cnn

Ian Goodfellow, Yoshua Bengio and Aaron Courville in the book Deep Learning by CNN definition say that the name suggests that the network uses a mathematical operation called convolution. Convolution is a specialized kind of linear operation. Convolutional networks are simple neural networks that use convolution in the multiplication of the general matrix in at least one of their layers. In mathematics, convolution is a mathematical operation that uses two functions and produces a third function that expresses how the shape of the first is modified by the second. CNNs are widely known for their dramatic superiority over other network architectures in image processing. As I explained at the beginning of the chapter, ANNs consist of three layer topologies, the input layer, output layer, and the hidden layer which can consist of one or more layers. Here a little clarification regarding the term DL, when the architecture of a network uses more than 3 layers in the hidden layer topology then it can be defined as a DL network. CNNs in the hidden layer topology, have 3 main distinct layer types that are differentiated in terms of their function. These types are deployed in a sequential layering structure of convolutional (CONV), pooling, fully-connected (FC) layers.