Introduction

In the previous chapter, we discussed Haar Cascades and how to use them for object detection problems. We also used a variety of cascade classifiers for problems such as eye detection, people counters, and others. In this chapter, we will take this one step further and learn how to track the objects that have been detected. This tracking can be done for a pre-recorded video or a live video stream.

Object tracking is quite easy to understand. You have an object that either you have selected manually or was selected by some other algorithm, and now you want to track the object. The beauty of object tracking comes into the picture when we understand the difficulty of carrying out this very simple task in different scenarios. To understand this better, think of a scenario where you are with your friend in a busy market. Among a huge crowd, you need to detect and recognize your friend the entire time, and also need to keep track of where he or she is so that you don&apos...