Building and Comparing AI-Powered Algorithms in Road Sign Detection

In this paper we conducted research with AI-Powered algorithms in order to build a model for image classification with high accuracy and compare it to existing studies. The neural network used is Convolutional Neural Network (CNN), the model was built using German Traffic Sign Benchmark dataset where we added different characteristics of our own to improve the accuracy. We train our model for a limited number of epochs, all the while checking the values of accuracy and loss and comparing the performances with each epoch. During the training time our model is getting better through forward propagation and backpropagation. Our end-goal of our model is to be trained well-enough to detect features, we achieve our goal of an acceptable high-accuracy rate. We compare our results with another study where the road sign detection is done with a predictive filter solution using SVM with gaussian kernel. Copyright (c) 2022 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)