Facial Recognition for People Wearing Masks

from Schneier on Security

The Chinese facial recognition company Hanwang claims it can recognize people wearing masks:

The company now says its masked facial recognition program has reached 95 percent accuracy in lab tests, and even claims that it is more accurate in real life, where its cameras take multiple photos of a person if the first attempt to identify them fails.

[…]

Counter-intuitively, training facial recognition algorithms to recognize masked faces involves throwing data away. A team at the University of Bradford published a study last year showing they could train a facial recognition program to accurately recognize half-faces by deleting parts of the photos they used to train the software.

When a facial recognition program tries to recognize a person, it takes a photo of the person to be identified, and reduces it down to a bundle, or vector, of numbers that describes the relative positions of features on the face.

[…]

Hanwang’s system works for masked faces by trying to guess what all the faces in its existing database of photographs would look like if they were masked.

More here.

Posted in Law, Technology and tagged , , , .

18 Comments

  1. The most interesting part about this article is how machine learning algorithms work and how it makes it easy to develop this type of facial recognition technology. Although they are just claims, and only use photos from a small group of 2,000 people, its hard to say whether anything could be considered true, or atleast the acknowledgment of some implicit bias. But what they are saying they have done to develop this is not something impossible. Machine learning algorithms and computers have recently been used in all types of situations where trends in data can be analyzed. Most interesting to me is the use of machine learning when it comes to what are known as “deepfakes”. Deepfakes are images or videos made using machine learning technology, and the basic gist of it is you can can edit any video or picture to make exact replicas of people. For example, there are many deepfakes of Barack Obama saying things that he never actually said. These deepfakes utilize this machine learning technology, which makes differences in real footage and deepfaked footage almost indistinguishable. This has very serious implications if put into the wrong hands or used maliciously, as fake news and media panic could be spread extremely quickly by abusing this technology. As seen with the reaction to the coronavirus pandemic, media panic can cause widespread societal panic and distress. As for facial recognition, it is already being used by certain countries to stop violence, for instance in Hong Kong, where protests have been raging for months and police are trying to identify protestors and political opponents, all who wear masks or clothes hiding their identity when protesting. Technology like this would spell the end for many protestors and give and incredible edge to whoever is using it.

  2. Throughout many of these blog posts, many of them relate to technology. That just goes to show that technology is continually becoming an important part of our society. In our current situation, technology has been and still is becoming an important part in combatting the COVID-19 virus. Hopefully, we can begin to upgrade our tools so we can be able to better test for carriers of the virus. But in the article, it is speaking on facial recognition technology being able to recognize citizens wearing masks. I think that is very convenient is especially in the country of where the virus originated, because the amount of people that have to use face masks is very high. So, by having this technology makes it much more convenient for whatever purpose it is used for. Whether it be for facial recognition in your phone or for scanning in at work, to have it in a time where you are encouraged to wear a mask is very good. The article also cites something about how the program functions, that it has to disregard some information to be able to recognize. It essentially takes the aspects of your face not covered by a mask as uses that. That being said, it is easy to recognize some controversy with it. What happens to people who wish to cover themselves from cameras or people who simply do not want to be recognized in a database. That technology will be able to recognize a larger collection of people. This speaks to the advancements technology is making, whether or not we like it. For you to believe that you still have total privacy, then you are a fool. Much of the technology you own is able to take a look into your life, even in your home. If you really wanted to have your own privacy, you would have to go off the grid and find a shack in the woods. It just goes to show that, the industry being what it is, should we continue to try and fight it or should we get with the times?

  3. It is crazy to think that technology such as the facial recognition program produced by Hanwang can recognize people that wear masks. It seems every day we get a little closer to being able to create a society such as Oceania, the dystopian society depicted in George Orwell’s book, 1984. I would even say that of all the major countries of the modern world the People’s Republic of China is the closest to creating such a society.
    In 2014, The Chinese government announced its plan to create a “social credit system”. In this system each citizen will be ranks according to their social credit score, this score can change depending on the actions an individual take. Those with low scores will face punishments and those with high scores will receive rewards. Some of the punishments for having a low score include, banning you from using various forms of travel, restricting you and your family’s access to certain schools and levels of education. Some of the rewards for “good citizens” include, discounts utility bills and banks offering better interest rates to you. The Goal of said system is “to reinforce the idea that keeping trust is glorious and breaking trust is disgraceful,” In clearer terms, this system will enable the Chinese government to curtail behavior that it doesn’t approve of and incentivize behaviors it deems valuable. This system is a more sophisticated nightmare of an Orwellian nature. Its not difficult to see such a system quash any remnants of freedom in China.
    This system is supposed to be fully implement throughout the country by this year. Technologies such as the one developed by Hanwang will only make this system more efficient and will provide more power to the government in its goals. With this information, the facial recognition technology will have a negative effect for the average Chinese citizen, as this will allow the government to punish any citizen that attempts to hide their identity from the government during protests, such as the ones that were happening in Hong Kong.
    I think it should be concerning to all people that live in free countries, that this technology has been developed. Though this technology can be used to figure out the identity of bank robbers and other criminals, it can also be used for nefarious purposes, such as it will be used for in China. Thus, we should restrict the use of such technology. If we allow such technology to be pair with a “social credit system”, it will inevitably lead us to a worst society.

    https://www.businessinsider.com/china-social-credit-system-punishments-and-rewards-explained-2018-4

  4. With many countries and states deeming face masks essential when leaving the house makes the need for facial recognition despite these covers increasingly necessary. Hanwang, a company responsible for being able to perform facial recognition on people wearing masks says its technology has reached 95% accuracy in identifying these mask wearers. Over the years, facial recognition has become a very standard procedure for security and have proved to be effective. This facial-recognition company has placed 2 million of its cameras at entrance gates across the world, started preparing for the possibility of this pandemic early in January. This company also holds a 99.5% accuracy rate for full face images. Demand for these unique products has soared as many essential businesses are looking to get a hold of them in order to get reliable footage of all individuals. Many authorities across the world have been trialing these devices in order to the address the need for “anti-terrorist measures” in certain regions. While the rapid rise of facial recognition technologies has been undeniable, various states and municipalities are enacting new, stringent laws regulating the use of facial recognition technology by commercial entities. Facial recognition technology involves the process of using “biometrics” to digitally map an individual’s facial geometry such as the distance between an individual’s eyes, which has essentially assisted the China company Hanwang in being able to recognize partial faces. It is very important that this company is able to provide evidence of these accurate faces in order to help with crimes or other needs. Does this new technology hold the ability to recognize people who wear sunglasses by looking at lower face geometry? This is a big step and this company has showed tremendous preparation by testing these new devices out promptly at the beginning of this pandemic. These facial templates that are created by these devices when testing recognition need to have a better accuracy rate in the coming future. The almost certain accuracy is beneficial but can still be fought in court. Facial recognition technology has significantly enhanced the operations of businesses across all industries tremendously including security/identity fraud prevention, access and authentication, and accessibility to accounts and services. Moving forward, with facial recognition being the fastest-growing biometric technology, we need to invest multiple studies into the development to ensure success throughout the world.

  5. I genuinely was just thinking about this idea the other day when I was binging Scandal all day. In the episode, they were using facial recognition to try and find the person that attempted to murder someone. I wondered what things were like now that not wearing masks is not an option. I think with the history of mask usage literally everyone looks like they just robbed a bank. Now that everyone looks unidentifiable how are they identified?
    Schneier on Security’s blog mentions a company figuring out how to combat facial recognition with the masks. The company created a facial recognition program that has reached 95 percent accuracy in lab tests and even claims that it is more accurate in real life, where its cameras take multiple photos of a person if the first attempt to identify them fails. The blog also states, ” A team at the University of Bradford published a study last year showing they could train a facial recognition program to accurately recognize half-faces by deleting parts of the photos they used to train the software. When a facial recognition program tries to recognize a person, it takes a photo of the person to be identified, and reduces it down to a bundle, or vector, of numbers that describe the relative positions of features on the face.” This then sparks my further questions like is recognizing half of a face reliable enough? Are the results going to be less accurate, especially if there are many people with the same features? How long is this facial recognition take? Lastly, how expensive is this creation? The blog post seems to be very open-ended and vague about the details of the system
    When reading the comments on the post many people had the same concerns. Some people even pointing out the fault in this technology which is that “It’s 95% accuracy with up to 50,000 faces, but actually trained with a database of 2,000 people, so the 50,000 is speculative. That’s 1 in 20 false positives with a population well under a million. Reading between the lines, it may also require people to aim their faces at the camera. Good enough for unlocking a door in a small office, not as useful for mass surveillance. (That’s what cell phones are for!).” I am no tech guy but this person who commented made me put things more into perspective. This system is only the beginning and it is not completely functional… yet. With time I feel like this can be a very, very useful thing for the government to use.

  6. I find this article interesting since it demonstrates the extent to which monitoring applies. As discussed with previous articles, the monitoring or surveillance of the population is a double bladed sword. Systems used for monitoring and surveillance can provide a higher level of security at the expense of privacy. It is fascinating how algorithms such as facial recognition can be brinquen into axioms. It is understood that the dilemma with the accuracy of face recognition algorithms is the fact that the portion of the face that is covered by the surface will not be useful since only the top half is exposed when a mask is being used. Emulating how the subjects would look when wearing a mask seems to be a very advance implementation to facial recognition systems. This would obviously provide security agencies with an advantage during this uncertain times. The Pandemic still poses a threat for the world. This article, however, does not take into consideration the ethical implications specially in the setting where this advanced face recognition systems are being implemented. The CCP is known for its strict restrictions and control over most of the Chinese population. Depending on the setting, the use of such advanced systems may be perjudicial and, regardless of the setting, there’s always the probability of the information gathered being used by the wrong people or for wrong purposes. This article focuses mostly on the operation of the algorithms and not on their application, which is the most important factor when it comes to a new technological development. The article fails to explain the specific purpose for this updated face recognition system. Almost with out a doubt we would assume that it’s implementation will be utilized as a security tool, but the higher the invasion of privacy gets, the more I question that security is the sole purpose.

  7. The Background. Facial recognition technology was first developed in the mid-1960s with funding from U.S. intelligence agencies and the military. In the very first version of facial recognition in the 1960s, Bledsoe created a system that could organize faces’ photos by hand using the RAND tablet. … People used that system to manually record the coordinate areas of facial features like eyes, nose, mouth, and hairline. Today, facial recognition technology uses artificial intelligence algorithms to identify people. It is used everywhere from airports, venues, shopping centers and even by law enforcement. This tech has become quite standardized across security platforms and even in everyday life. Growing up facial recognition was something out of movies for me, and today every time I unlock my phone and purchase a new app, I use facial recognition software. As we face the Coronavirus pandemic, I would’ve thought that having clear visibility only of half a person’s face would limit the capabilities of this technology. Hanwang however has developed the tech to combat this problem. Their system has a 95% success rate in recognizing mask-wearing faces. This is a big step for facial recognition technology. While I certainly see this innovation as a helpful asset for security it leads me to think about how quickly our privacy is disappearing this day in age. the facial recognition industry is expected to grow $3.2 billion in 2019 to $7.0 billion by 2024 in the U.S. The most significant uses for the technology being for surveillance and marketing. China is using facial recognition to fine jaywalkers, imagine this tech hitting a place like New York City. Are we free if we can’t even cross the street without being flagged? Legislation and regulation are in the conversation but if the tech proves to be a good investment it is likely that we will not stay ahead of or even, try to keep up with privacy concerns.

  8. The title of this article really caught my attention as I thought about our current situation and the immense amount of people wearing masks to work or out in public. Of course, with current facial recognition technology in cellphones, you would need to remove the mask to unlock your phone. So facial recognition that could recognize your face with a mask could be very beneficial. On one hand, it could be a very positive technological development, but on the other, it may cause some problems for users.
    Hanwang, the Chinese company claiming they have developed this technology, says that the program has reached “95% accuracy in lab tests” and could be even more effective in real life. This could be a huge convenience for health care professionals or anyone working with a mask regularly. My brother currently works at a local Dunkin Donuts which requires him to wear a mask and I’m sure he would love if his iPhone could recognize his masked face. My first question would be: how would this work? I was under the impression that facial recognition used all of the unique features of one’s face to accurately recognize the person attempting to unlock their device. It seems that a study at the University of Bradford has concluded that they can “train a facial recognition program to accurately recognize half-faces by deleting parts of the photos they used to train the software.” If this works well with smartphones, it would be a great advancement in facial recognition technology.
    While this could prove to be a beneficial advancement, I could see there being some issues for users. With current facial recognition technology, it sometimes fails to recognize you if you are wearing a hat, or closing your eyes. So, it seems like with only half of the face to recognize, it could pose more problems as there is less of the face to differentiate. If another person with similar eyes was also wearing a mask, could they maybe unlock your device. Perhaps, this new technology will focus a lot on the uniqueness of each person’s eyes to help improve the accuracy. Also, there are always issues when we talk about facial recognition and privacy. If devices can now recognize people with masks covering the majority of their face, a level of privacy is infringed upon. This may worry people who are already worried about our privacy in the age of the internet. Overall, I believe this advancement in facial recognition could be very effective for the time we are living in, but may cause issues in the future and will have to be perfected.

  9. This article stood out to me right away, mainly because these last few weeks I have had to wear a mask where ever I went due to Covid-19, In New Jersey all businesses have signs up that require customers and employees to wear masks, with that being said who knows how long we will be required to wear masks, now every time Im wearing my mask Ive noticed my iPhone cannot recognize facial ID in order to unlock my phone, it hasn’t been to much of in convince as I will just swipe up and manually enter my passcode but I keep thinking that since most of the world has been forced to wear masks for there safety over the last month, smart phone companies like Samsung and Apple would create some sort of software update that makes it so facial recognition technology that only scans the part of your face that is not covered by your mask. In this article it seems the Chinese facial recognition software company “Hanwang” has claimed it has created technology that can recognize people wearing masks, this software system has the camera take multiple photos of the person if the first attempt at recognizing fails, the article also makes note of that if you train these facial recognition algorithms to be able to I.D. faces wearing masks it will have to throw data away, this means you could actually train an algorithm to recognize half of a face (or one wearing a mask) by deleting parts of the original photo that was used to train the algorithm. The way “Hanwang” is able to make its system work by trying to guess what all the faces in its all ready existing database of photos would look if they were masked. I think this would be like Apple training its facial recognition algorithm to be able to recognize someones face wearing a mask based on the existing facial scan it already has of you in it database and what they think that would look like with a mask, I can see this system being flawed mainly because not everyone wears the same type of mask, some take up more of the face then others, but I feel Apple and Samsung need to follow in the footsteps of “Hanwang” in order to train its facial recognition algorithm to be able to recognize a face wearing a mask.

  10. Believe it or not, it’s not that difficult for facial recognition technology to recognize people wearing masks that covers 50% of their face. Facial recognition technology would just have to analyze one’s iris along with what’s visible of their face instead of their whole face where companies like how Apple recognizes faces with their Face ID where the camera maps out the structure and depth of your face with invisible dots.

    Iris recognition is nothing new, it has in fact been around for more than a few decades along with facial recognition software and overtime, it’s just going to get better and better along with the tech industry. This is also not the first time where facial and eye recognition technology has been employed to recognize people with face masks or other types of identity concealing objects. Iris recognition technology has been used throughout the protests in Hong Kong by the government in order to arrest and detain the lawbreakers as well as being used in China by the government during the pandemic to quarantine those who are sick or showing symptoms.

    In the article, it is said that Hanwang’s facial recognition software are able to recognize masked people with a 95% accuracy along with a 99.5% accuracy for unmasked people in another article: (https://www.reuters.com/article/us-health-coronavirus-facial-recognition/even-mask-wearers-can-be-idd-china-facial-recognition-firm-says-idUSKBN20W0WL). The statistics shows how crazy technology is nowadays as well as making us wonder whether our privacy is decreasing as technology and facial recognition software advances. Furthermore, Apple is also coming to address the problem of their masked customers not being able to unlock their phone with Face ID with an update that solves this problem. In the end, I think that this pandemic accelerated development of facial recognition technology faster than usual as part of a plan to adapt to the pandemic (instead of waiting it out).

  11. The facial detection features that some companies are coming up with are incredible. The number of information companies receives from one camera is incredible. Many people were worried about governments using this type of software because it could be a breach of privacy. Just think at every corner in a city there could be a camera taking your photo and now telling somebody where you are and what you are doing. When the virus outbreak came along, facial recognition took a backseat and has allowed them to just improve and improve in the background. Now they are saying these companies can recognize people even while they are wearing masks. This is incredible and will be the start of something good or bad.
    This could be used for good because it can limit crime. “The company now says its masked facial recognition program has reached 95 percent accuracy in lab tests, and even claims that it is more accurate in real life, where its cameras take multiple photos of a person if the first attempt to identify them fails.” This means that the system could perform even better in public than it is in testing. Not many people are going to be able to successfully steal or rob a store if the store has cameras that can identify people even if they are wearing a mask. This could lead to a lot less crime in general because you could always know who is where when something goes down as long as a camera can see it. This can be scary however because no there is no way to hide. Someone is always going to know where you are and what you’re doing when you go out. The only safe place is at home but then we have to deal with different types of screens that track our every move.

  12. It is a strange, Syfy-movie feeling to think that people are now dedicating time to creating technology that will help with facial recognition for those wearing masks. In Pennsylvania at least, there is a requirement to wear a mask any time you go into a store and in the few times I have been out, this is a problem I have faced when picking up my phone. In China, where the pandemic originated and has resided for much longer, this is a problem that has persisted long enough that people have developed ways for these systems to recognize faces, even though there are part blocked off. What is surprising about this is that this is done through the system trying to figure out what the faces that they have in the database would look like if they had a mask. However, the article states that this can lead to problems of people not being identified correctly because so little of their faces are being shown. The system works best on a smaller population of 50,000 people rather than billions of people. While I understand that companies want to keep themselves safe by having cameras recognize their employees, I felt that there were potential issues when the protesters in Hong Kong were brought up. Though Mr. Huang said that he was not involved I the potential use of facial recognition of these protesters, it is very probable that this technology will, or already has been used by the police against the protesters.

  13. Although a very short summary of this piece of technology, no more is needed to understand how a tremendous of a feat this would be to completely accomplish and roll out onto market. The United States has never seen a need like this before the coronavirus. It is fitting that this was being developed in China because wearing masks in Asia has been a normal for a long time now even before the virus. The pollution in their country is on an entirely different level than the rest of the world which is why it baffles me why many public figures always call out the United States instead of China. Nonetheless, with the Chinese people having been wearing masks as a culture for a while now, can imagine they are not too pleased with not being able to use one of arguably the coolest features of the new cell phones which is facial recognition to unlock. With this being an issue around the world now because of the virus, this will be able to give people that much needed convenience that they cannot do without. The only other places that come to mind where this can be such a big help to is women in a strict Muslim culture. They are often forced to wear garments over their face for their religion and this would be a major help to them as well. Not only would this have practically in this specific climate today, this would be used with endless applications. This could start being used to identify people that commit crimes and are wearing masks using databases that show faces and match them. This would almost start making it impossible to get away with a crime in a day and age where it already is almost impossible to commit one depending on what area your in. This might also have some practically in doing rendering of faces. Many times, we have artist’s sketch what their interpretation of a face might be, where this might be able to be used in the future to more accurately depict that. The possibilities are truly endless with this.

  14. This article explains the intricacies of computer programming and how artificial intelligence changes facial recognition. As technology becomes deeply integrated into every facet of society, we have to understand exactly how it affects us, and its implications on privacy and surveillance. While there are many advantages to utilizing facial recognition technology, as seen in cellphones and laptops, mass surveillance like we see in China has become akin to the oppression seen in George Orwell’s 1984. This scares me far more than the potential upsides, because the availability of this technology does not make me confident that my own privacy will be prioritized. As we have discussed in class, I see this as nothing more than another way to control people and train them to behave in a way that only the government would approve of. There needs to be heavy scrutiny placed on this new technology as it is and will continue to be abused to control people. Surveillance has been a cornerstone of the Chinese government, tracking billions of peoples every move with the help of hundreds of millions of cameras. They have implemented toxic systems which dramatically reduce the standard of living of citizens, as they are unable to actually make decisions for themselves and have to run it through a mental filter of whether or not the action would please the government. The social credit system affects the ability of people to get loans and jobs, so the role of government influence is reinforced in every way. Albert Maysles said that “Tyranny is the deliberate removal of nuance” and I think that this perfectly exemplifies the situation in China right now. I hope that as we become an increasingly transparent world the horrifying influence China has over its citizens will be reduced to humane levels and the people will be allowed freedom of choice. While this advancement can be worrisome on the surface, it shows us a broader trend that will hopefully lead to a world with an increased accountability for how governments treat their people.

  15. Currently, amongst the COVID-19 pandemic we collectively are experiencing firsts as a society. From remaining at home, to wearing facial masks, gloves, waiting in lines to enter our local supermarkets and grocery stores, and distancing ourselves from each other, these are all firsts for us. In the past century, we have not faced a pandemic and all of us remain in uncertainty and worry. In an age of tech savvy, we face new challenges with our facial recognition smartphones or touch identification phones. Because of the masks many local and state authorities require civilians to utilize, the ability to unlock our phones with merely our faces has become impossible. Currently, the facial detection software is unable to read facial measurements and features due to the masks and similarly, the touch identification software is unable to read fingerprints because of gloves. Of course, everyone is capable of merely entering their passcode but this can contaminate phones and fingers with bacteria and potentially the corona virus while shopping. This could cause the potential spread of the virus through a subconscious touch of the mucus membranes. If this program from Hanwang truly is capable of recognizing facial features of individuals while wearing a mask, it could simplify various things for healthcare workers and everyone else as well in the upcoming months. At my local grocery store they have introduced a mobile shopping app, which allows customers to scan their items while adding to their carts and at the end pay for their groceries through contactless checkout. Speaking for myself, not only is it annoying not being able to benefit from the facial recognition software, but also having to consciously remind myself to disinfect my phone. This could be a superficial update to the existing software and render useless in reality, but I would appreciate not having to touch and potentially contaminate the screen of my phone as the screen of my phone is what comes to contact with my face when making calls. The only way to test the usefulness of this is to simply use it and determine for ourselves if it is a worthy and useful update.

  16. “Facial recognition for people wearing a mask” is an article written on 18 April 2020 in which Schneider on security writes a blog. According to this write-up, Hanwang is a Chinese company that claims that it is capable of recognizing people wearing masks. The company claims that the results that were taken to the lab are 95 percent accurate. This recognition is conducted in the process where a photo of the person to be identified is taken and is reduced down to a bundle or vector of numerical, which describes the relative position of the feature on the face.

    The program works through guesses of the faces in its existing database of photographs would be like when they have won masks. In my opinion, this recognition program seems possible because a person can be identified by looking at them. After all, people have different eyes regardless of the color. This identification program cannot work in open public places because people are always on the move, and their recognition may seem harder. It is a time-consuming process since it requires a comparison of the taken photo with all the images in the database and, in turn, may delay feedback. The program is also expensive because many photos should be taken and be stored in the database so that it can be used in the future for comparing the images to get the person in question. This database requires skilled personnel to handle the database systems as well as a vast workforce to conduct the whole process of recognizing people who have won a face mask.

  17. The COVID-19 pandemic has brought a lot of issues to the surface. However, I am impressed with the solutions to these issues. This article discussed how a facial-recognition company, Hanwang placed two million of its cameras at entrance gates across the globe in January in preparation for the coronavirus. Hanwang’s chief technical officer, Huang Lei explained that he first received requests from hospitals in Hebei province, who were dealing with the outbreak. The hospitals asked him to update the software so that nurses wearing masks can be recognized. Besides the technological side of the facial-recognition company being successful due to its accuracy in recognition, I think the ability to quickly act on an issue has truly added to the company’s success. For example, Lei stated; “We wouldn’t wait until something explodes to act. If three or five clients ask for the same thing?.?.?.?we’ll see that as important.” I think that this attitude should be applied to any business. Without clients, every business would fail. Hanwang does not wait for a hundred clients to request something, instead when they start receiving even a few requests they quickly work to meet those requests, which results in making their company more efficient and their clients loyal. A lot of the time company’s get overwhelmed, lazy or do not want to spend the money to resolve issues that clients are pointing out and this results in their clients losing trust in them and eventually cutting ties with them. Since Hanwang replies to their clients so effectively, the demand for their product has reached a new high, resulting in a broad range of clients. Now, police stations, railway stations and all the office towers use Hanwang’s cameras to screen employees. Hanwang now says its masked facial recognition has a 95 percent accuracy in lab tests and affirms that it is more accurate in real life, where its cameras take several photos of a person if the first attempt to recognize them fails. I think it’s amazing that China built a company that can accurately identify individuals wearing masks, meanwhile American companies such as Apple with their iPhone cannot recognize the owner of the phone wearing a mask when that person tries to unlock their phone. I think the U.S. has something to learn from and strive from by observing Hanwang’s technological advancements. However, there is always a chance for error. For example, misidentifying a masked face is more likely to happen than misidentifying a full face because there is less data to work with. There may be several people with similar features around their eyes and nose and this could lead to an error in identification. To combat this issue, Hanwang tries to predict what all the faces in its current database of photographs would look like if they had masks on. Hanwang asked nearly 2,000 employees for photos of themselves with masks on, but also made an additional database of about six million photos of people with artificially generated masks, displaying different types of masks commonly worn in China. I think all of these efforts truly displays how smart, savvy and innovative Hanwang is. This company tries and tries until they get the most accurate result they possibly can and then they still try to improve upon their accuracy. I think it’s important for other companies to learn by Hanwang’s actions that a company is never finished. There is always room for improvement. Lastly, it is important to point out that Hanwang’s market primarily targets entrance gates, meaning it is not looking to apply its cameras to public spaces such as large stores, due to the larger degree of probable inaccuracy of different people in and out of stores unlike the workplace that has the same employees day in and out. I think this leaves room for hopes in future of even greater technological advancement to apply cameras in public spaces, while remaining accuracy of facial recognition.
    https://arstechnica.com/tech-policy/2020/03/how-china-built-facial-recognition-for-people-wearing-masks/

  18. Wearing masks in the United States, at least in the North East, has been made mandatory by most states in order to prevent the spread of Covid-19. So the question of how are we going to be able to tell who is who and identify people that commit crimes comes into play. This has never before been a problem within the United States, but times are changing and 2020 is a crazy year making this type of software needed. This company has a great software that can have 95% accuracy identifying people with a mask on and to me that is absolutely amazing. Technology is becoming so advanced that this software can take half of someone’s face and be able to tell exactly who someone is.
    This article came out before the riots and looters started, in retaliation to the death George Floyd, and now this software can be put into full use. Since the death of Floyd people have been protesting for black lives matter and against police brutality and even though most people are protesting peacefully there are always some people that take it to far. Therefore, people are raiding, vandalizing, and burn down local business with masks and shirts over their faces thinking they can’t be identified. Making this software a perfect way to catch people that are doing the wrong thing to prove a point. They are not bringing justice to George by destroying their communities, but are just committing crimes and blaming what they are doing on that situation. Now there is a way to fix and bring justice to those people harming the communities.

Leave a Reply

Your email address will not be published. Required fields are marked *