/Traffic-Sign-Recognition

Shallow network with combined pooling for fast traffic sign recognition (Zhang et al)

Primary LanguagePython

Traffic-Sign-Recognition

Shallow network with combined pooling for fast traffic sign recognition (Zhang et al)

Overview:

A Deep Learning model that utilizes Convolutional Neural Network (CNN) along with dropout and multi scale features to identify traffic signs in images. The dataset used is the German Traffic Sign Dataset. The trained model achieves around 96% accuracy on the test set. Details given in report.pdf

Dependencies:

Python
Tensorflow
cPickle

Running the project:

  1. Download the dataset from here and put it in the root folder
  2. Setup Tensorflow on local machine
  3. Run "python tsr.py" from root directory