Deep Learning Based Entropy Controlled Optimization for the Detection of Covid-19

Emerging technological advancements open the door for employing deep learning-based methods in practically all spheres of human endeavor. Because of their accuracy, deep learning algorithms can be used in healthcare to categorize and identify different illnesses. The recent coronavirus (COVID-19) outbreak has significantly strained the global medical system. By using medical imaging and PCR testing, COVID-19 can be diagnosed. Since COVID-19 is highly transmissible, it is generally considered secure to analyze it with a chest X-ray. To distinguish COVID-19 infections from additional infections that are not COVID-19 infections, a deep learning-based entropy-controlled whale optimization (EWOA) with Transfer Learning is suggested in this paper. The created system comprises three stages: a preliminary processing phase to remove noise effects and resize the image, then a deep learning architecture using a pre-trained model to extract features from the pre-processed image. After extracting the features, optimization is carried out. EWOA is utilized to combine and optimize the optimum features. A softmax layer is used to reach the final categorization. Various activation functions, thresholds, and optimizers are used to assess the systems. Numerous metrics for performance are utilized to measure the performance of the offered methodologies for assessment. Through an accuracy of 97.95%, the suggested technique accurately categorizes four classes, including COVID-19, viral pneumonia, chest infection, and routine. Compared to current methodologies found in the literature, the proposed technique exhibits advantages regarding accuracy.