Coslets for Classification of Covid-19 X-Ray Images

COVID-19 (Corona virus) endures to have disastrous effects on the lives of human beings. Corona virus was first reported in the year 2019 and later widely spread throughout the world. The machine learning approaches are extensively used to analyze the characteristics of Covid-19 pneumonia in chest X-ray images. In this paper, a sequential approach is implemented to study and classify theCovid-19 pneumonia case, non-Covid-19 case and Normal cases. Discrete Cosine Transformation (DCT) and wavelet decomposition methods are employed to extract the coslets features from the chest X-ray images. The discriminative properties are retained by employing Singular Value Decomposition (SVD) method. Finally, Probabilistic Neural Network (PNN) and Generalized Regression Neural Network (GRNN) is applied to classify the given chest X-ray images into three classes such as Covid-19 pneumonia case, non-Covid-19 case and Normal case. The performance of the proposed method is evaluated in terms of accuracy, sensitivity and specificity. The result shows that method is promising and encouraging.