AI / Machine Learning
-
March 1, 2022

Fingerprint Recognition: the most popular Biometric

Fingerprint recognition is one of the most commonly used and secure biometric technologies today. The speed and ease of fingerprint recognition make it one of the best biometrics for verification purposes.

Fingerprint recognition is also widely used for identity applications, best exemplified by the databases administered by the FBI to track and identify people of interest. 

In this article, we will look at the following key aspects of fingerprint recognition technology:

  1. Fingerprint recognition history.
  2. Characteristics of fingerprint recognition systems.
  3. How fingerprint recognition systems analyze data.
  4. Specifications and quality checks in fingerprint recognition systems.

The history of fingerprint recognition as a biometric

As a form of biometric, the use of fingerprints has been around since 2000 BC, when it was used as a form of signature in business transactions.

The first true research paper that attempted to examine the unique structures of the fingerprint was published in 1694, and the first fingerprint classification system evolved in 1823 by a scientist known as Jan Purkinje.

At the turn of the 20th-century, law enforcement agencies in the United States started to use fingerprints as the primary means to track down criminals. 

To automate fingerprint-based searches across levels of law enforcement, the FBI devised the Integrated Automated Fingerprint Identification System, or IAFIS, which has enhanced features to keep track, compare, and identify individuals all across the world.

Fingerprint recognition
Fingerprint recognition

Now that we have a little background on fingerprint recognition as a biometric let's go into how a fingerprint recognition system would work today.

The levels of fingerprint recognition

In the world of biometrics, the details of the fingerprint are broken down into three distinct levels:

  1. Level One: At the first level, fingerprint recognition software will image the fingerprint patterns as a whole, rather than emphasising the specifics of the print.
  1. Level Two: The minutiae points of the fingerprint are gathered in greater detail, from which the bulk of the unique features is extracted.
  1. Level Three: This includes the shapes and the images of the ridges and their associated pores. This is the most sophisticated and detailed level and is generally not widely adopted in use cases. 

How does fingerprint recognition work?

Fingerprint recognition works by capturing an image of a fingerprint, emphasizing its unique patterns, shapes, and details. This image is then compared to a database or existing image of the fingerprints to verify an individual. 

Biometric fingerprint recognition has been widely used for years, if you’ve owned a smartphone over the last six to seven years, odds are you are more than familiar with using a fingerprint sensor. So, let's go a bit deeper into how the fingerprint recognition system works.

It is important to note that most biometric-based fingerprint systems collect images at Levels 1 and 2, the more powerful fingerprint recognition systems collect Level 3 details. Fingerprint recognition systems consider a wide variety of details when verifying a user. 

The Level 1 specific features include the following:

  • Arches - These are the ridges that flow in one direction, without doubling back, or going backwards. These comprise about 5% of the features on a fingerprint.
  • Loops - In this feature, the ridges go backwards, and go either left to right or vice versa. There are two distinct types of loops; the radial loops that go downward, and the ulnar loops that go upwards on the fingerprint. The loops make up 65% of the features within the fingerprint.
  • Whorls - The ridges in the fingerprint make a circle around a core comprising 30% of the features in the fingerprint.
Fingerprint recognition features
Fingerprint recognition features

In addition to the above features collected by a fingerprint recognition system, the number and orientation of ridges can also be used as part of the data in fingerprint recognition systems when verifying an individual.

Other distinctive features which can be extracted, but are not as commonly used, include the following:

  • Prints/Islands - These are the very short ridges found on the fingerprint.
  • Lakes - These are the unique indentations and depressions located right in the middle of the ridge.
  • Spurs - These are the actual crossovers from one ridge to another.

The steps of fingerprint recognition software

Fingerprint recognition software generally follows a distinct methodology which can be broken down into the following steps. 

  1. Raw Data Acquisition

The first phase of the fingerprint recognition system is a process we are all familiar with. At this stage, the user presses their finger down on a sensor to capture images of the fingerprint, like on a smartphone.

The raw images of the fingerprint are acquired through the sensor, whereby a quality check is passed or failed. This means that the biometric system examines the raw images to find extraneous data in the fingerprint image, interfering with the print acquisition.

If an obstruction is found on the image, the fingerprint device will automatically discard the image and prompt the user to try placing their finger again. If the raw images are accepted, they are processed and extracted in the next step.

  1. Extraction 

With the raw images accepted in the previous step, the image needs to be extracted and processed to map the unique features of the print. Once those details have been extracted and stored, the system can move to the next step.

  1. Template Creation

After the successful extraction of the details, the data becomes the enrollment template of the print. The enrollment and verification templates are then compared to one another, to determine the degree of similarity or non-similarity in the final step.

  1. Comparison

Here the enrollment template and verification template are compared against each other to validate a match. The comparison is broken down into a percentage of how similar the fingerprints are. The comparison is valid if the percentage rate is above a set threshold.

Optimized fingerprint recognition systems

Every image taken of the fingerprint must be of a high enough quality so the user can be easily verified. There are specific aspects to a fingerprint recognition system and its software that must meet a baseline of criteria to ensure that the imaging of the prints is satisfactory. 

Fingerprint recognition system
Fingerprint recognition system

Below is a list of the most crucial parts of the system that need to be optimized to ensure this quality.

  • Resolution - This refers to the total number of dots per inch (DPI) or pixels per inch (PPI). Most fingerprint algorithms require a DPI or PPI resolution of at least 250 to 300, and for the FBI, a minimum of 500 is required.
  • The number of fingerprint pixels -This refers to the total number of pixels in the scanned image.
  • Area - This is the actual size of the scanned and captured fingerprint image. The specifications set forth by the FBI mandate a minimum scanned fingerprint area of at least one square inch.
  • Frames per second - The total number of raw images the fingerprint recognition device sends to the processing unit every second. The higher number of frames per second means a much greater tolerance for any unwanted movements of the fingerprint during the imaging phase.
  • The dynamic range - This refers to the possible ranges available for the encoding of each pixel value. The FBI mandates at least eight bits.
  • Geometric accuracy - These are the geometric differences between the enrollment and the verification templates. This is calculated via the deviations from the X and Y axes on the fingerprint template.
  • Overall image quality - This refers to the identifying of unique features in the fingerprint, such as the ridgeline patterns and the other characteristics of the fingerprint we discussed earlier.

The most critical aspects of fingerprint recognition

Let’s recap some of the most important aspects of a fingerprint recognition system. 

First, there are the key characteristics and details of a fingerprint that we touched on. Some of the most important aspects of a print are the Arches, Loops, and Whorls. Then for a more detailed fingerprint recognition system, it's also possible to track the Islands, Lakes, and Spurs. 

There are four steps the system must go through for every fingerprint recognition system to verify the fingerprint. They are:

  1. Raw Data Acquisition - the imaging of the fingerprint by the sensor of the device. 
  2. Extraction - the mapping and identification of the unique characteristics in the fingerprint
  3. Template Creation - the actual data of the fingerprint is created
  4. Comparison - the template data is then compared to the existing template to verify the user

Finally, it's important to emphasize that every high-quality fingerprint recognition software maintains the minimum requirements for an effective recognition system. The main criteria and specifications for a fingerprint recognition system are.

  • Maintaining a high resolution of the image either measured in DPI or PPI.
  • An overall acceptable range of fingerprint pixels in the total image.
  • A sufficient area of the print is covered to capture the complete print.
  • The number of images taken of the print every second.
  • The encoding of each pixel value into bits through the dynamic range.
  • The level of geometric accuracy when comparing the two templates.
  • The overall quality of the image and how easily viewable the characteristics of the print are.

Given all that we have covered on the benefits of this biometric, it's highly likely that fingerprint recognition will most certainly continue to be used throughout the tech and government industries for verification purposes.