If someone came up with a specific method of embedding a physical watermark in infills which was easily detectable via nondestructive methods (i.e, without sawing the object in half -- perhaps something that would leave a pattern on X-ray or ultrasound?), but which didn't significantly affect the model's density, that would be nifty and likely patent-worthy. (In the best possible way: it'd necessarily be innovative, and would only cover the specific method described.)
The method described in the patent, on the other hand, is largely uninnovative -- it simply describes a "phone-home" licensing scheme. It doesn't even appear to describe the "object code" as being encrypted, which would make it rather trivial to bypass.