Research on the Optimal Machine Learning Classifier for Traffic Signs

University of Macau, Faculty of Science and Technology, Macau, 999078, China

^* Corresponding author. Email: db92823@umac.mo

Abstract

Now autonomous driving is a hot topic, and the identification of traffic signs is also extremely important for autonomous driving. This paper mainly compares the difference of the Support Vector Machine (SVM), Multilayer Perceptron (MLP), and Logistic Regression (LR) Classifier in the traffic sign classification. The effect of the initial image processing on classification accuracy is also studied. The paper found that sharpening the image significantly improved the accuracy of the image classification. Based on the results of various situations, the author found that, in this paper, SVM is the classifier with the best classification effect, but the effect of LR classifier is not much worse than that of SVM when the image is sharpened.