Pneumonia is a common respiratory disease caused by numerous sorts of microscopic organisms, infections, and parasites. It is the driving cause of dreariness and mortality around the world, especially among newborn children beneath the age of five and the elderly. Chest X-ray imaging is commonly utilized to analyze pneumonia, as it can uncover critical indications, such as expanded lung murkiness and combination. Be that as it may, it can be troublesome to decipher a chest X-ray (CXR) since pneumonia side effects can be unpretentious and cover with other lung maladies. In this proposed investigation, chest X-ray pictures from each lesson are recognized, assisting CNN engineering is built. Advancement the SGD shows with optimizer is compiled as SGD, misfortune work, SparseCategoricalCrossEntropy, and measurements as exactness. CNN design utilizing VisualKeras has been performed. Thus from dataset the misclassified pictures are distinguished and visualized. In terms of precision and misfortune bends, the approval comes about is displayed. This proposed chest X-ray picture classification framework demonstrated that profound learning has a few competitive focal points which may be utilized in diverse areas of restorative picture investigation like pneumonia.