/Traffic-Sign-Classification

CNN model built using TensorFlow and Keras for the purpose of detection of traffic signs.

Primary LanguageJupyter Notebook

Traffic-Sign-Classification

CNN model built using TensorFlow and Keras for the purpose of detection of traffic signs.

Setting up

  1. Download the GTSRB dataset from here.
  2. Place the above files in the folder just downloaded, containing train, test, and meta folders.
  3. Run the notebook to generate the Keras model(95% accuracy can be achieved with pre-made layers).
  4. The last block of the notebook can be used to generate a TensorFlowLite model for use in mobile devices.

Feel free to use the model on different datasets, and do share your work!