Implementation of CNN using Keras for a ImageClassifier. We use a similar model as the one exposed by Yann Le Cun. A triple stack of Conv + MaxPool layers, a Flatten layer previous to the Dense (64) layer, a Dropout of 50% and a final Dense layer with a Softmax function of two encoded classes. Conv and Dense (64) layers use ReLu activaction functions and we use the last version of the Adam optimizer for compiling.

The idea is to use pre-classified images to train a Convolutional Neural Network (CNN) in order to classify later images.

In order to run the project you need:

- Python 2.7 or 3.5
- Tensorflow
- Keras