Detecting cavities in medical imaging is essential for the early diagnosis and treatment of various diseases. Machine learning techniques, particularly convolutional neural networks (CNNs), have shown great promise in automating this process. This paper investigates the application of advanced object detection models, specifically YOLOv8 and YOLOv5, for cavity detection in medical images. Additionally, these models have been deployed on edge AI devices, such as the Jetson Nano, to facilitate real-time inference at the point of care. The study builds on extensive medical imaging and machine learning research, particularly object detection, by incorporating findings from previous studies. The effectiveness of deep learning in detecting dental cavities underscores its potential to enhance detection accuracy and efficiency. The study also examines the benefits of deploying machine learning models on edge devices for real-time medical image analysis. Experimental results demonstrate the efficacy of the YOLOv8 and YOLOv5 models in detecting cavities across various medical imaging modalities, including X-ray. Furthermore, the performance of these models was evaluated on the Jetson Nano and Jetson Orin, showcasing real-time inference capabilities suited for point-of-care applications. YOLOv8 achieved a higher precision of 65%, while YOLOv5 reached 62%.