arubiales/traffic-signals-tf

Traffic Signal Recognition

What are we going to do?

In this repository we are going to deploy Deep Learning models from scratch (almost from scratch) with Tensorflow. We have a dataset with 62 signal classes and we have to predict it. We are going to use the following models:

1. Artificial Neural Network
2. Convolutional Neural Network
3. Convolutional Neuroal Network (advanced)

This is a very practical Notebook, to develop deep learning models in Tensorflow from Scratch, there are not technicals or mathematics explications. And obviously you can train an image classification neural network faster and better if you do with Keras or Tensorflow Layers

Repository

• BelgiumTSC_Testing and BelgiumTSC_Training are the train and the test sets, you can download from here. But don't worry about it, i do the download process in the notebook :)

• Models_from_scratch: here we develop the three models.