Today Face Recognition Technology (FRT) has made a significant step forward and seems to slowly become more and more widely used in our everyday lives. It’s supposed to be the most natural of all biometric measurements. And for good reason – we recognize ourselves and other people not by looking at fingerprints or irises, for instance, but by looking at faces.
But what exactly is facial recognition? How does it work? Facial recognition is a type of biometric software that is able to identify or verify a person from a digital image by mapping out his/her features mathematically and saving the information as a fingerprint. This technology uses deep learning algorithms to compare these images for ensuring correct individual’s identity, and this way having the same aim as fingerprint matching, retina scanning and voice recognition.
Stages of image processing:
Facial recognition technology is able to perform its task following several stages of image processing:
1. Capture. The first step for the system is to collect physical or behavioral samples in predetermined conditions and during a stated period of time.
2. Extraction. Then, all this gathered data should be extracted from the samples to create templates based on them.
3. Comparison. After the extraction, collected data is compared with the existing templates.
4. Matching. The final stage of face detection technology is to make a decision whether the face’s features of a new sample are matching with the one from a facial database or not. It usually takes just seconds.
Where face recognition is being used?
And as long as your face is somewhere on the web Facial Recognition Software can find it, analyze it, and store it. A lot of industries and organizations use this technology, some you’re probably aware of, and some you’re probably not. So here are some examples of ways how FRT is being used today:
1. Access control
One of the most obvious ways of Face Recognition usage is access control of personal devices, residences, vehicles, offices and other premises alike. And a great example of it is Apple’s iPhone X using FRT for unlocking a smartphone. It’s possible thanks to infrared and 3D sensors that work non-stop within a front-facing camera.
2. Shopping online
Alibaba, a prominent Chinese ecommerce company, plans to use the Alipay platform to let users makes purchases over the Internet. And as a first step, Alipay has already launched a ‘Smile to Pay’ facial recognition system at a KFC in Hangzhou. The system recognizes a face within two seconds, and then verifies the scan by sending a mobile alert. ‘Smile to Pay’ is also able to identify people wearing make-up or wigs as a disguise.
3. Finding lost people
Such projects as Helping Faceless use FRT with the aim to bring lost kids back to their families and prevent child trafficking. Adults use a special app to take pictures of children begging for money on the street. The photos are automatically uploaded to the project’s database, where the facial recognition system tries to match them with photos already stored on the server. This data is also shared with special agencies that can help these kids.
4. Helping addictive gamblers
In casinos the system merely matches the faces of people playing the slots with self-proclaimed problem gamblers. When the system finds a match, it alerts the security team, which then discreetly approaches the gamblers and escorts them off the property.
5. Finding your soulmate
Who knew facial recognition could help you find the man or woman of your dreams?
Some dating sites use facial recognition to suggest their users to get acquainted with people with similar facial features. And they really made a business out of it.
6. Tracking down criminals
And this one comes as no surprise. Facial recognition is a crime-fighting tool that law enforcement and security agencies use to recognize people. For example, with the help of MORIS (Mobile Offender Recognition and Information System) – a handheld biometric device attached to a smartphone – the officer just has to snap an image and voila.
Several States like New York, for example, use facial recognition to find criminals who steal other people’s identities so that they can get driver’s licenses. And facial recognition is a powerful tool to unmask imposters.
7. Recognizing regular customers
Hotels use a facial recognition system to recognize their customers. That way, the concierge can remember customers’ names and can greet them accordingly making their guests extra special. Also, luxury retailers in Europe use FRT to recognize their VIP customers to offer special care. Airports in Australia use this technology to allow passengers fly without carrying their ID-card.
8. Picking out underage buyers of alcohol and cigarettes
Underage drinkers and smokers can now be easily recognized. A supermarket chain uses face recognition cameras to prevent staff from selling alcohol and cigarettes to people under 18. The system has customers’ photos stored in its database, and if it recognizes someone who has been unable to prove his/her age before the system will alert a cashier about it.
9. Taking attendance in school
Schools in the UK use FRT to take attendance. This has been going on in the UK for a while, but will certainly spread to other countries as well. According to the reports, both students and teachers in the UK love the new system that scans faces with an infra-red light and matches them up with archived images.
10. Organizing photos
Finally, the most widespread way to use this technology: Apple, Google, and even Facebook use their own face recognition systems to distinguish a portrait from a landscape, find a user in photo, and sort images by categories. And every time we upload a photo and tag our friends on it we all provide an enormous help for the facial recognition algorithm.
Challenges that FRT faces today
All of the above are just examples of the good that facial recognition technology can do. Quicker the progress is, more amazing and interesting our world becomes. Nevertheless, every technology has its drawbacks. Let’s look at those of FRT:
1. Data Storing
Nowadays, it is more and more expensive to store huge amount of data for further usage. Either it is a picture in a low resolution or a high-quality video – everything needs space. This means that in order to be efficient FR systems process only about 10-25% of data. Some companies use clusters of computers to process all the data and minimize the time spent on it. But every new computer needs all this data to be transferred via network and this lowers a processing speed.
2. Image Size & Quality
Algorithms used by facial recognition technology needs HQ digital cameras to operate accurately. A face on the picture or on the screen-shot from a video is captured by a face-detection, after that the relative size of the face image is compared with the size of the enrolled one. So, the quality of the photo plays the major role in FR process ensuring that the recognized identity is clear. And the better the photo’s quality is the more correct the results of FRT’s work are.
Moreover, processing of different face image sizes is a processor-intensive task. Most systems provide the identification of a face-size range so as to exclude false recognition and accelerate processing of images. But this face tracking software needs a substantial investment. However, it will pay back in spades.
3. Surveillance Angle
Also the identification process is strongly dependent on the surveillance angle that is responsible for the capture of target’s face. Multiple angles of face are being used while enrolling the face through the recognition software – profile, frontal, 45-degree, etc. But for generating a clear face template, only a frontal view is needed. The higher resolution photo has and the more direct its angle is (goes for both enrolled and compared images) the more accurate resulting matches would be.
Then, there are also problems with, for example, facial hair or sunglasses. One can still fool the FRT with a suddenly appeared or removed beard, same goes for obscuring face’s parts with glasses or masks. So system databases must be regularly updated with the most up-to-date images to avoid such failures.
While there may be some initial concerns with facial recognition, it is clear that this technology is likely to continue its rapid growth. It is possible that in just a few years such systems will be advanced enough to process expressions and hand gestures within a matter of seconds. We are just to wait to see it.
If you have questions or want to add something feel free to leave your comments below!
Thanks for reading!