Deep Learning-Based Detection and Severity Analysis of Corona and Pneumonia Using CNN with Wolf Optimization Algorithm

Effective identification and severity analysis of two serious respiratory diseases-pneumonia and corona-are desperately needed, and this study attempts to fill that gap. Early diagnosis and treatment of many conditions depend critically on prompt intervention. Diagnoses are made slowly and subjectively by existing procedures, which frequently rely on human interpretation of medical imaging data. The existing methods for severity analysis do not have a cohesive structure. Our study presents an all-encompassing strategy to get around these restrictions. First, to improve the quality of the images, grayscale-based pre-processing is applied. Next, marker-based watershed segmentation is used to precisely locate anomalies. Feature extraction and categorization are done using a convolutional neural network, or CNN. Specifically, for severity analysis, we optimize the CNN parameters by including the Wolf Optimization Algorithm (WOA). Grayscale pre-processing guarantees the best possible picture quality; marker-based watershed segmentation improves the accuracy of lesion identification; CNN streamlines feature extraction; and WOA supports the study of disease severity. Python is used in the study's implementation, which takes advantage of its many machine learning frameworks and tools. The outcomes show how well the suggested methodology works for precisely identifying and evaluating the levels of corona and pneumonia. By integrating WOA with CNN, the shortcomings of current techniques are addressed and diagnostic accuracy and severity evaluation are improved. When compared to alternative approaches like CNN-SVM, Bayesian CNN, DCNN with BPNN, the suggested CNN with WOA (99.3%) shows an average accuracy gain of about 11.68%, demonstrating higher performance in disease diagnosis and severity analysis. This innovative method offers a methodical and effective framework for better patient treatment, which has the potential to advance the field of respiratory disease diagnostics.