Now that we have an intuition as to what the LBP captures in our image, let's try to apply this to the facial images that we have been working with. But, we aren't simply going to create the histogram of the (uniform pattern) LBP codes like we have been doing so far. We are going to make use of the spatial information that is inherent in faces.
What does exploiting spatial information in faces really mean? When you look at the image of a face, you expect to see some facial features at some distinct locations of the face. For example, the nose will almost be at the center, with the eyes on either side, the chin would cover the lower part of the face image, and so on. Now, when you create a histogram (of the LBP codes) out of the entire face image, you lose such spatial information. And how exactly are we losing spatial information? Consider two pixels-one of which is present near the left eye and the other near the center of the lower lip. If they have...
