Deep Learning with Microsoft Cognitive Toolkit Quick Start Guide

Network architecture used for image classification

Convolutional networks used for image classification typically contain one or more convolution layers followed by pooling layers, and usually end in regular fully-connected layers to provide the final output as shown in the following screenshot:

This image is from: https://en.wikipedia.org/wiki/File:Typical_cnn.png

When you take a closer look at the structure of a convolutional network, you'll see that it starts with a set of convolution and pooling layers. You can consider this part a complex, trainable photo filter. The convolution layers filter out interesting details that are needed to ...