FPDCnet: A Multi-Classification Model for Classifying COVID-19, Pneumonia, and Normal Chest X-ray Imagery Using Fractional Partial Differential Algorithms

Globally, the coronavirus (COVID-19) has had a negative impact on economies and healthcare systems. Clinical practitioners may become confused while identifying this new form of flu because the symptoms of COVID-19 are similar to those of existing chest conditions as pneumonia, tuberculosis (TB), lung cancer (LC), and pneumothorax. To address this problem, this study created a model to categorize various chest infections. In medical practice, chest x-ray examinations are the most common diagnostic technique and the main means of identifying these different types of chest infections. Researchers and paramedics are working hard to develop an accurate and reliable technique for the early diagnosis of COVID-19 to save lives; however, the diagnosis of COVID-19 remains highly idiosyncratic and varies significantly. Therefore, a multi-classification method was created and tested in this work based on deep learning (DL) and a mathematical algorithm model for automatic differentiation of COVID-19 cases from pneumonia based on chest x-ray images. These two diseases were diagnosed using a hybrid model named FPDCnet, which integrates a convolutional neural network (CNN) and Fractional Partial Differential models. This model was used alongside publicly available benchmark data to identify the diseases, and the performance was evaluated using several objective metrics, where it recorded an impressive accuracy of 98.1 %; the model also demonstrated remarkable precision (0.982), recall (0.980), and F1-score (0.981) values.