Classification of Traffic Signs using Convolutional Neural Networks

Main Article Content

Dr. Shubhangi Vaikole, Makarand Bhalerao, Parth Nimbalkar, Soham Moghe

Abstract

Advancements in the field of Artificial Intelligence have optimized nearly everything in almost every industry. Automated and precise reliable models have been efficient in either assisting the humans, reducing the error of margin or completely taking over the tasks, speeding up processes and making them more efficient. Self-driving vehicles have been one of them. The following paper presents an approach towards a Traffic Sign Classifier with the help of Convolutional Neural Networks and has been tested on the standard dataset named as German Traffic Sign Recognition Benchmark (GTSRB)  consisting of a total of 51839 images. Adam optimizer has been used to decrease the overall loss and increase the accuracy of the neural network. Dropout regression has been employed to prevent overfitting.

Article Details

Section
Articles