What is a cascade of regressors?

Question

Python

Learning to work with dlib . It turned out to train the machine to find and recognize the desired object.
Now I want to study the algorithm itself, which is used at the heart of the library. It is described in detail in this document.
From there, it is not entirely clear to me that there is a cascade of regressors. For now, I imagine it like this: the cascade is a certain section of the image, the regressor is a guide that indicates the direction to the next cascade.
An explanation with a concrete example is highly desirable.