Traffic-sign Recognition and Detection using Yolo-v8

In order to ensure road safety and navigation, au-tonomous vehicles (AVs) and advanced driver-assistance systems (ADAS) depend on traffic sign detection and recognition (TSDR).The effectiveness of YOLOv8, a state-of-the-art deep learning framework for object detection, in real-time TSDR applications is examined in this research.Through a thorough methodology that includes data pre-processing (resizing, normalization, augmentation), YOLOv8 ar-chitecture (backbone network, feature pyramid network, predic-tion head), training (loss functions, optimizer selection, learning rate scheduling), and evaluation using standard metrics (mean Average Precision (mAP), precision, recall) for each traffic sign class, we investigate YOLOv8's performance on a well-established dataset.The model's performance is examined in detail in the results and discussion section, along with the effects of architectural changes or hyperparameter tuning (learning rate, batch size) on the model's strengths and shortcomings.Next, we discuss the difficulties that come with TSDR, such as occlusions, varying lighting and weather, and processing constraints in real-time for ADAS/AVs.Future research objectives and potential solutions are suggested, including the investigation of lightweight YOLOv8 models for embedded system deployment, the integration of depth information from LiDAR sensors, and domain adaption techniques.The study concludes by highlighting YOLOv8's efficacy for TSDR and emphasizing its importance for the advancement of AV and ADAS.We recognize the limits of the work and provide directions for future investigation to improve the robustness and generalizability of TSDR systems based on YOLOv8.