Dental caries is a widespread oral health concern and one of the foreseen issues in India. An undiagnosed dental caries can lead to severe complications if not diagnosed and treated early. Traditional diagnostic methods, such as visual inspection and radiographic imaging, are often subjective, time-consuming, and prone to human made errors among dental professionals. With the current rise in the advancements in artificial intelligence (AI) and deep learning, automated early caries detection system has become a needed solution for improving diagnostic accuracy and mass deployment. To address the aforementioned problem, we propose a deep learning-based approach for the early detection of dental caries using 3D intra-oral scans. Furthermore, we have also introduced a novel dataset, curated in collaboration with Indira Gandhi Government Dental College, Jammu. The dataset consists of high-resolution 3D intra-oral scans acquired using the Helios-500 dental scanner and meticulously annotated by experienced dental professionals. These annotations provide ground truth labels for training and evaluating our model with real-life dental cases. Our approach leverages modified YOLOv5, a state-of-the-art object detection framework, to accurately identify carious lesions from 3D intraoral scans. The scan samples in the dataset are converted into top-view images from the 3D model and are carefully labelled, ensuring optimal performance for deep learning-based analysis. We have adhered YOLOv5 and modified it to detect and classify caries, distinguishing between 3 categories i.e. filling, carries and suspected carries (risk of having carries). Along YOLOv5, we have also experimented with different pre-processing modules utilizing transformation to new color spaces and extracting features from existing channels. We have evaluated the performance of our proposed system using precision, recall, F1-score, and mean average precision (mAP). Experimental results demonstrate that our approach enables faster, more consistent, and highly accurate early detection of dental caries, supporting timely diagnosis and scalable deployment with minimal clinical intervention.