Augmented YOLOv5 for Indian Traffic Signs Identification

Traffic signal comprehension and adherence are essential for road safety and efficient traffic management. This research focuses on developing a method for identifying Indian traffic signs using the YOLOv5 object recognition algorithm. Utilizing a publicly available dataset of Indian traffic signs, the YOLOv5 algorithm was trained for real-time sign identification and classification. The dataset preprocessing is carried out to ensure adequate sample size and uniform representation across all sign classes. Proposed method achieves an accuracy of 96.7% for traffic signs. The efficiency of the proposed method is analyzed in terms of accuracy and time efficiency, as evaluated by metrics including class, epoch, accuracy, recall, f1-score, and frames per second (fps). This method’s application extends to real-world scenarios such as driverless vehicles, traffic management systems, and driver assistance systems, contributing significantly to road safety and accident reduction efforts.