We have described how the LBP operator is applied to images, and we have also discussed some variants of it. Now obviously, if we are studying LBP in such great detail, we would definitely be applying it somewhere! We are going to use the LBP operator on the (cropped and aligned) facial images that we obtained in the last chapter. But before we jump in and start running the LBP code on our face images, let's take a step back and ponder upon a very important question, "What does the LBP capture?"
We have already gone through the mechanics of calculating the LBP code for a pixel, and we have seen that the end result is a 256 (or a 58) dimensional histogram. But, what does that histogram tell us about the image? Well, you can say that histograms give us frequency counts and you would be right. It was easy to visualize this in the case of the image histograms from Chapter 4, Image Histograms where the histogram bins were the grayscale values. Hence, it was easy to intuitively...