Convolutional neural network learning summary (1)

Today I learned about CNN (Convolutional Nerual Netwaor) convolutional neural network.

Understand the learning process of convolutional neural networks: first extract the features of pixels, extract the edge features of the object, then extract the local features of the object (such as human eyes, human ears, nose, etc.), and finally get the object model.

The common building methods for convolutional neural networks are: 1. Input picture 2. Perform a layer of convolution 3. Perform maximum pooling 4. Perform second convolution 5. Perform second maximum pooling 6. Perform two full Connection 7. Classification.