Traffic and road sign recognition using deep convolutional neural network

Abstract

Automatic road sign recognition is one of the most important steps to help a driver prevent accidents. In this paper, the deep convolutional neural network (Deep CNN) was used for the autonomous traffic and road signs detection and recognition system. The proposed system works in real-time recognition of the image of road signs. The contribution of this article is that it consists of a newly developed datasets for 42 different road signs using combination of existing datasets and collected local road signs. The images were taken from different angles and included other parameters and conditions. More than 30K images were collected to create the dataset, which is named Uzbekistan traffic and road signs (UZ-TRS-2021). The CNN architecture was used with various parameters to achieve the best recognition speed. Experimental results show that the proposed CNN architecture achieved 98% accuracy.