This article is part of a series of articles Figure neural networks, article directories as follows:
- From FIG. (Graph) to FIG convolution (Graph Convolution): Talk FIG neural network model (a)
- From FIG. (Graph) to FIG convolution (Graph Convolution): Talk FIG neural network model (II)
- From FIG. (Graph) to FIG convolution (Graph Convolution): Talk neural network model of FIG. (C)
In the previous blog , we simply introduce two important cycle diagram model based on neural networks, In this, we will introduce a lot of ink convolution operation Figure convolution neural networks . Next, we will first introduce about frame chart convolution neural network, need to show the difference between it and the neural network-based map of the cycle. Then, we will start from scratch introduces the reader to the basic concepts of convolution, and its meaning in the physical model. Finally, we will introduce two different convolution operation in detail, namely spatial convolution and time domain convolution , corresponding classical model. There does not need any reader basic signal processing, the Fourier transform concepts are described in detail herein.