/Roadsign-Recognition

TensorFlow CNN that can classify 43 types of traffic signs with 94% accuracy

Primary LanguagePython

Roadsign-Recognition

This brief project was created in order to explore applications of machine learning to autonomous vehicle planning. A TensorFlow convolutional neural network model learned to classify 43 different types of traffic signs with about 94% accuracy. Training and testing data was provided by the German Traffic Sign Recognition Benchmark (available in this repo's 'gtsrb' folder).