An Efficient Neural Network Model For Pneumonia Detection

Abstract

Healthcare image classification is crucial in anomaly diagnosis. The disease Pneumonia is a form of acute respiratory infection that affects either one or both lungs. It is one of the most life-threatening diseases and is the leading cause of death among children. Conventional methods of detection have reached their performance limits. Radiologists on average observe hundreds of chest x-rays per day. Furthermore, the use of the conventional methods requires significant time and effort to extract and select categorical features. In image classification, deep neural networks dominate the field with efficient results using convolutional neural network architectures. Deep networks demonstrate their potential in a variety of classification tasks. Nevertheless, clinical image datasets are difficult to obtain because it is a tedious task that requires expertise to label them. This paper investigates identifying an efficient convolutional neural network (CNN) architecture for x-ray based image classification and developing a robust AI model based upon it to identify the presence of pneumonia in chest x-ray images.