Face ID is a biometric authentication facial-recognition system designed and developed by Apple Inc. for the iPhone and iPad Pro. The system can be used for unlocking a device,[1] making payments, accessing sensitive data, providing detailed facial expression tracking for Animoji, as well as six degrees of freedom (6DOF) head-tracking, eye-tracking, and other features. Initially released in November with the iPhone X, it has since been updated and introduced to all iPhones outside of SE models and all iPad Pro models from onwards.[2] Users on iOS 18 and newer can choose to lock specific apps, requiring Face ID to access them.[3]
Sihovision are exported all over the world and different industries with quality first. Our belief is to provide our customers with more and better high value-added products. Let's create a better future together.
The Face ID hardware uses a TrueDepth Camera[2] that consists of a sensor with three modules; a laser[4] dot projector that projects a grid of small infrared dots onto a user's face, a module called the flood illuminator that shines infrared light at the face, and an infrared camera that takes an infrared picture of the user, reads the resulting pattern, and generates a 3D facial map.
Face ID has sparked a number of debates about security and privacy. Apple claims that Face ID is statistically more advanced than Touch ID fingerprint scanning.[5] It exhibits significantly fewer false positives. Multiple security features are in place to limit the risk of the system being bypassed using photos or masks, and only one proof-of-concept attempt using detailed scans has succeeded.
Debate continues over the lack of legal protections offered by biometric systems as compared to passcode authentication in the United States. Hackers have been able to use combinations of FaceID data and SMS messages to enter various locked information on Apple users iPhones protected by FaceID technology. Privacy advocates have also expressed concern about third-party app developers' access to "rough maps" of user facial data, despite rigid requirements by Apple of how developers handle facial data. Privacy concerns also exist regarding the use FaceID data to retrieve other personal information stored on Apple technology.[6] Use of FaceID technology and biometric data in criminal cases as been of much debate due to lack of legal regulation. FaceID has been compared to fingerprint and passcode locking mechanisms to evaluate the ethics behind use of FaceID in criminal cases. Finally, infiltration on Apple products has been a concern of the public as twins and close relatives have been successful in fooling the FaceID technology. Facial replication into realistic masks has been an infiltration concern, but has thus far been unsuccessful.
With the onset of the COVID-19 pandemic, it was noted that Face ID was unable to recognize users wearing face coverings on some devices.[7][8] Apple responded to criticism by offering faster fallback to passcode input, and the option for Apple Watch users to confirm whether they intended to unlock their iPhone.[9] In March , Apple released iOS 15.4 which adds mask-compatible Face ID for iPhone 12 and later devices.[10]
Apple announced Face ID during the unveiling of the iPhone X on September 12, .[11] The system was presented as the successor to Touch ID, Apple's previous fingerprint-based authentication technology embedded in the home button of the iPhone 8 and earlier devices in addition to the second and third-generation iPhone SE.[12] On September 12, , Apple introduced the iPhone XS and XR with faster neural network processing speeds, providing a significant speed increase to Face ID. On October 30, , Apple introduced the third generation iPad Pro, which brings Face ID to the iPad and allows face recognition in any orientation.[13] iOS 13 included an upgraded version of Face ID which is up to 30% faster than Face ID on previous versions.[14]
In February 19, , following the announcement of the iPhone 16e and the discontinuation of the iPhone SE (3rd generation), all the iPhone models have transitioned from the Touch ID to Face ID.[15]
Face ID's technology is based on PrimeSense's previous work with low-cost infrared depth perception that was the basis of the Kinect motion sensor for the Xbox console line from Microsoft; Apple had acquired PrimeSense in after Microsoft started to wane on the use of Kinect.[16][17]
Face ID is based on a facial recognition sensor that consists of two parts: a vertical-cavity surface-emitting laser dot projector module that projects more than 30,000 infrared dots onto the user's face, and an infrared camera module that reads the pattern.[18] The pattern is projected from the laser using an Active Diffractive Optical Element which divides the beam into 30,000 dots.[19]
The TrueDepth Camera [2] uses infrared light to create a 3D map of users unique facial identity. This map is compared with the registered face using a secure subsystem, and the user is authenticated if the two faces match sufficiently. The system can recognize faces with glasses, clothing, makeup, and facial hair, and it adapts to changes in appearance over time. Concerns regarding the safety of longterm infrared facial screening has been debated and studied.[20]
The pattern is encrypted and sent to a local "Secure Enclave" in the device's CPU to confirm a match with the registered face.[21][22] The stored facial data is a mathematical representation of key details of the face, and it is inaccessible to Apple or other parties.[21] To avoid involuntary authentication, the system requires the user to open their eyes and look at the device to attempt a match, although this can be disabled through an accessibility setting.[21] Face ID is temporarily disabled and the user's passcode is required after 5 unsuccessful scans, 48 hours of inactivity, restarting the device, or if both of the device's side buttons are held briefly.[23]
Apple claimed the probability of someone else unlocking a with Face ID is 1 in 1,000,000 as opposed to Touch ID at 1 in 50,000.[12][24] During initial setup, the user's face is scanned twice from a number of angles to create a complete reference map. As the system is used, it learns about typical variations in a user's appearance, and will adjust its registered face data to match aging, facial hair growth, and other changes using the Neural Engine. The system will recognize a face wearing hats, scarves, glasses, most sunglasses,[25] facial hair or makeup.[26] When significant facial changes occur, FaceID may not recognize the person when comparing the image to stored data. In such cases, the user will be prompted to verify using their passcode and the facial recognition data will update to the changes.[2] It also works in the dark by invisibly illuminating the whole face with a dedicated infrared flash module.[27]
Authentication with Face ID is used to enable a number of iOS features, including unlocking the automatically on wake, making payments with Apple Pay, and viewing saved passwords. Apps by Apple or third-party developers can protect sensitive data with a system framework; the device will verify the user's identity and return success or failure without sharing face data with the app. Additionally, Face ID can be used without authentication to track over 50 aspects of a user's facial expression and positioning, which can be used to create live effects such as Animoji or camera filters. In recent years, third-party developers have developed more use cases for FaceID such as e.g. Eyeware Beam, an iOS app that provides a reliable and precise, multi-purpose head and eye-tracking tool. It is used to enable control of the camera angle through head-motion-in games and eye-tracking to share attention with audience in streams, but also augmentative and alternative communication (AAC) and biometric research.[28]
Face ID uses an infrared flood illuminator and laser infrared dot projector. Prolonged exposure to infrared light has been known to cause harm to skin and the eyes. Apple has stated that the output is low enough that it will cause no harm to the eyes or skin, and meets 'international safety standards'.[20] FaceID infrared output has been compared to what is put out from TV remotes.[20] They do not, however, recommend the sensor be repaired by third parties, citing security concerns. There is also an in-built feature to deactivate Face ID should unauthorized components be found.[29]
Inconsistent results have been shown when testing Face ID on identical twins, with some tests showing the system managing to separate the two,[30] while other tests have failed.[31] The system has additionally been fooled by close relatives.[32] Apple states that the probability of a false match is different for twins and siblings, as well as children under 13 years of age, as "their distinct facial features may not have fully developed".[33]
Face ID has raised concerns regarding the possibility of law enforcement accessing an individual's by pointing the device at the user's face.[34] United States Senator Al Franken asked Apple to provide more information on the security and privacy of Face ID a day after the announcement,[35] with Apple responding by highlighting the recent publication of a security white paper and knowledge base detailing answers.[36][37]
In August , the FBI obtained a warrant to search the property (which includes electronic devices) of a man, Grant Michalski, accused of transmitting child pornography; they unlocked the suspect's iPhone by holding it up to his face, without needing his passcode.[38]
The Verge noted that courts in the United States have granted different Fifth Amendment rights to keycode and biometric unlocking systems. Keycodes are considered "testimonial" evidence based on the contents of users' thoughts, whereas fingerprints are considered physical evidence, with some suspects having been ordered to unlock their phones via fingerprint.[39] Debates are ongoing regarding the use of FaceID technology in law enforcement and criminal cases as laws and regulations have not yet been put in place. Currently, law enforcements' need for evidence is to be balanced with individual privacy rights in regards to biometric data.
Verification experts claim that if biometric technology does not account for skin texture or blood flow, sophisticated masks may be successful in bypassing FaceID technology. However, many people have attempted to fool Face ID with sophisticated masks, though most have failed.[40] In November , Vietnamese security firm Bkav announced in a blog post that it had created a $150 mask that successfully unlocked Face ID, but WIRED noted that Bkav's technique was more of a "proof-of-concept" rather than active exploitation risk, with the technique requiring a detailed measurement or digital scan of the iPhone owner's face, putting the real risk of danger only to targets of espionage and world leaders.[41][42]
If the user explicitly grants a third-party app permission to use the camera, the app can also access basic facial expression and positioning data from Face ID for features such as precise selfie filters such as those seen in Snapchat, or game characters mirroring real-world user facial expressions. The data accessible to third parties is not sufficient to unlock a device or even identify a user, and Apple prohibits developers from selling the data to others, creating profiles on users, or using the data for advertising. The American Civil Liberties Union and the Center for Democracy and Technology raised privacy questions about Apple's enforcement of the privacy restrictions connected to third-party access, with Apple maintaining that its App Store review processes were effective safeguards. Jay Stanley, a senior policy analyst with the ACLU, has stated that the overall idea of letting developers access sensitive facial information was still not satisfactorily handled, with Stanley telling Reuters that "the privacy issues around of the use of very sophisticated facial recognition technology for unlocking the have been overblown. The real privacy issues have to do with the access by third-party developers".[43][44]
Mobile hackers have been able to combine data from FaceID and SMS one-time verification codes to access information from other accounts of Apple devices. Bank accounts of users in Asia and the Pacific Islands have been breached in isolated attacks by mobile hackers using FaceID data.[6] Hackers were able to use facial images, stored in FaceID data, to make deepfake images that open information secured on Apple users devices with FaceID security. Combining FaceID data and one-time SMS verification codes gave hackers access to various FaceID protected accounts. [6] While these cyber attacks have been isolated in Asia and the Pacific Islands, it raises concerns about the security of FaceID technology.
On the dark web, users have been selling their personal FaceID images and identity for acute financial gain. Hackers are using identities bought off the dark web for "sophisticated impersonation fraud".[45] This type of "hacking" or fraud is extremely difficult to detect because the FaceID documents submitted are real, as they are coming directly from the user, meaning they match biometric FaceID data almost perfectly.[45]
iProov, a biometric data verification service, has suggested various ways to prevent FaceID biometric data from being successfully used to hack user data. "Embedded imagery and meta data analysis" can be used detect if FaceID images are a real person or a media image.[45] Technology can be used to quickly detect and respond to threats on verification systems using FaceID through ongoing monitoring and proactively searching for threats. With adequate training, engineers can learn how to reverse potential hacking situations to better understand how to prevent them.[45] Without understanding how hackers are using FaceID biometric data to bypass verifications, ways to prevent cyber attacks can better be initiated to protect users.
During the COVID-19 pandemic, face masks were employed as a public and personal health control measure against the spread of SARS-CoV-2. Face ID at the time was incompatible with face masks, with Apple stating "Face ID is designed to work with your eyes, nose and mouth visible."[46] With the release of iOS 13.5, Apple added a feature that automatically brought up the passcode screen if it detected that the user was wearing a mask.[47][48] Apple was criticized for not addressing these issues with the release of the iPhone 12, but was praised for the lack of inclusion of Face ID in favor of Touch ID integration into the power button on the fourth-generation iPad Air.[7][49] In April , Apple released iOS 14.5 and watchOS 7.4 with an option to allow Apple Watch to act as a backup if Face ID fails due to face masks.[9] In March , Apple released iOS 15.4 which adds mask-compatible Face ID for iPhone 12 and later devices.[10]
Facial recognition—the software that maps, analyzes, and then confirms the identity of a face in a photograph or video—is one of the most powerful surveillance tools ever made. While many people interact with facial recognition merely as a way to unlock their phones or sort their photos, how companies and governments use it will have a far greater impact on people’s lives.
When it’s a device you own or software you use, you may be able to opt out of or turn off facial recognition, but the ubiquity of cameras makes the technology increasingly difficult to avoid in public. Concerns about that ubiquity, amplified by evidence of racial profiling and protester identification, have caused major companies, including Amazon, IBM, and Microsoft, to put a moratorium on selling their software to law enforcement. But as moratoriums expire and the technology behind facial recognition gets better and cheaper, society will need to answer big questions about how facial recognition should be regulated, as well as small questions about which services we’re each willing to use and which privacy sacrifices we’re each willing to make.
Most people have seen facial recognition used in movies for decades (video), but it’s rarely depicted correctly. Every facial recognition system works differently—often built on proprietary algorithms—but you can sort out the process into three basic types of technology:
The detection phase of facial recognition starts with an algorithm that learns what a face is. Usually the creator of the algorithm does this by “training” it with photos of faces. If you cram in enough pictures to train the algorithm, over time it learns the difference between, say, a wall outlet and a face. Add another algorithm for analysis, and yet another for recognition, and you’ve got a recognition system.
The diversity of photos fed into the system has a profound effect on its accuracy during the analysis and recognition steps. For example, if the sample sets mostly include white men—as was the case in the training of early facial recognition systems—the programs will struggle to accurately identify BIPOC faces and women. The best facial recognition software has started to correct for this in recent years, but white males are still falsely matched less frequently (PDF) than other groups; some software misidentifies some Black and Asian people 100 times more often than white men. Mutale Nkonde, fellow of the Digital Civil Society Lab at Stanford and member of the TikTok Content Advisory Council, notes that even if the systems are operating perfectly, issues with gender identification remain: “Labels are typically binary: male, female. There is no way for that type of system to look at non-binary or even somebody who has transitioned.”
Contact us to discuss your requirements of face recognition tablet. Our experienced sales team can help you identify the options that best suit your needs.
Once a company trains its software to detect and recognize faces, the software can then find and compare them with other faces in a database. This is the identification step, where the software accesses a database of photos and cross-references to attempt to identify a person based on photos from a variety of sources, from mug shots to photos scraped off social networks. It then displays the results, usually ranking them by accuracy. These systems sound complicated, but with some technical skill, you can build a facial recognition system yourself with off-the-shelf software.
The roots of facial recognition formed in the s, when Woodrow Wilson Bledsoe developed a system of measurements to classify photos of faces. A new, unknown face could then be compared against the data points of previously entered photos. The system wasn’t fast by modern standards, but it proved that the idea had merit. By , interest from law enforcement was already creeping in, and such organizations appear to have funded Bledsoe’s continued research—which was never published—into a matching program.
Throughout the ’70s, ’80s, and ’90s, new approaches with catchy names like the “Eigenface approach” (PDF) and “Fisherfaces” improved the technology’s ability to locate a face and then identify features, paving the way for modern automated systems.
Facial recognition’s first dramatic shift to the public stage in the US also brought on its first big controversy. In , law enforcement officials used facial recognition on crowds at Super Bowl XXXV. Critics called it a violation of Fourth Amendment rights against unreasonable search and seizure. That year also saw the first widespread police use of the technology with a database operated by the Pinellas County Sheriff’s Office, now one of the largest local databases in the country.
Skip ahead a few years to , when Illinois’s Biometric Information Privacy Act went into effect, becoming the first law of its kind in the US to regulate the unlawful collection and storage of biometric information, including photos of faces. Jennifer Lynch, surveillance litigation director at the Electronic Frontier Foundation, describes BIPA as the model for commercial regulation. “Illinois requires notice and written opt-in consent for the collection of any kind of biometric,” she says. “At this point, Illinois is the only state that requires that.”
The s kickstarted the modern era of facial recognition, as computers were finally powerful enough to train the neural networks required to make facial recognition a standard feature. In , facial recognition served to confirm the identity of Osama bin Laden. In , Facebook publicly revealed its DeepFace photo-tagging software, the same year facial recognition played a key part in convicting a thief in Chicago and the same year Edward Snowden released documents showing the extent to which the US government was collecting images to build a database. In , Baltimore police used facial recognition to identify participants in protests that arose after Freddie Gray was killed by a spinal injury suffered in a police van.
Facial recognition first trickled into personal devices as a security feature with Windows Hello and Android’s Trusted Face in , and then with the introduction of the iPhone X and Face ID in .
Things have ramped up since then:
However, other, new players have entered the arena. Clearview AI made news in early when The New York Times revealed that the company regularly ran its recognition software against a database of photos scraped from sources across the internet, including social media, news sites, and employment sites—which Wirecutter, and many others, were able to confirm with testing—in a process that it used to identify suspects. In May , the ACLU announced a lawsuit against Clearview AI in Illinois state court alleging that it violated the privacy rights of Illinois residents under BIPA. Clearview AI is an outlier only in that it has faced public scrutiny: Equally less ethical software companies exist—companies that will sell their software to local law enforcement, usually with no oversight or public scrutiny into where the photos come from or how the identification algorithms work.
Proponents of facial recognition suggest that the software is useful because alongside identifying suspects, it can monitor known criminals and help identify child victims of abuse. In crowds, it could monitor for suspects at large events and increase security at airports or border crossings. The most long-running type of facial recognition software runs a photo through a government-controlled database, such as the FBI’s database of over 400 million photos, which includes driver’s licenses from some states, to identify a suspect. Local police departments use a variety of facial recognition software, often purchased from private companies.
There’s a long list of benefits facial recognition can offer outside of law enforcement, adding convenience or security to everyday things and experiences. Facial recognition is helpful for organizing photos, useful in securing devices like laptops and phones, and beneficial in assisting blind and low-vision communities. It can be a more secure option for entry into places of business, fraud protection at ATMs, event registration, or logging in to online accounts. Advertising and commercial applications of facial recognition promise a wide array of supposed benefits, including tracking customer behavior in a store to personalize ads online.
Brenda Leong, senior counsel and director of artificial intelligence and ethics at Future of Privacy Forum, suggested in an interview that proponents point to facial recognition as a replacement for loyalty programs or gated access: “You just walk through a set of cameras and all those things happen very seamlessly, sports arenas, event venues, amusement parks, all those places either are using or would have ideas of ways to use it similarly.”
Opponents don’t think these benefits are worth the privacy risks, nor do they trust the systems or the people running them. The first point of contention lies in the act of collection itself—it’s very easy for law enforcement to collect photos but nearly impossible for the public to avoid having their images taken. Mug shots, for example, happen upon arrest but before conviction. Error rates in recognition are also problematic, both in a false-positive sense, where an innocent person is falsely identified, and a false-negative sense, where a guilty person isn’t identified.
The facial recognition software that law enforcement agencies use isn’t currently available for public audit, and the algorithms that power the detection and identification software are often closed-box proprietary systems that researchers can’t investigate. When the public doesn’t know how these facial recognition systems work or how accurate they are, the public doesn’t know whether these systems are being used appropriately, especially in law enforcement. Joseph Flores, a software developer who in his free time uses machine learning for art projects (disclosure: I’ve worked on related artistic projects with Flores, for fun, not for profit), explained to me how he often intentionally biases his data sets to produce the results he wants, something law enforcement could also do: “You could do the same with your law enforcement facial recognition data to make sure that your friends were unrecognizable and your enemies were misidentified as criminals.” Flores adds, “It’s hard to challenge the legality or the reliability of math that you can’t review. Especially with the data scale we’re talking about. With no review everything is falsifiable and just modern phrenology.”
Another growing issue is law enforcement’s interest in real-time recognition in live video feeds or police body-cam footage. But even cities that enthusiastically moved forward with the technology, such as Orlando, Florida—where the police department used Amazon’s Rekognition software to attempt to identify suspects in real time from video streams—have dialed those efforts back after the technology failed to live up to expectations. But just because real-time facial recognition still suffers from hiccups on a large scale in live testing doesn’t mean it won’t become widespread in the future. The idea is so appalling to some communities that the practice is already temporarily banned in California, Oregon, and New Hampshire.
Generally speaking, the future of facial recognition can take any of three possible forms: no regulation at all, some regulation, and banning.
The Black Mirror episodes illustrating a world devoid of facial recognition regulation write themselves. Brenda Leong provided a few examples: “It’s very easy to create very Orwellian futures, where things are tracking you everywhere you go by your face because cameras are everywhere. If you’re a student it could be literally watching whether you’re focusing on your work versus daydreaming. If you’re an employee, monitoring your engagement on your computer or telling whether you wandered off somewhere else.” The list of surveillance possibilities is nearly endless, with China’s “Social Credit Score” or the London police force’s use of facial recognition cameras in real time offering a glimpse of one particularly grim reality.
As of this writing, there’s one proposed US law on a federal level banning police and FBI use of facial recognition, as well as another that allows exceptions with a warrant. Still another bill requires businesses to ask consent before using facial recognition software publicly, and yet another bans its use in public housing. Although facial recognition is certainly having a moment, it’s still unclear which of these bills, if any, will have enough support to become laws.
When anyone talks about regulating facial recognition, they need to divide the idea into two parts: regulating commercial use and regulating government use, including that of law enforcement.
For commercial use, Leong stresses, the main thrust of regulation concerning any commercial feature—a loyalty program, theme park VIP access, or whatever else—should be consent. Facial recognition “should never be the default,” she says. “It should never be part of the standard terms of service or privacy policy. And it should never be like the thing that happens that you have to then go opt out of.” The easiest way to see how such regulation might work in practice on a federal level is to look at Illinois’s BIPA, which requires consent before an entity can collect and use biometric data (including faceprints) and imposes requirements upon the storage of that data.
Consent can be tricky, though. It’s one thing for a store to ask if you want to skip showing your ID to enter and another when the store uses this technology to track shoplifters across all its franchise locations. As an example, the EFF’s Jennifer Lynch points to a recent case of a business district in London where a company placed cameras in a privately run area that people who worked nearby passed through: “You could see that the business district might say, ‘Oh, well, we put up signs,’” Lynch says. “And so people know that when they walk in this area or their face is being recorded and captured, but I don’t really believe that people can actually meaningfully consent in that situation. If you are working in that area, you may not have a choice of working somewhere else.”
When it comes to the government’s use of facial recognition, suggested policy approaches diverge. Leong says that although Future of Privacy Forum’s main focus is on commercial use of facial recognition, the group would want to see regulation of government use, too. “We would very much like to see overt, intentional regulatory guidance around how the government can and should use facial recognition,” she says, “even if it’s just things like being really clear about what levels of warrant or probable cause are required for agencies to access it.”
Other groups, including the EFF, don’t think regulation of law enforcement can go far enough.
Lynch, along with the EFF, argues that regulation isn’t sufficient. “We are pushing for a ban or at least a moratorium at the federal, state, and local level on government use of face recognition,” Lynch says. “It is a really game-changing technology and I think we’re at a key point in history where we could prevent broad government use of face recognition.”
Even as facial recognition addresses its diversity problem, there are still too many potential issues concerning how it’s used. “The security and policing industries are predicated on this idea that Black people are dangerous,” Mutale Nkonde says. “And so when thinking about tools for policing or tools for security, there is going to be this disproportionate deployment against Black people.” That is why Nkonde supports banning the software’s use outright: “I would want to see a ban around human subjects, just because I think the privacy trade-offs are too huge.”
Although policy changes, whether in the form of regulation or bans, offer the clearest way forward on a national scale, enacting such changes takes time. Meanwhile, there are smaller but not insignificant ways people interact with facial recognition on a daily basis that are worth thinking deeply about.
“I think that the concerning thing and the place where the distinctions sort of blur is that the more we use face recognition, the less we start to think of it, the less we think of it as risky out in the world, we become accustomed to it,” says Lynch. “I think it’s a slippery slope from using face recognition on your to the government using face recognition to track us wherever we go.”
Where society goes from here promises to be a mixture of policy and tweaks to people’s personal habits, but the conversation concerning the technology likely isn’t going anywhere for a long time. Like any technology, facial recognition is itself just software, but as Mutale Nkonde notes, how society uses it is what matters: “It’s the way the tool impacts our civil and human rights that is my point of intervention, because I think that all technology is agnostic.”
For more information, please visit 15.6 inch touch screen monitor.