The previous article in this series served as an introduction to fingerprint recognition. This article will delve into the quality control checks that are necessary to ensure accuracy in the use of fingerprint recognition.

The need to collect high quality, raw fingerprint images the first time is of paramount importance. The following are the quality control checks for fingerprint recognition:

  • Resolution:  This refers to the total number of dots per inch (DPI), or also known as the total pixels per inch (PPI). Most fingerprint algorithms require a DPI or PPI resolutions of at least 250 to 300 DPI, and to meet the stringent requirements of the FBI, a minimum of 500 DPI is required.
  • 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. Anything less than that will not produce a good quality, raw image.
  • Frames Per Second:  This is the total number of raw images the fingerprint recognition device sends to the processing unit. Of course, the higher number of frames per second means a much greater tolerance for any unwanted movements of the fingerprint on the platen.
  • The Number of Fingerprint Pixels:  This refers to the total number of pixels in the scanned image.
  • The Dynamic Range:  This refers to the possible ranges which are 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, and this is calculated via the deviations from the X and Y axes on the fingerprint template.
  • The Image Quality:  This is the variable which refers to the identifying, unique features in the fingerprint, such as the ridgeline patterns and the various minutiae which are extracted.

Methods Of Fingerprint Collection

Fingerprint images can be collected using one of two methods: offline scanning or the use of Live Scan Sensors.

Offline scanning methods have been in existence since fingerprints have been used in law enforcement. These traditional methods have involved using an inked impression of an individual’s finger, and then placing that impression onto a piece of paper. Subsequently, cards were utilized instead of paper to store the inked fingerprint image. These were kept in file cabinets, and when a suspect was apprehended, their fingerprints were then compared to the inked fingerprint image stored on the card.

Over time, as technology improved, these two-dimensional images were then scanned into a digital format, translated into a 500 DPI, to meet the FBI requirements and specifications. But, with today’s fingerprint recognition systems, a law enforcement officer in the field with a wireless biometric device can scan a suspect’s fingerprint, and then have that image automatically uploaded into a central server, from where it can then be converted into a proper fingerprint digital image.

As one can imagine, the use of these traditional methods definitely has its fair share of flaws. For instance, it takes great skill to get a good, inked impression of a fingerprint, too little ink used means that the print area of a particular finger can be missed, and too much ink used could very well obscure any features which were attempted to gather. But the offline methods do possess one great advantage over the live scan method. And that is, a “rolled” impression of the fingerprint can be gathered, which means that a much larger image of the fingerprint can be captured, versus the much smaller scan areas of the live scan sensors.

To this degree, the question often gets asked, is how are the latent fingerprints captured?  Latent fingerprints are those which are left behind at a crime scene and are collected later in the course of an investigation. Offline scanning methods are used to collect these latent fingerprint images at a later point in time, and chemical reagents and processes are often used to collect the fingerprints.

With live scan sensors, two methods of collecting fingerprints images are used: touch and sweep.

With the touch method, direct contact is required of the finger onto the biometric sensor. But using this type of method possesses some grave disadvantages. For instance, the platen can become dirty or smudged quite easily, any fingerprints left behind on the sensor or platen by another user previously can lead to substantial errors, and in terms of dollars and cents, there is a positive correlation between the size of the scan area and the actual cost of the sensor. Also, some sensors may not be able to capture raw images if the finger is off by more than 20 degrees from the platen.

Unlike the touch method where only finger is scanned, the sweep method captures images of multiple fingerprints as they are read by the fingerprint sensor, and from that point, multiple images are captured at the end of the sweep. From these multiple images, one composite image is created. The quality of the raw images which are captured are also partially dependent (as well as the other variables just described) on the accuracy of the reconstruction algorithm whose primary function is to compile the multiple, raw images into one image of the fingerprint.

The biggest disadvantage of the sweep method is that a much higher rate of error is introduced due to the variance in the sweeping rates and the angles. It is the sweep method, which is primarily used by fingerprint recognition devices today, and there are three types of sensors which are available:

  • Optical sensors:  These are the most commonly used.
  • Solid state sensors:  The image of the raw fingerprint is captured onto a silicon surface, and the resultant image is then translated into electrical signals.
  • Ultrasound sensors:  This is where acoustic signals are sent towards the finger, and then a receiver subsequently digitizes the echoes from the acoustic signals.

The Matching Algorithm

As mentioned earlier, it is the matching algorithm which compares the enrollment template with the verification template, and in order to ascertain the degree of similarity or closeness between the two, a certain methodology must be followed. That methodology is as follows:

  1. Whatever data is collected from the raw image of the fingerprint must have some sort of commonality with the enrollment biometric template which is already stored in the database. This intersection of data is known as the core, which is also the maximum curvature in a ridgeline.
  2. Any extraneous objects which could possibly interfere with the unique feature extraction process must be removed before the process of verification/identification can occur. For example, some of these extraneous objects can be the differences found in the size, pressure, and the rotation angle of the fingerprint, and these can be normalized and removed by the matching algorithm.
  3. In the final stage, the unique features collected from the raw data (which becomes the verification template) must be compared to that of the enrollment template later. This is the point at which the matching algorithm does the bulk of its work. The actual matching algorithm can be based upon the premise of three types of correlation:
    • Correlation Based Matching: When two fingerprints are overlaid on each other, or superimposed, differences at the pixel level are calculated. Although it is strived for, perfect alignment of the superimposed fingerprint images is nearly impossible to achieve. Also, a disadvantage with this correlation method is that performing these types of calculations can be very processing intensive, which can be a serious strain on computing resources.
    • Minutiae based matching: In fingerprint recognition, this is the most widely used type of matching algorithm. With this method, the distances and angles between the minutiae are calculated and subsequently compared with another. There are methods that focus on global minutiae matching as well as local minutiae matching, and the latter method focuses on the examination of a central minutiae, as well as the nearest two neighboring minutiae.
    • Ridge Feature Matching: With this matching method, the minutiae of the fingerprint are combined with other finger based features such as shape and size, the number and the position of various singularities, as well global and local textures. This technique can be of great value if the raw image of the fingerprint is poor in quality, and these extra features can help compensate for that deficit.

Up Next: The next article in this series will examine the advantages and disadvantages of Fingerprint Recognition, as well as its commercial applications.

Join the conversation.

Keesing Technologies

Keesing Platform forms part of Keesing Technologies
The global market leader in banknote and ID document verification

+ posts

Ravi Das is a Cybersecurity Consultant and Business Development Specialist. He also does Cybersecurity Consulting through his private practice, RaviDas Tech, Inc. He also possesses the Certified in Cybersecurity (CC) cert from the ISC2.

Previous articleCounterfeit Euro Banknotes in the COVID Era
Next articleHistory of the U.S. Passport: Part 1