What is Your Favorite Biometric Modality?
I asked a bunch of people (via Twitter, to be honest), “For being identified before being allowed access to a public space (airport, courthouse, etc.), which #biometric would you prefer to use?” Here are the results:
It’s fascinating to me that people gravitated towards fingerprints as the preferred biometric modalities. I did not expect that at all. I thought that with face and iris recognition coming out on phones and the general increase in speed and convenience involved, that those modalities would be more popular. But "Minority Report" aside, fingerprints are still the biometric most people are familiar with. We know that criminals get fingerprinted, and we see references to fingerprint databases and latent prints on popular TV shows. A lot of people have been fingerprinted for job related background checks. We like things that are familiar. So which one really is best?
There are three main biometric modalities in use today. They are fingerprint, face image and iris image. But there are many more biometric modalities in existence. Any human trait that is unique enough to be usable to identify people can be a biometric modality. For example, there are knuckle biometrics, palm vein, walking gait biometrics, typing biometrics as well as ear biometrics. Biometrics for identification have advantages over traditional methods like passwords, PINs or physical cards like a driver’s license. Biometrics are something you are, so you can’t lose or forget them. (There are disadvantages too, but we’ll talk about those another day.)
The reliability of each biometric modality is different and depends on factors such as ergonomics, lighting, age and ethnicity. People who work with their hands can lose their fingerprints, for example, and a very small number of people just don’t have fingerprints that are usable with current sensors. Different industries and environments demand different types of biometric modalities based on different situations.
There are other factors that influence which modality is best. The most accurate biometric in a lab environment is probably not the best for workers at a construction site. Human factors, performance and price are all important considerations in the real world. Some important factors to consider include:
- Ease of Use
- User acceptance
Maybe "What is your favorite biometric modality?" is kind of a nerdy question. I mean,*I*have one, but I'm not surprised that most people do not. To help you pick out *your* favorite, here is a summary of the most popular biometric modalities:
If you have ever watched a TV show, you probably are familiar with fingerprints. They are the most recognizable biometric modality around the world, but outside of a law enforcement applications, more accurate, hygienic, and reliable modalities have been invented. The uniqueness of this modality is determined by mapping the ridges and furrows on a finger. This modality is practical and convenient for identification and the first commercially widespread biometric sensor which is why it is widely used in law enforcement, forensic departments, government projects, banking, and workforce management.
Examples of current fingerprint deployments include:
- Fingerprints collected at crime scenes to help identify a criminal
- Most cell phones use fingerprints as an unlock mechanism
- Iraq used fingerprints on election day to solve multiple voting fraud
The pros and cons of this modality are:
- Relatively inexpensive
- High availability
- Unique: Everyone including identical twins have unique fingerprint patterns
- High accuracy
- Poor skin integrity, ethnicity, age, and environment can negatively affect results
- Enrollment is difficult
- Easier to spoof
- Negative end user stigma associated to criminality
Face recognition is familiar because everyone has experienced having their picture taken. We are also very familiar with face recognition in general - it's a basic skill we are all born with and use to identify our friends and family. It's very accurate in controlled environments and less accurate but still good enough "in the wild" (when the lighting and pose angle is not controlled).
Examples of current face recognition include:
- Every time your ID is checked when entering a bar
- Face is now part of the FBI criminal database
- Some cell phones use face recognition to unlock
- Ease of Use: having your photo taken is easy and familiar
- Can be captured at a distance
- Accuracy is dependent on controlling lighting and pose
- Face will change over time more than other modalities
- Some systems are easy to spoof with a picture
- Can be captured without user's consent or knowledge
Iris recognition is widely considered to be the fastest and most accurate method of biometric identification. It works by capturing photos of your eyes and maps your unique iris pattern to verify your identity. Iris recognition has a strong acceptance rate because it’s non-invasiveness and non-contact. This means it’s more hygienic in nature. Places such as hospitals (where providing accurate care and protecting privacy are paramount) increasingly rely on iris recognition for patient identification. This modality is also used in highly secure environments like military bases, data centers, bank vaults, border security, passport checks, and national ID programs.
Examples of current iris recognition deployments include:
- UAE’s land, air, and sea ports of entry
- India’s UID Initiative where registration for the Aadhaar card includes iris recognition for multi-factor authentication
- Google uses iris scanners to for data center access control
- Accuracy: 1 in 10⁷⁸ chance that iris pattern of two individual matches
- Stability: The structure of the iris remains the same throughout our lifetime
- Fastest matching of any modality
- Relatively expensive, requiring high resolution cameras
- Usually requires coordination with user
- Unfamiliar to users, confused with invasive retina scanning
Every biometric modality has its own set of benefits and drawbacks. Some are just considered to be “cooler” than others. Fingerprints may currently be the most popular modality, but iris recognition has the most promise for the future of biometric recognition because of its accuracy and positive reputation. Face is still my favorite modality because the sensors are widely available and people are familiar with having their picture taken.
Thank you to those who participated in the poll and provided us with the input for this post. I had some interesting conversations with people who adamatly opposse any use of biometrics - that is another blog article for sure! Please follow us on Twitter for details on our next biometric related poll! Participate and be a part of our community!