It is not always possible to just compare images pixel-wise and decide whether an object is present in an image or not, or whether an object has the expected shape or not, and many more similar scenarios that we can't even begin to list here. That is why the smarter way of interpreting the contents of an image is to look for meaningful features in it, and then base our interpretation on the properties of those features. In computer vision, a feature is synonymous with a keypoint, so don't be surprised if we use them interchangeably in this book. In fact, the word keypoint is better suited to describe the concept, since the most commonly used features in an image are usually key points in that image where there is a sudden change in color intensity, which can happen in corners and edges of shapes and objects in an image.

In this section, we'll learn about some of the most important and widely used keypoint-detection algorithms, namely the corner- and edge-detection algorithms that are the basis of almost all of the feature-based object detection algorithms we'll be learning about in this chapter.