In this chapter, we started by showing how image classification models can be created using standard machine-learning techniques; however, this has limitations as the images get larger and more complex. We can use convolutional neural networks to combat this issue. Using this approach, we demonstrated how we could perform dimensionality reduction and make it more computationally efficient to train a classification model on image data. We built a model with one convolution and pooling layer and then showed how we could make the model even deeper by adding further layers. Lastly, we used dropout layers and early stopping to avoid overfitting our model. Using all of these tactics in concert, we are now able to build models for classifying any type of image data.
In the next chapter, we will learn how to code a multilayer perceptron. The multilayer perceptron is a feedforward...
