A Novel Hybrid Optimization Enabled Densenet for Covid-19 Classification using CT Images

Corona Virus disease 2019 (Covid-19) is an acute disease that affects the respiratory system alveolar of the human, which leads to serious illness and may cause death. When an infected person coughs, sneezes, speaks, sings, or breathes, the virus can spread from their mouth or nose in minute liquid particles. Disease diagnosis using Computed Tomography (CT) images is widely utilized to detect infection; still, accurate detection is a challenging task. Hence, this research introduces a novel hybrid optimization enabled DenseNet for Covid-19 classification. The hybrid optimization, named Gradient Mutated Leader Algorithm (GMLA) is proposed for tuning the weights of DenseNet by combining the Mutated leader's behavior in guiding the group member in position updation with the gradient descent algorithm for the acquisition of computation efficiency with stable convergence, which helps to obtain the global best solution for tuning the DenseNet's weights to make the classification more accurate. In addition, the proposed GMLA_DenseNet utilizes the data augmentation technique using rotating, shifting, zooming, and flipping to obtain balanced data with enormous data sizes, which makes the classification more efficient. The performance of GMLA_DenseNet is evaluated using True Negative Rate (TNR), True Positive Rate (TPR), Precision, and segmentation accuracy and obtained the maximal values of 0.920, 0.919, 0.919, and 0.919, respectively.