Transfer Learning-Driven Deep Neural Network for Pneumonia Detection from Chest X-Rays

Pneumonia is recognized as a serious medical condition with pronounced effects among children and older adults. Chest X-ray imaging is commonly employed as a primary clinical tool for identifying pneumonia cases. In this work, we systematically examine three learning-based approaches for pneumonia classification: a VGG16-based pretrained model, a custom convolutional neural network (CNN), and a baseline multi-layer perceptron (MLP). We propose LungXpert, a VGG16-based framework fine-tuned for chest X-ray classification. Experimental results show that the proposed model attains improved performance relative to previously reported approaches, reaching an accuracy of 97.22% with consistently strong precision, recall, and F1-score. The reported results demonstrate the practical value of pretrained representations in medical imaging tasks and indicate the potential of LungXpert as a dependable computer-aided diagnostic tool for pneumonia classification.