Overview

Convolutional Neural Networks are neural networks that are configured to be used with images. convolution

Data Format

Images are rectangular arrays of pixels. They can be visualized using a simple picture.


The image here is a 3 pixel by 3 pixel image. However, neural networks require a vector input. Therefore, image data is usually reshaped into a vector.

Layer Types

• Convolution Layer
• Pooling Layer

Structure

The typical structure of a convolutional neural network is to feed the inputs to a series convolutional layers, each followed by a max pooling layer and then finally passing to a set of fully connected neural network layers.