Jaya-tunicate swarm algorithm based generative adversarial network for COVID-19 prediction with chest computed...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Coronavirus disease (COVID-19) emerged as a significant respiratory illness.
• Early prediction of COVID-19 is crucial but remains a challenge.
• Chest computed tomography (CT) scanning is vital for COVID-19 monitoring.

Purpose of the Study:

• To develop an effective prediction mechanism for early-stage COVID-19 detection.
• To propose a novel Jaya-tunicate swarm algorithm driven generative adversarial network (Jaya-TSA with GAN).
• To enhance the accuracy and efficiency of COVID-19 patient identification.

Main Methods:

• Lung lobes segmentation using Bayesian fuzzy clustering.
• Feature extraction and dimensionality reduction for efficient processing.
• Generative Adversarial Network (GAN) trained with the proposed Jaya-TSA for prediction.
• Optimization of GAN training using a fitness measure.

Main Results:

• The Jaya-TSA with GAN achieved high effectiveness in COVID-19 prediction.
• Key performance metrics included specificity (0.8857), accuracy (0.8727), and sensitivity (0.85).
• Feature optimization accelerated the training process.

Conclusions:

• The proposed Jaya-TSA with GAN demonstrates a promising approach for early COVID-19 prediction.
• The integration of Jaya algorithm and tunicate swarm algorithm enhances prediction capabilities.
• This method offers a significant advancement in utilizing chest CT scans for infectious disease detection.