Pneumonia is an infection in respiratory system that can be triggered by bacteria or viruses, significantly impacting many persons, especially in developing and underdeveloped countries. These regions often experience high pollution levels, poor living conditions, and overcrowding, compounded by insufficient medical resources. Pneumonia can lead to Hydrothorax, a condition where fluid is accumulated in the lungs, which results in breathing difficulties. Diagnosis of pneumonia at early stages is essential for effective treatment and for improving the rate of survival. Chest Radiograph is the most commonly utilized technique for pneumonia diagnosis however interpreting these images can be complex and subject to individual interpretation differences. In this study, we created a computer-assisted diagnostic system designed to automate the process of identifying pneumonia using X-rays of lungs. We made use of deep transfer learning to address the limited availability of data and created a Convolutional Neural Network (CNN) model by merging four transfer learning techniques: CovXNet, MobileNet, and VGG16. The existing method using ResNet 50 did not achieve satisfactory accuracy, prompting the exploration of alternative transfer learning methods. The effectiveness of the approach that was proposed was assessed using a pneumonia x-ray dataset that is available publicly.