查看文章

In general, pneumonia affects children under 5 years and adults over 65 years of age which targets the lungs and fills the alveoli (air sacs) with liquid. In this paper, we employ convolutional neural networks (CNNs) of varying configurations on a machine learning based binary classification task with a given dataset of chest X-rays that depicts affected and unaffected cases of pneumonia. This paper primarily focuses on putting forth the performances of different simple CNN architectures and selecting the best architecture based on optimum corresponding minimum loss and maximum accuracy which can serve as a viable tool for physicians and the medicine community to correctly identify and diagnose viral, bacterial, fungal-caused and community acquired pneumonia given only the chest X-ray of the patient.