Machine-Learning-CNN-Traffic-Sign-Classifier

• The main goal of this project is to learn Keras and Convoluted Neural Network including back-propogation, pooling and dropout.
• Model is designed to recognize and classify traffic signs.
• This model can be further used in automated driving cars.
• Dataset is taken from The German Traffic Sign Recognition Benchmark. http://benchmark.ini.rub.de/?section=gtsrb&subsection=dataset.
• For loading and precossing of images, code is adopted from https://chsasank.github.io/keras-tutorial.html.
• Used Amazon Web Services because of massive computations involved.