Avaliação comparativa de modelos de visão computacional para reconhecimento de placas de trânsito

Abstract

Intelligent vehicle automation is an area of study that has been growing significantly with the evolution of artificial intelligence applied to computer vision. This area includes everything from driver assistance systems till the development of completely autonomous vehicles. In this context, the detection of traffic signs plays a fundamental role in providing a intelligent system with relevant information that can assist in driving. In this work, is briefly discussed the object detection models who are most used in current literature, as well as the metrics used to evaluate these models. Therefore, this work aims to compare and analyze some of the most relevant object detection models, in order to conclude which approach is most suitable for the task of detecting traffic signs, providing aditional information and knowledge that can be expanded to other similar applications.