Traffic-Sign-Recognition-using-CNN

This work is a implementation of CNN model for Traffic Sign Recognition. The fun part of this work involves network pruning by analyzing the contribution of the nodes to the final outcome. Determining the appropriate number of layers and number of hidden nodes within a layer is ambigious and involves a huge amount of trail and error. To some extend, neural network pruning is a pretty good alternative by getting rid of the redundant part of the Neural Network.