CNN

A convolutional neural network (CNN) is a feed-forward neural net, with the input being an image. The architecture of the net comprises of convolutional layers, pooling layers, activation layer, and fully connected layers.

A series of convolution operations take place at every layer, which extrapolates the pertinent information from the images. At every layer, multiple convolution operations take place, followed by zero padding and eventually passed through an activation layer and what is outputted is an Activation Map.

There are many standard architectures which work great for many standard problems. Examples being AlexNet, GoogleNet, InceptionResnet, VGG etc.