RETRACTED ARTICLE: Covid-19 classification using sigmoid based hyper-parameter modified DNN for CT scans and chest...

Area of Science:

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

Background:

• Coronavirus disease (COVID-19) diagnosis using Computed Tomography (CT) and Chest X-rays (CXR) faces challenges like overfitting and early detection.
• Accurate and timely classification of respiratory conditions is crucial for effective patient management.

Purpose of the Study:

• To develop and evaluate a novel deep neural network (DNN) model for classifying COVID-19, Pneumonia, and normal cases from CT and CXR images.
• To enhance diagnostic accuracy and address limitations in current imaging-based detection methods.

Main Methods:

• Image pre-processing using an adaptive Gaussian filter for noise reduction.
• Implementation of a Sigmoid Based Hyper-Parameter Modified DNN (SHMDNN) model.
• Hyperparameter optimization utilizing the Adaptive Grey Wolf Optimization (AGWO) algorithm.

Main Results:

• The SHMDNN model successfully classified CT and CXR images into three categories: normal, Pneumonia, and COVID-19.
• Achieved a high classification accuracy of 99.9%.
• Demonstrated superior performance compared to other deep neural network architectures.

Conclusions:

• The proposed SHMDNN model with AGWO optimization offers a highly accurate and effective approach for diagnosing COVID-19 and Pneumonia from medical images.
• This method shows significant potential for improving early and accurate diagnosis in clinical settings.