Showing posts with label adversarial attacks. Show all posts
Showing posts with label adversarial attacks. Show all posts

Sunday, February 25, 2024

Your fingerprints can be recreated from the sounds made when you swipe on a touchscreen — Chinese and US researchers show new side channel can reproduce fingerprints to enable attacks

Bad news! Well it is not actually fingerprints, but sounds from swiping!

"An interesting new attack on biometric security has been outlined by a group of researchers from China and the US. PrintListener: Uncovering the Vulnerability of Fingerprint Authentication via the Finger Friction Sound [PDF] proposes a side-channel attack on the sophisticated Automatic Fingerprint Identification System (AFIS). The attack leverages the sound characteristics of a user’s finger swiping on a touchscreen to extract fingerprint pattern features. Following tests, the researchers assert that they can successfully attack “up to 27.9% of partial fingerprints and 9.3% of complete fingerprints within five attempts at the highest security FAR [False Acceptance Rate] setting of 0.01%.” This is claimed to be the first work that leverages swiping sounds to infer fingerprint information. ..."

Your fingerprints can be recreated from the sounds made when you swipe on a touchscreen — Chinese and US researchers show new side channel can reproduce fingerprints to enable attacks | Tom's Hardware Researchers claim they can successfully attack up to 27.9% of partial fingerprints.


Credits: Last Week in AI

Tuesday, November 05, 2019

How 'adversarial' attacks reveal machine learning's weakness

Food for thought, but no reason to panic or be alarmist! It is by now well understood why adversarial attacks work and it is only a matter of time before most of these attacks will be defeated!

To publish research papers on adversarial attacks is almost trivial or like a quick and easy/lazy hit job to get published! I am not sure why these papers are still accepted in such large numbers at prestigious conferences! Is it a fad?

How 'adversarial' attacks reveal machine learning's weakness | VentureBeat: Adversarial attacks are emerging as a major threat to progress in computer vision and machine learning as researchers race to find a defense.