The goal of this research is to identify and recognize traffic signs for autonomous vehicles in real-time. We will use the Deep Learning, multi-layer Convolutional Neural Network, and Keras libraries to design a model that categorizes the traffic signs in the dataset into different groups. Later with the help of GUI (Graphical User Interface) model will be used to detect and recognize the uploaded images and predict the symbol using speech output.

The dataset is based on German traffic signs that contains more than 50,000 images of different traffic signs available on the road. There are 39,209 training images and 12,630 test images (Fig.2) in the German Traffic Sign Recognition Benchmark (GTSRB) dataset, which consists of 43 classes of traffic signs. Some of the images are taken under varying light conditions such as blurred, poor visibility, deformed to train the model under different circumstances.