Things You Probably Don’t Know About Biometric Technology

Despite the increasing use of biometric recognition in day-to-day life, people generally do not know about many aspects of this technology, which lead to misinformation and confusion. When it comes to biometrics, people generally relate it to fingerprint, face, voice and iris recognition. Biometric technology, however, neither limited to these popular modalities, nor as superficial as people know it to be. There are many aspects that people ordinarily do not know about biometrics. This article sheds light on some of these uncommon aspect and facts associated with biometric technology.

A few decades ago, biometric technology was primarily used either in forensics or deployed only in specific applications. However, personal recognition using biometrics has become a commonplace now. It may seem as if biometrics has emerged in recent years, the core technology, however, is way older than you may know. We are not talking about a couple of years or decades here, roots of biometric technology date back to the early 19the century. The beginning of 19th century witnessed increasing adoption of fingerprint recognition in forensics and law enforcement applications.

Before fingerprinting, an identification method called Bertillon system prevailed in some parts of the world for criminal identification. It was the first identification method somewhat similar to biometric identification. This system was based on the science of anthropometry, which focuses on the meticulous measurement of different parts of the human body. Alphonse Bertillon, who developed this system, believed that these measurements can be used to uniquely identify a criminal. Bertillon system was later replaced by fingerprint identification due to many inadequacies and cases of misidentification of criminals (e.g. in popular Will and William West case, it was found that two persons can have similar physical measurements).

So, despite the recent emergence of biometrics, this technology has a history of more than hundred years.

Most people, even the end users, who are not associated with the field biometric technology in any way, know it to be a technology that makes use of fingerprints, face, voice or eyes to recognize the users. The level of familiarity with biometrics has improved recently due to rise of biometric recognition on mobile devices. Biometric recognition, however, is not limited only to a few popular modalities. There are several types of biometric modalities, even those which are very unusual and are still in development. DNA profiling, vein pattern, gait pattern, behavioral pattern, etc. are some of the biometric modalities used in specific applications.

• Pattern formed by lips
• Body odor
• Shape of ear
• Gait
• Thermography

There may even be more biometric modalities yet to be studies. Some biometric modalities may have all characteristics to be a biometric modality but are not leveraged due to technological limitations.

More information on uncommon biometric modalities can be found here. We have also captured several types of biometric modalities here.

Superficially, it may seem as if a biometric system scans or capture images during the enrollment and match those images during identification or authentication process. However, this common perception is only partially true. Biometric systems do capture or scan image of a person’s biometric identifiers, however, they do it to extract the features of the scanned identifier. Before doing that, these images are enhanced and processed to generate a machine-friendly image and the recognition algorithm finally generates a digital template, which neither looks anything like the original image nor can be reverse engineered to generate the original image.

So, contrary to the popular belief, biometric systems do not directly work on images to match a biometric identifier.

Be it any type of biometric application, it has either identification or verification of identity at its core. Biometric technology has been deployed in a variety of applications in a number of industry types, government organizations, institutions and even in domestic use cases. Today, biometrics is used by border control agencies to identify international travellers while businesses use it for employee time and attendance to network security and access control. Payment apps are also using the same technology to identify their users. From ATM security to credit card security, biometric technology has taken over traditional identification approaches.

New applications of biometrics are now emerging faster than ever and it is expected to be deployed in more and more use cases. Innovative use of biometrics is expected to keep us amazed for many more years to come.

Other than popular biometric identification approaches like fingerprint, face and voice, there are certain modalities that has been identified, developed and patented by technology firms. These identification technologies can only be used by the patent owners or by their patent licensees. No other companies can develop the similar systems so that intellectual property rights of the patent owners can be protected.

For example, EyeVerify, a USA based biometric solution provider, has developed recognition system called Eyeprint ID, which makes use of blood vessels pattern visible on the whites of the human eye and other micro-features of the human eye for personal identification and authentication. Fujitsu’s finger-vein recognition is also one of such patented technologies that can only be produced by Fujitsu or a partner company with the patent licence.

There are many other biometric solution providers that offer patented biometric solutions, which no other tech firm can produce or make use of.

We know biometric authentication to be instant. We witness it happening on phones, computers or the office attendance machine everyday. Biometric identification, however, is the entirely different story. Authentication is a 1:1 process, in which a user claims an identity and the biometric sample acquired from the user is matched against the stored identity data. In biometric authentication, users also provided with unique information that they can present at the time of authentication. This information is generally a unique number, which is presented to the biometric systems to fetch the user’s identity record. Then the users provide his/her biometric sample (e.g. finger scan, iris scan, etc.) and this scan is matched against the fetched one.

On the other hand, biometric identification is the process of looking up identity data of an unknown biometric sample. This biometric sample is looked up against all available identity records. In this process, user may or may not be present; and if the user is present, he/she does not claim any identity. His/her identity data is searched in the database of already identified biometric samples. This is also done in case of latent prints collected from a crime scene. This process can take from a few seconds to several minutes, even hours, depending on the size of the database and efficiency of the system.

Most people think that biometric recognition on phones was first introduced by Apple with the launch of iPhone 5s. Biometric fingerprint recognition on iPhone 5s was a huge success; however, it was not the first time that a phone manufacturer had offered biometric ability on a mobile device.

Pantech, a South Korean phone manufacturer, had done it way back in 2004 with their Pantech GI100 feature phone. Surprisingly, it was not even a smartphone. Later, other manufactures also offered fingerprint sensors. Motorola was the first to introduce it on Android with Motorola Atrix handset. It was also launched before the iPhone 5s.

So Apple may be the first to do it right, but certainly not the first to introduce biometric recognition on mobile phones.

Did you know that the way you operate your devices like smartphones, tablets, etc. is a type of biometrics? Even your typing rhythm on a keyboard can be used as a biometric identifier. The usage pattern on these devices is considered to be persistent and unique once the user has developed the habit of using it in a particular way. Using device and sensor data like how a user holds the phone, swipes the screen, pressure and size of the finger, etc. along with behavioral biometric techniques, a user profile can be created which will be unique to the user.

This technique is being used by many banking and financial service organizations to safeguard their users as well as service from unauthorized access. For example, a bank may choose to incorporate a behavioral biometric solution in their mobile banking app. This app will not have any sign that it is equipped with behavioral biometrics and users may sign-in using their credentials like just any app/service. However, when behavioral biometrics comes into play, it can enroll users in the background and make sure that the device is in right hand. If it detects anything unusual, it can ask the user to re-verify his/her identity.

Despite being super accurate, fast, and more efficient than other identification and authentication approaches, biometrics is not one-size-fits-all solution.

Though biometrics can identify people on the basis of “who they are”, it may not cover everyone in the target population. Larger the population, greater the possibility that some subjects will fall out of coverage. For example, if identity of a large group of people needs to established using their fingerprints, not all of the individuals may be assumed to have their fingerprints in biometrically usable condition. Due to friction, certain chemicals, injury or skin diseases, people may lose their fingerprints partially or completely. Some work conditions may require people to repetitively touch objects (e.g. manual loading/unloading of goods) that may cause people to wear off their fingerprints.

Many other biometric modalities like retina, iris, voice, etc. also suffer from the problem of population coverage and one modality may not work for all. In that case, other biometric identifiers that are good enough to be leveraged, can be collected.

Did you know that AI (Artificial Intelligence) powered biometric systems are going to be a reality?

AI is marking its presence in many applications and systems to improve their performance, efficiency and reliability with usage. It is going to be the next big thing in the field of information technology, embedded systems and robotics. AI based systems and applications do not require explicit improvements by human intervention, they can learn with user interactions and improve themselves over time.

Modern biometric systems have been already making use of machine learning techniques to improve decision making, biometric fusion, etc. Subspace projection, fusion methods, distance to template, etc. are some of the commonly used machine learning techniques on biometric systems.

Manufacturers are also looking at future artificial intelligence applications to enable these systems to improve themselves with usage. AI can help improve security and user experience by understand variations in fingerprint sample. It will not only improve the fault tolerance but can also help reducing probability of false acceptances and rejections.

When it comes to IoT, people mostly talk about connected vehicles or appliances and biometric devices get completely overlooked. Despite the ignorance, IoT is coming to biometrics.

Imagine a world where your phone can communicate with your biometric door lock and remotely grant access to a friend who you will not be able to attend personally. In another instance, you book a driverless taxi on your phone app (authenticated with your fingerprint) and it allows you to get in by scanning your fingerprint you used during booking the taxi.

Presently, it may sound like the story of some sci-fi movie however, this is going to be the reality in some point of time in future. This will be made possible by IoT or the Internet of Things.

Present day internet is mostly the internet of information connected with IT system. IT devices do communicate with each other; however, these devices are mostly computers, peripherals and devices directly connected to a network or the internet. However it is still uncommon to see a home air conditioning system communicating with a car (so that it can turn on automatically when it finds the vehicle approaching home). The scenarios discussed above are very practical cases, which will be commonplace in the future. Most things will work automatically and you will not have to turn them on, off or set instructions to make them work. They will also improve themselves with usage (thanks to machine learning and AI as we discussed above).

IoT is generally associated with vehicles and appliances, however, industry specialists know that IoT can make dramatic changes in biometric technology and improve security, efficiency and take convenience to the next level. You probably did not know that biometrics is IoT ready, did you?

Biometric technology systems have already claimed a huge chunk of market share in identification and authentication applications and it continues to grow. Biometric technology improves security as well as convenience; however, people tend to overlook many aspects associated with this technology. Knowing these aspects not only gives us a chance to have deeper understanding of this technology; it also helps us take informed decisions when it comes to choosing the right technology solutions.

Biometric technology, however, is still not limited to aspects covered in this article. We will keep coming up with more to keep our readers informed.