According to UNICEF, at least a child dies every 45 s from pneumonia. Pneumonia is arespiratory infection resulting from various bacterial or viral infections which affects the lungs and causes breathing problem. Pneumonia is usually identified by chest X-ray, but chest X-ray of other pulmonary diseases could be similar to pneumonia. Deep learning, particularly Convolutional Neural Networks (CNNs), has become instrumental in medical imaging for disease identification. When the COVID-19 pandemic emerged in late 2019, CNNs were widely employed to detect it using chest X-rays. In this study, we introduce a CNN model trained from scratch to classify and detect pneumonia in chest X-ray images sourced from Kaggle. Unlike methods relying solely on transfer learning or traditional techniques, our CNN-based Computer-Aided Diagnosis (CAD) system extracts features and classifies images by leveraging Convolution, RELU, Pooling, and Fully Connected Layers. We enhance classification accuracy through normalization and data augmentation techniques, achieving an impressive 96.23% testing accuracy. Furthermore, we calculate precision, recall, and F1-score, facilitating meaningful comparisons with other models trained on the dataset.