US researchers have announced that, using a sophisticated artificial intelligence system, they have created convincing artificial fingerprints that can mimic real ones. They presented the artificial fingerprints called “DeepMasterPrints” at a security conference in LA.
The technique uses a generative adversarial network (GAN) to learn the underlying structure of fingerprints, and then generates a synthetic fingerprint that can be used to fool fingerprint recognition systems.
DeepMasterPrints can be used to create a “master print” that can unlock multiple fingerprints that are similar but not identical. This is because the GAN learns the underlying structure of fingerprints and generates a synthetic fingerprint, used to fool fingerprint recognition systems. The synthetic fingerprint is to create a “master print” that can unlock multiple fingerprints that are similar but not identical.
The researchers tested the technique on a dataset of fingerprints and found that it was able to fool a fingerprint recognition system with a success rate of about one in four attempts. They also found that the success rate increased with the synthetic fingerprint used in combination with a real fingerprint.
The researchers note that the technique is not foolproof. It is likely that fingerprint recognition systems will be able to adapt and improve their defenses against DeepMasterPrints. However, the technique highlights the need for robust fingerprint recognition systems that are able to adapt and defend against attacks.
Did you like that story? In that case you may also find this one interesting.
Author: PC-GR
The World of Technology
