A Quantitative Analysis Technique for Early Detection of Respiratory Virus in Lung Region Using PNN

Abstract

Respiratory detection system (RDS) is one of the foremost driving causes of around the world passing and requires appropriate therapeutic treatment. As a result, the aim of this paper is to categories lung MRI channel images as undesirable or commonplace. Our most recent suggested Probabilistic neural network (PNN) appearance makes use of image mixture and PNN approaches. To begin with, multiple preprocessing operations have been used by using multi-focus image combination in order to improve the accuracy of MRI images. Energize, preprocessed images are energized into the recently proposed 13-layer PNN structure for RDS classification. Two experiments on two separate databases were used to assess the consistency of our PNN protocol. The MRI image dataset is divided into 20% research and 80% preparation sets in the first attempt, but 10-fold cross-validation of the image dataset is done in the second attempt. The classification exactness obtained by our methodology on dataset 1 in the first test is 98.33 percent, and in the second test, it is 98.77 percent, while in dataset 2, the exactness obtained in tests 1 and 2 is 92.22 percent and 93.33 percent, respectively.