While a part of the world is heading into the realm of crime. Nothing is completely secure, the age-old, as well as the latest locks, can be broken into, safes can be picked, and digital passwords can be deciphered and hacked. The other part of the same world is getting secure with the increasing use of biometrics.
The latest biometric systems are reliable for both; monitoring and authenticating in day-to-day activities. Out of all biometrics, the most widely used is the use of fingerprint biometrics. The most significant factor behind the popularity of the technology is the perfect combination of security it provides along with the convenience of usage. Today’s blog will focus on this particular technological marvel, the history of biometric fingerprint technology, how it functions and the accuracy biometric fingerprint technology provides.
How Biometric Fingerprint Technology Works
First things first, if you are about to implement biometric fingerprint technology for your business or your residence, you should understand how the technology actually works. This section will throw light on how a biometric fingerprint system reads, processes and assess a fingerprint.
A control board is present in these biometric fingerprint systems that processes the scanned fingerprints and assesses and authenticates from a database and then proceeds to depend on the function; which can be either to record the entry and/or to garner entry to a particular function. This control board will be the central communication point between the biometric scanner and the other units of the system.
Each biometric fingerprint device consists of a scanner that provides fingerprint scans for the machine to read the details on a fingerprint so that it can differ between the biometrics of one person from another.
There are three main methods of scanning fingerprints. Each of these methods uses a different technology to process the image, make the right calculations and then compute the image of the finger.
An optical scanner works by throwing light on a fingerprint and then taking a digital photograph of the finger pressed upon the scanner using a CCD or a CMOS sensor. A typical scan from an optical biometric fingerprint scanner will look somewhat like a photocopied fingerprint, but a detailed crisper image for that matter. The integrated computer will then process the image using pattern-matching software and convert the fingerprint into a code that will be used in the biometric fingerprint device.
If you are a tech-savvy person, you might have heard about capacitive touch screens on the smartphones we are using in this era. Capacitive scanners on biometric fingerprint systems work in a similar fashion.
A capacitive scanner measures a finger electrically when a finger is pressed upon the scanner’s surface. If you look closely at your fingers you will find micro ridges on your skin. These ridges touch the scanner surface while the gaps between the same ridges will not. Now the capacitive scanner will make an image of the fingerprint by calculating the distance between these ridges. The benefit of using a capacitive scanner over an optical scanner is the speed it provides. But at the same time, a capacitive scanner will not be able to work effectively in the presence of moisture.
As the name suggests, ultrasonic scanners use high-frequency sound waves to create an image of a fingerprint instead of the conventional use of light to map the fingerprint. Though these types of scanners are usually slower than optical or capacitive scanners ultrasonic scanners provide the highest level of security as they scan the images in three dimensions against regular two-dimensional images procured in the two methods mentioned above.
Using these biometric fingerprint scanners, the control board scans and collects the relevant personnel data which it uses later.
History Of The Development Of Biometric Fingerprint Recognition
The term biometric is derived from the Greek words, “Bio” which means “life” and “metrics” which translates to “measure”. The history of fingerprint biometrics can be dated back more than thirty thousand years when fingerprints can be found in cave paintings by prehistoric men who used them as signatures to authenticate.
1788- Fingerprints Deemed Unique
In 1788, a German anatomist, JC Mayer explained how ridges on fingerprints in each person no matter how similar but are always unique, in a published paper “Anatomical Copper-plates with Appropriate Explanations”
1858- First Systematic Use of Handprints
Again, a civil worker in British-occupied India, William Herschel captured a handprint to make sure that poverty-stricken people do not claim to be factory workers and collected wages in the British industries at the time. This method is the first systematic use of handprints for authentication.
The 1870s- The Bertillon System
The French criminologist, Alphonse Bertillon devised the Bertillon system that was used to identify people on the basis of measurements of their body parts. This system was specifically used to manage a record of serial offenders.
The 1880s- The Henry System
But it took more than a hundred years after the publication of JC Meyer’s paper when in the 1880s fingerprints started to be used as a way to identify criminals as well as to verify signed contracts when an Indian with the name Azizul Haque created the first reliable fingerprint indexing system in India for Inspector General of Police in the state of Bengal, Edward Henry. This system was given the name, “The Henry System”.
1903- US Penitentiary Uses
In 1903, the New York Civil Service Commission began fingerprinting the applicants in order to prevent test fraud in the state. The same system was later copied by the State Prison System for the identification of criminals.
1969- FBI Involvement
By 1969, the Federal Bureau of Investigation (FBI) understood the applications of biometric fingerprint recognition and how it can be used effectively in multiple avenues in day-to-day life. They started financing research to develop a system to automate the process to identify biometric fingerprints. This funding brought in the National Institute of Standards and Technology (NIST) to study the process.
1974- First Commercial Systems for Hand Geometry Recognition
The world began adopting biometric fingerprint systems for purposes other than preventing fraud and managing criminal databases. These commercial uses mainly included attendance in factories and offices.
1999- Recognition of Biometric Fingerprint as an International Standard
Till now, the use of biometric fingerprints was mainly limited to the United States of America, but then the International Civil Aviation Organization started a study on the use of biometric fingerprint technology in the MRTD (Machine Readable Travel Document) inspection processes around the world, making it an internationally recognised system.
2013- Apple introduced Touch ID
Biometric fingerprint technology came into the hands of the direct everyday consumer when apple introduced Touch ID in their devices. The touch ID allowed users to unlock their devices as well as authenticate purchases on their iTunes account.
Several companies followed Apple and it became a norm to have biometric fingerprint technology in smartphones.
Facial Recognition Algorithm Assessment
Besides assessing the accuracy of fingerprint algorithms, NIST also provide assessment on facial recognition algorithms, another very popular form of biometric. Vendors often showcase the accuracy of facial recognition algorithms by participating in frvt 1:1 verification testing. Due to the number of test variations being so diverse, there are plenty of example recipients that showcase their variety of facial recognition algorithm production. Neurotechnology are an example recipient, for more information visit here.
Accuracy of Biometric Fingerprint Devices
This is a known fact that biometric fingerprint technology provides extremely high accuracy, but not many know that there are several factors that should be taken into consideration. Recognition accuracy generally depends on factors such as input image quality and the matching algorithms used by the developers of the hardware and the software of the biometric fingerprint devices.
While the probability of false positives i.e. providing access to unauthorised personnel is nearly zero, nonetheless chances of a false negative i.e. not providing access to authorised personnel is a thing that most of us have experienced first-hand. It should also be noted that the chances of false positives can be avoided to almost the extent that the biometric fingerprint recognition systems can provide an accuracy of 99% using live scanning.
Scientists and researchers at the National Institute of Standards and Technology (NIST) have recently tested 34 commercially available systems from around the world and the results of the test were phenomenal. Out of all, the best system came out to an accuracy of over 98.6 per cent on single-finger tests, 99.6 per cent of the time on two-finger tests, and came out to be correct at 99.9 per cent in all the tests involving four or more fingers. While the calculated percentage for false positives stood at a rate of 0.01 per cent in the same tests performed by NIST.
How Does NIST Evaluate The Algorithm?
Biometric fingerprint recognition can not be termed 100% accurate. The accuracy of a biometric system can be evaluated by using well-known performance indicators. The most common performance indicators are False Accept Rate (FAR), False Reject Rate (FRR), and Equal Error Rate (EER). Depending on the application and test design, different evaluation metrics are appropriate
NIST uses the Fingerprint Vendor Technology Evaluation (FpVTE) to evaluate algorithms and grade and asses assess the capability of fingerprint systems to meet requirements for both large-scale and small-scale real-world applications by putting to test fingerprint matching, identification, and verification systems provided by the developers.
NIST, in their verification ROC (Receiver Operating Characteristic), uses the 1:1 matching performance statistic to measure and compare the accuracy of systems for small-scale biometric fingerprint scanning purposes. On the other hand, for the large-scale fingerprint matcher, NIST applies the 1:N strategy, which helps them in putting out many candidates in a faster manner to minimize the overuse of multiple resources.
Apart from the type of test strategy, another important factor that NIST takes care of while performing these tests during the Fingerprint Vendor Technology Evaluation is the choice of sample data as some devices might work efficiently in a particular set of data while others may lack the same. FpVTE consists of multiple tests performed with combinations of fingers i.e. single fingers, two index fingers, four to ten fingers and different types and qualities of operational fingerprints i.e. flat livescan images, multi-finger slap livescan images.