Practical Convolutional Neural Networks
上QQ阅读APP看书,第一时间看更新

Convolutional layer

The main objective of convolution in relation to ConvNet is to extract features from the input image. This layer does most of the computation in a ConvNet. We will not go into the mathematical details of convolution here but will get an understanding of how it works over images.

The ReLU activation function is extremely useful in CNNs.