Vehicle recognition and identification is a critical aspect in many modern applications, including autonomous driving, traffic monitoring, and smart city infrastructure. With the increasing number of vehicles on roads, ensuring efficient and real-time recognition of vehicles is essential for optimizing traffic flow, reducing accidents, and improving overall safety. Traditional object detection models face challenges related to high computational costs and time complexity, especially when dealing with diverse vehicle types. To address these challenges, we propose a novel vehicle recognition system based on YOLOv8, augmented by an innovative object grouping mechanism that significantly reduces time complexity from \(O(n)\) to \(O(k)\) , where \(k\) represents the number of unique vehicle types rather than the total number of vehicles represented by \(n\) . . By employing a COCO dataset and introducing a counter-based grouping technique, our system minimizes the number of direct comparisons needed during vehicle identification. We extract features using Convolutional Feature Maps (CFMs), Region of Interest (RoI) Pooling, and Scale-Invariant Feature Transform (SIFT), which serve as the basis for clustering vehicles into groups. The proposed system achieves a precision of 93.5%, recall of 92.3%, and mAP@0.5 of 91.7%, while reducing inference time from 38 to 25ms. Comparative analysis with recent state-of-the-art models such as RCNN, SVM, and CNN-based models demonstrates the superiority of our approach in terms of both accuracy and efficiency. Our research demonstrates a promising pathway for real-time vehicle recognition systems, offering potential applications in urban traffic management, surveillance, and autonomous systems.