The newly released Galaxy Note 8 has received very positive reviews, which have made us forget the mess of the Note 7, at the root of one of the toughest crises ever for Samsung. But there is a problem: facial recognition, which is much less sophisticated than it is legitimate to think. Whether it's a fingerprint or a face, the reason why these unlocking systems are used is precisely their inseparable link with the user: if a password can be stolen, the same cannot be done with the voice or the features. Well, according to a test carried out by the developer and UI expert Mel Tajon, a photograph is enough to deceive the face recognition of the South Korean device. Which would be paradoxical, because at this point a simple password would be much more secure than resorting to an image found without too much effort, perhaps from the Facebook or Instagram profile of the rightful owner. Yes, because even low-resolution images, according to Tajon's experiment, are enough to unlock the phone.
Of course, facial recognition is not mandatory, and it is possible to safeguard the security of your Note 8 using your fingerprint, retina or the classic password. The iPhone 8 (or iPhone Edition: the name will be known only at the time of its presentation, next September 12) has instead focused a lot on face recognition, even with a considerable delay compared to the competition - according to the malicious, precisely to leave the missteps to others in such immature technology. Android phones for at least six years have boasted a smartphone unlocking system via face recognition, but the risk, as we have seen, is that cases like that of the Note 8 will occur. Apple has acquired two companies in this sector, as it is now used to do when it identifies promising startups: RealFace and PrimeSense, established thanks to the technology developed for the Kinect of Microsoft's XBox. Given that, according to experts, the fingerprint recognition system on new iPhones is the most advanced currently on the market, it is likely that even for those who want to access the contents of their device by showing their face, Apple has in store, if not other, a safer and more intuitive approach. Proof of this is the fact that facial recognition is not considered reliable by Samsung itself, which considers it more of a convenience for the user but not an alternative to "real" security systems, so much so that it is not possible to use it for Samsung Pay payments; on the contrary, net of surprises, Apple seems to intend to link this technology to payments made through Apple Pay, demonstrating a notable faith in the inviolability of the system, a negative episode would be enough to cause a real disaster in terms of communication and marketing. However, the prospects for facial recognition are many, and certainly not limited to device security or access control. Disney, for example, is experimenting with a deep learning approach to artificial intelligence that can record viewers' reactions to a film. The approach, based on Factorized variational autoencoders (Fva), after a few minutes of observation allows us to predict with surprising accuracy - according to the Disney engineers - the facial expressions of the viewer for the rest of the film. This is a typical big data problem (in total 16 million points were collected from 3.179 viewers) where the deep learning approach, which exploits neural networks, makes it possible to automatically identify significant patterns. For an optimist, this could be a great way to guide studios to bring more interesting stories to the screen, evaluating the onset of emotions that conflict with what was expected; on the other hand, however, the risk is to sacrifice creativity and the unexpected in the name of a sure laugh or a tear at the right time. However, it is not the first time that the visual entertainment industry has focused on facial recognition: the AdBeacon technology of the American AdMobilize, for example, has developed a system capable of interpreting the face of those who look at a screen and derive variables, such as race, gender and age, in order to then transmit advertising content tailored to the target. Faces for social media and payments: the Polygram and KFC cases Polygram is, on the other hand, a new social network - yet another attempting to break into an already largely saturated market - based precisely on facial recognition: the smartphone camera of a member identifies the expression and suggests an emoji (or animated gif) to use as a comment. Thanks to this technology, Polygram promises more possibilities for special effects, Snapchat (dog ears and nose or the like that we have all experienced sooner or later) thanks to more advanced graphic algorithms.
Welcome to the future!