Now that we know that keypoints refer to the interesting regions in the image, let's dig a little deeper. What are keypoints made of? Where are these points? When we say "interesting", it means that something is happening in that region. If the region is just uniform, then it's not very interesting. For example, corners are interesting because there is sharp change in intensity in two different directions. Each corner is a unique point where two edges meet. If you look at the preceding images, you will see that the interesting regions are not completely made up of "interesting" content. If you look closely, we can still see plain regions within busy regions. For example, consider the following image:

If you look at the preceding object, the interior parts of the interesting regions are "uninteresting".

So, if we were to characterize this object, we would need to make sure that we picked the interesting points. Now, how do we define "interesting points"? Can we just say that anything that's not uninteresting can be an interesting point? Let's consider the following example:

Now, we can see that there is a lot of high frequency content in this image along the edge. But we cannot call the whole edge "interesting". It is important to understand that "interesting" doesn't necessarily refer to color or intensity values. It can be anything, as long as it is distinct. We need to isolate the points that are unique in their neighborhood. The points along the edge are not unique with respect to their neighbors. So, now that we know what we are looking for, how do we pick an interesting point?

What about the corner of the table? That's pretty interesting, right? It's unique with respect to its neighbors and we don't have anything like that in its vicinity. Now this point can be chosen as one of our keypoints. We take a bunch of these keypoints to characterize a particular image.

When we do image analysis, we need to convert it into a numerical form before we deduce something. These keypoints are represented using a numerical form and a combination of these keypoints is then used to create the image signature. We want this image signature to represent a given image in the best possible way.