Deep Learning 1: Foundations of Convolutional Neural Networks
Convolutional neural networks have three types of layers: Convolutional layer, Pooling layer, Fully connected layer. Usually, a convolutional layer and a pooling layer are called the same layer since a pooling layer doesn't have parameters remained to be fitted.
1. Convolutional layer
Parameters: weights and bias
Hyperparameters: padding , stride
, filter size
, filter number
-
Padding and stride
Convolution without padding has two problems:
- shrink output
- throw away information from the edge of the image
Stride decreases the displacement of images.
-
Convolutionover volume
An RBG image using a filter becomes an image with one channel.
-
Example: one convolutional layer
2. Pooling layer
Pooling layer can be considered as a max/average filter.
3. Fully connected layer
4. LeNet-5