If you are familiar with the basics of machine learning, you will certainly know what supervised and unsupervised learning is all about. To give a quick refresher, supervised learning refers to building a function based on labeled samples. For example, if we are building a system to separate dress images from footwear images, we first need to build a database and label it. We need to tell our algorithm what images correspond to dresses and what images correspond to footwear. Based on this data, the algorithm will learn how to identify dresses and footwear so that when an unknown image comes in, it can recognize what's inside that image.

Unsupervised learning is the opposite of what we just discussed. There is no labeled data available here. Let's say we have a bunch of images, and we just want to separate them into three groups. We don't know what the criteria will be. So, an unsupervised learning algorithm will try to separate the given set of data into 3 groups in the best possible way. The reason we are discussing this is because we will be using a combination of supervised and unsupervised learning to build our object recognition system.