical imaging datasets for COVID-19 detection.
• Comparative performance analysis against state-of-the-art diagnostic techniques.

Main Results:

• The proposed MOAOA-enhanced BiLSTM model achieved a diagnostic accuracy of 95.32% and specificity of 95.09%.
• Significant improvements in accuracy, efficiency, and overall performance metrics were observed compared to existing methods.
• The model demonstrated robust performance on diverse medical datasets.

Conclusions:

• The MOAOA-enhanced BiLSTM model represents a significant advancement in AI-driven COVID-19 diagnosis from CT scans.
• This approach offers a more accurate and efficient solution for automated medical image analysis in the context of infectious diseases.
• The findings suggest potential for broader application in clinical settings for rapid and reliable disease detection.