Traffic Signs Identification by Deep Learning for Autonomous Driving

• DOI: 10.1049/cp.2018.1382
• Publication Date: 2019-03-11T10:08:15Z
• Authors: W. Farag; Z. Saleh
• Keywords: Identification

In this paper, we have proposed and developed a comprehensive Convolutional Neural Network (CNN) classifier “WAF-LeNet” to be used in traffic signs recognition identification that empowers the autonomous driving technologies. The implemented CNN architecture is deep fifteen-layer network has been selected after extensive trials. got trained using Adam's optimization algorithm based on scholastic gradient descent technique. learning process carried out well-known “German Traffic Sign Dataset”. data partitioned into training, validation testing sets. Additionally, more random images are collected from web further test robustness of classifier. paper goes through development details shows image processing pipeline harnessed development. approach proved successful identifying correctly 96.5% set 100% set.