CNN Study Notes

Basic concepts

Convolution core: That is, the filter, is a nxn coefficient matrix (n is odd), the function is to extract some characteristics of the input image, and then the resulting features output to the neuron.

In convolutional networks, the number of parameters of the hidden layer is only related to the size of the filter and the type of filter, regardless of the number of hidden neurons

The number of neurons in the hidden layer is related to the original image, the size of the filter and the sliding step in the image.

CNN Study Notes