How can you combine convenience, security, and hygiene into a single authentication solution? If your answer was contactless biometrics, you’re right. This is why biometric technologies such as the face, iris, fingerprint, voice, and hand geometry recognition are on the rise in a post-pandemic world. Of these, palm scanning seems to have the edge.
But is paying with your palm really better than a fingerprint scan, for instance? And what does it take to develop palm scanner software?
We provide companies with senior tech talent and product development expertise to build world-class software. Let's talk about how we can help you.
Contact usIn this article, we’ve taken the new Amazon One palm scanner as a starting point for exploring the issues and tech behind palm recognition. With our experience in fintech software development services at Relevant, this tech is close to our heart. We wanted to share our discoveries and help you decide whether a palm scanner could be the next step for your business.
Table of Contents
As the digitalization revolution continues, the importance of authentication systems that are highly secure (in both data and health terms) is rising exponentially. Digital banking, safer shopping, contactless transactions, and hygiene standards in healthcare are just some examples of why biometric scanners look set to be a new tech boom.
And the numbers back it up: by 2027, global biometrics technology revenue is predicted to reach $55.42 billion, with the identity verification component taking $18.1 billion. Factor in that these estimations were pre-COVID-19, and you’ll see why biometrics is worth a look if you’re considering new development areas. With the coronavirus marching through countries, nobody wants to touch anything unless it’s absolutely necessary. This subconscious fear won’t disappear instantly, but the development of contactless biometric technologies can significantly ease it.
Given all this, it’s not surprising that the Amazon One palm scanner, released last September, has been making waves. And while the idea is not that innovative, recent developments in computer imaging and data processing will probably make the Amazon hand scanner a go-to solution for many industries.
Let’s take a closer look at this contactless palm recognition system.
Amazon One is a biometric device that scans your palm with computer vision to either create an image (on registration) or compare the scan with the stored image (for authentication). The palm scan includes both your external hand “drawing” and its internal network of veins. The scanned image can be associated with personal data such as a credit card number or loyalty card, or anything else you provide upon registration.
When shopping with Amazon One, you can just hover your hand over a palm scanner (no touching!) to enter the store, present loyalty cards, pay for purchases, and so on. Paying with your palm is convenient, and Amazon One states that customers save time, touch fewer things, and shop more efficiently.
Imagine you’re going to a rock concert, and all you have to do to squeeze into the stadium is scan your palm. Or let’s think bigger, and consider biotech institutions, surgical departments, or other locations with exceptional hygiene and security requirements. A contactless biometric system is a perfect choice.
But what about data protection? After all, the company gets access to all your personal details. Amazon addresses this serious concern with scan encryption and secure cloud storage. Images do not remain on an Amazon One device, and you can delete all data associated with you at any time.
Why should retailers opt for palm recognition solutions over others, you may ask. Retina and iris detection may seem more secure than a palm scan, but the hard truth is, they aren’t.
Retina and iris detection, voice recognition, and fingerprint scanning are all hackable. With iris scans, for example, hackers can take a photo using night mode cameras, print the picture, and place a wet contact lens over the iris image. It’s also possible to produce an iris code using recognition software and create an almost identical iris template. As for fingerprints, they can be gathered from surfaces and then replicated using silicone.
Some biometric methods are also simply less accurate than others, which is frequently exploited.
Comparative Accuracy and Ease of Use of Different Biometric Methods
With a palm scan, however, it’s not the surface you are scanning—or rather, not only the surface. The combination of wrinkles, lines, and veins makes your palms unique and, theoretically, harder to copy.
To understand in more detail why palm authentication is better and how it works, let’s first explore the difference between palm vein and fingerprint recognition.
While fingerprint identification has been around since the 19th century, vein scanning is a new technology. Vein structure, revealed by infrared light, is unique to each person, differs on both palms, and remains stable during your lifetime. It also allows for contactless reading. Here are some more benefits of palm vein authentication over fingerprinting.
Let’s start with the fact that fingerprints can be collected without your consent. Also, they can be easily replicated, as we’ve mentioned earlier. A palm vein network, on the contrary, cannot be captured as easily—it’s internal and doesn’t leave a physical print. Besides, as infrared light captures blood flow and not static lines or wrinkles, it’s impossible to forge a replica—your veins must have blood running through them. This makes palm vein authentication far more secure than fingerprinting.
Two questions are used to measure the accuracy of a biometric system:
In these terms, palm recognition solutions have the lowest indexes compared to other authentication methods:
Method | FRR | FAR |
---|---|---|
Fingerprint | 3 to 7 in 100 (3-7%) | 1 to 10 in 100,000 (.001-.01%) |
FaceRecognition | 10 to 20 in 100 (10-20%) | 100 to 1000 in 100,000 (.1-1%) |
VoiceRecognition | 10 to 20 in 100 (10-20%) | 2000 to 5000 in 100,000(2-5%) |
Iris | 2 to 10 in 100 (2-10%) | >=.001% |
Hand Geometry | 1 to 2 in 100 (1-2%) | 10 to 20 in 1000 (1-2%) |
Amazon cites ease of use and customer satisfaction as the key reasons for creating an Amazon palm scanner. While fingerprinting is more familiar to the public, it often fails to grant access instantly (you’ve probably tried to unlock your smartphone with wet fingers or a scratch on your fingertip). Palm vein scanning requires a simple “hovering” motion, and the infrared light can get through the wear and tear of your palm right to the vein. Better yet—you don’t have to touch anything.
Direct contact with a sensor was acceptable before COVID-19, and smartphone sensors are commonplace. But touch-based biometric scanners in public places have always had the potential to spread infection, especially when it comes to large facilities such as stadiums, airports, and large shopping malls. Given the pandemic’s effects, it seems that developing palm scanner software will be a frequent request from hospitals, food facilities, pharma manufacturers, big retailers, and basically anyone with security or hygiene concerns.
Given this, a new question arises: what does the technical challenge of building a palm recognition solution like Amazon One look like? And should a “pay with your palm” system be your company’s next investment?
While Amazon hasn’t detailed how their palm scanner works and the parts it consists of, it’s possible to make an educated guess about the technology inside a biometric payment system like this.
Let’s start with what we know: palm recognition solutions like Amazon One are based on both palm print and vein authentication. These are two separate processes.
There are two approaches to palmprint recognition based on image data type, whether grayscale image or 3D multispectral image ( 3D is also used for face and iris recognition). The system uses several palm features:
The recognition process consists of two phases: training and testing. The first phase involves capturing the palm with the biometric scanner and generating an image. This image is then cleaned of background noise and used as training data. The result of this phase is the extracted feature data. In the testing phase, the system captures a palm image as before, but then the extracted features are compared to the training data to determine a match.
Here’s how it happens in more detail:
If it seems that a palm print recognition system is not that complicated, note that every ROI feature has to be detected with a different recognition technique based on the type of the feature. And if we return to our Amazon One model, it combines both palmprint and vein recognition.
Biometric sensors capture veins with the help of near-infrared light using a technique called near-infrared spectroscopy. To be more specific, hemoglobin without oxygen absorbs the rays, and then the palm scanner takes a photo of the light that scatters back from your palm. The scanner can use one of two methods: either reflection or transmission.
After the image has been captured, the palm recognition solution uses image processing to extract features and compare patterns, similar to palmprint processes.
Getting back to our Amazon hand scanner, it captures both palm print and veins, encrypts the ROI to reduce data size and provide faster access, and fuses the data from both images with the help of a biohashing algorithm. As a result, the system gets a biohash code and compares it with the templates in its database using matching algorithms. More specifically, the input palm is compared to all the registered templates in the database.
These matching algorithms are highly accurate. This makes the palm recognition solution both reliable and secure, which means it’s well suited to meeting stringent hygiene demands and fintech security requirements.
The high accuracy and security of the Amazon authentication system make it an attractive solution for many companies in fintech, healthcare, retail, and more. Amazon has not revealed any third-party agreements, but we can assume it won’t be long until we hear about the wider deployment of palm scanners. Here are just a few potential applications of contactless palm recognition systems in different industries:
With Amazon actively promoting its palm recognition software, we can expect to see more use cases in all of these areas shortly.
Amazon One looks set to be the solution for making offline shopping more attractive to customers in a post-pandemic world. What’s more, it’s a test case that will probably bring more contactless tech into retail, finance, healthcare, and other fields. Paying with your palm is fast, convenient, and safe—what more could consumers ask for? So why not consider this option during your next financial app development?
At Relevant, we develop complex custom software for sectors such as fintech, healthcare, retail, and travel. Just get in touch, and let’s discuss your ideas!
Do you know that we helped 200+ companies build web/mobile apps and scale dev teams?
Let's talk about your engineering needs.
Write to us