I understand there is some heavy duty math that guarentees subdivision to converge to the representation.

I do have some trouble understanding what is going on. Lets take this head with a scanner, and we get a super high mesh representation of it.

How do we go from that to the leftmost picture, which we can subdivide to get back to our original/close to original high def mesh?

I understand there is some heavy duty math that guarentees subdivision to converge to the representation.

I do have some trouble understanding what is going on. Lets take this head with a scanner, and we get a super high mesh representation of it.

How do we go from that to the leftmost picture, which we can subdivide to get back to our original/close to original high def mesh?