would a microscope "fingerprint" for watches actually work? looking for holes in my idea
Superclones are getting scary good. Serials get copied off, cards and nfc tags can be faked or moved, and frankenwatches can slip through a lot of peoples guard. Most checks are on the paperwork, not the watch itself, and even when some preowned dealer opens the case-back, modern super clones are genuinely getting so good.
I'm 19, doing mathematics/CS at University of Sydney, and ive been working on a different approach. Under a microscope every watch has tiny random marks, like the edges of the rehaut engraving, dial printing, randomness in Geneva striping, other parts of the movement etc. Two identical refs look the same but at that level theyre as different as fingerprints, and a cloner cant copy marks on a watch theyve never seen. I have tested this on three Rolex datejusts, 3235 movement references.
How it works: a small microscope camera photographs a few fixed spots through the crystal, nothing opened or via opening the case back and taking key images there, and software saves them against the serial. When the watch is resold or serviced, someone photographs the same spots and the AI algorithm confirms it's the exact same watch.
I got the software working, testing on a few datejusts now as stated. Not selling anything.
What am i missing, and what would make you not trust it? Would this be beneficial enough for a pre-owned dealer because at the current rate, super clones are going to continue to improve, so maybe we should not rely on human authentication as much. Any advice would be awesome!
