Deep learning for Scalable Object Detection

In this section, we will learn how to make image recognition using TensorFlow. Also, we will be using transfer learning which is the ability to use the network weights of a pre-trained model that was trained on large dataset like ImageNet), usually people use transfer learning when they have small datasets. So we will be starting with a pre-trained model and use it on another problem. We will retrain this model on a similar problem because if we started from scratch then it will take days.

Figure 1: Flowers dataset (Image by TesnorFlow, Source: https://www.tensorflow.org/images/daisies.jpg)

There are lots of pre-trained networks that you can start using and they already packed with TensorFlow. In this section, we will be using Inception V3 network which is trained for the ImageNet (http://image-net.org/ ). This network can differentiate between 1000 different classes.

Initially, you'll need a set of images to teach the network about the new classes...