TRAFFIC SIGN DETECTION USING MACHINE LEARNING

Abstract

Our road system will not be complete without traffic signals. They offer important guidance to road users,
and often persuasive advice, which allows them to change their driving behaviour to ensure compliance with whatever
road regulations are currently in effect. To identify and classify traffic signs in the past, traditional computer vision
techniques were used, but this involved intensive and time-consuming manual work to handcraft essential features in
pictures. Instead, we build a model that accurately classifies traffic signals using deep learning, which learns to find
the most suitable features for this problem on its own. This seminar covers the identification and understanding of
road signs, which is essential for a variety of expert programs, such as driver assistance and self-handling. In this
article, the method of identification and recognition for Indonesian road signs was examined. A detailed study of
different classification methods for detecting signs is carried out, with the findings demonstrating the reliability and
accuracy of each process in terms of Precision and Recall.