Medical imaging is an essential component of modern healthcare. Medical imaging techniques such as X-ray, computed tomography (CT), and magnetic resonance imaging (MRI) provide complementary views of anatomical structures and pathological changes, often serving as the first line of evidence in clinical decision-making. The growing demand for faster and more accurate interpretation of medical images has increased interest in artificial intelligence, in particular, convolutional neural networks (CNNs). CNNs have achieved high performances in many computer vision tasks, but their effectiveness can vary depending on the imaging modality, data quality, and the disease context. The images used in this experiment include dental X-rays, bone fracture X-rays, brain stroke CT scans, and Alzheimer’s MRI images. The goal of this study is to conduct a comparative evaluation of CNN architectures across multiple two-dimensional medical imaging modalities. The results showed a strong overall performance, with high accuracy and balanced precision-recall tradeoffs in most datasets, and particularly strong outcomes from the brain stroke and dental datasets. The model consistently achieved competitive AUC values, underscoring its robustness and adaptability across diverse imaging modalities.