Convolutional Neural Network: II

The last chapter discussed the units of a Convolutional Neural Network and introduced LeNet. This chapter takes the discussion forward and presents an overview and implementation of some of the famous CNN architectures like LeNet, AlexNet, and Google LeNet (Inception Net). The simplicity of LeNet gives a good idea of how things work in CNN. However, to classify complex images and to accomplish advanced image analysis tasks, we need deep, more complex structures. The advancements in the 2010s were aimed at handing the problems in the then-popular CNNs and gave us the architectures that have since become immensely important for all image-related tasks: both supervised and unsupervised.