Here’s the paper where they explain it. Basically, they make subtle fluctuations in layer height, adding or subtracting small amounts that are not visible to the naked eye, to encode 0s and 1s. So, maybe in principle this could run at the firmware level on your printer. Then, someone can use a microscope to read off the code from pieces of the printed part.
I would have some doubts about how reliable this is, given the relatively large tolerances in fdm printing, but they have a section about that in the paper, so I guess they at least have thought about it.
Fluctuations in layer height that are not visible? Dude, most printers are not even able to achieve a layer height consistency that would be invisible to the naked eye if they wanted to.
Plus what about adaptive layer heights that automatically change based on the features of each layer?
Or just acetone/MEK/whatever smoothing your parts?
So when I turn on variable layer height this whole thing goes out the window. Or if I post process and sand the parts. Oh I know what if I use a slicer and firmware like klipper to bypass everything. No 3d printer that isn’t locked down already by closed ecosystem will never have this “code” in the parts.
Sand paper wouldn’t really help. They could just cut a part in half. But yeah no way in hell this will ever end up in any of the open source printer firmwares. So it’s a moot point.
The fact that this could only work in 100% locked down ecosystems was my thought too. About cutting the part up, that seems to be what this particular paper is most proud of: they did a bunch of math to make some codes that they could still figure out even when they were cut into pieces and mixed up—like if a person broke their printed part after using it. Sort of like error-correcting codes I guess, but able to be reassembled from fragments.
Why would you cut it into pieces. Just burn it i’d say…
you know how easy it is to circumvent this?
relatively easy, its relatively easy. now you know.