DETECTION AND RECOGNITION OF ROAD SIGNS USING YOLOv5

In the field of deep learning, a convolutional neural network is a class of artificial neural networks that became dominant in various computer vision tasks, which is widely used to solve complex problems in various areas, including driver assistance systems in the auto- motive field. Convolutional neural networks overcome the limitations of others conventional machine learning approaches since they are designed to automatically and adaptively learn the spatial characteristics of features in an image. In this paper, we are going to evaluate the inference and accuracy of YOLOv5s, for effective traffic sign detection in various environments. The results generated upon five classes gives satisfaction by 63.7% for the mean average precision, and over 80% in accordance to 5 categories set in this study. This article compared to YOLOV4 based CSP-DarkNet53 using Indonesia Traffic Signs generate better precision.