Dental caries, an infection affecting the tooth and its supporting structures, is the leading cause of tooth loss. This issue occurs due to insufficient dental care. Initial signs can include persistent bad breath or an unpleasant taste, bleeding, or other indicators of gum disease, as well as toothaches or mouth discomfort and heightened sensitivity to hot or cold food and beverages. Identifying caries early in children is beneficial, as advanced cases can lead to severe pain and infection, often requiring tooth extraction. This research aims to create a simpler and quicker method for diagnosing dental cavities through deep learning and soft computing techniques, focusing on early detection with computer vision and digital colour images. The study will utilize Roboflow for data annotation and preprocessing while employing the YOLOv8 model, which is a convolutional neural network, to determine the stages of caries. This model uses machine learning algorithms to detect and classify objects in real-time scenarios, enhancing dental image analysis. Data annotation will be carried out on Roboflow, a platform used to manage, annotate, and create datasets for computer vision tasks. Two labels, namely Healthy and Decay, will be created and annotated manually on all images in the dataset for the detection of dental caries. Medical imaging in oral health can prevent complex dentistry. This non-invasive approach, which relies on digital colour images, aims to improve the comfort of patients undergoing diagnosis without sacrificing the precision of the outcomes.