Most present research focuses solely on the identification of cavities in the teeth in X-ray scans. We have developed a dental cavity detection system using convolutional neural networks (CNNs) to analyse both colour and X-ray images. The proposed system will first preprocess the images to augment the dataset and create more image samples. The CNN algorithm will next be trained on a huge dataset of annotated dental pictures to understand the characteristics of dental cavities. The trained model will be able to accurately detect cavities in both colour and X-ray images. The model for colour images gives an average accuracy of 95%, and the model for X-ray images gives an accuracy of 83%. The system will be evaluated on a test set of dental images and give predictions for cavity detection. Also, we can provide dentists and other healthcare professionals with a reliable tool to assist in the early detection of dental cavities with limited treatment costs.