Convolutional neural network

Convolutional neural network (CNN) is an artificial neural network consisting of multiple layers also known as neurons. A typical structure of the CNN contains convolution, nonlinearity, subsampling repeatedly connected with fully connected layers. Modern CNN framework consists of a large number of layers (deep layers) containing convolutional and subsampling layers followed by one or more fully connected layers.
Convolution is a mathematical operation on two functions that produces a third
function, which is the integral of the product of the two functions with one of them
flipped. Convolution is the main building block of the CNN. It is considered as
sliding a filter along its width and height. The convolution filter with desired kernel
size is applied in input data resulting in the feature maps. Convolution is done at
each point of input without overlap. Convolution filters are widely used in image
processing, digital signal processing.