Fingerprint biometrics – dispelling the misconceptions

May 22nd, 2017, Published in Articles: EE Publishers, Articles: EngineerIT

 

There is today no doubt that going the biometric route is the only way to verify that the person who presents the information is the person he or she says he or she is, irrespective of which system is deployed. Be it finger print, iris, voice or any other biometric system, it is the most secure method to prove the liveness of the information at the time it is being presented. Smart cards, tokens, passwords or pin numbers can be faked or shared.

Currently the most deployed system is fingerprint imaging, yet there are many misconceptions about its robustness. It is by far the oldest technology, and was first introduced to solve crimes. As it developed from a paper and ink based system to a digital system it became a very useful technology to identify a person – but the initial system had its limitations. Placing a finger on a sensor came with all sort of problems. The finger surface may be contaminated with dirt, hand cream and all sorts of other contaminates that would obscure the fingerprint, resulting in an incorrect reading and failure to grant access. With conventional sensors pressure was another problem, pressing too light it would not read and pressing the finger down too hard would flatten the ridges and valleys of the print and result in a reading failure. These factors may have led to the misconception that finger print biometric technology is not robust enough to work every time.

According to Greg Sarrail, VP at HID Global, the technology has come a long way since the introduction of the first electronic fingerprint system. HID is a leader in access control and secure identity solutions, and recently released a white paper on multispectral fingerprint image acquisition.

Fig. 1: A MSI finger print sensor installed at a copper mine. It has been flawlessly operating for over three years despite the dust and dirt.

There are a number of different techniques for capturing a fingerprint image including optical, capacitive, radio frequency, ultrasound, and thermal methods. One common shortcoming of many conventional fingerprint sensing technologies is the frequent occurrence of poor-quality images under a variety of common operational circumstances. Though each particular imaging method has different sensitivities, in general poor images may result from conditions such as dry skin, worn surface features of the finger, poor contact between the finger and sensor, bright ambient light, and moisture on the sensor. Many imaging technologies are also unable to provide strong affirmation that the fingerprint image is collected from a living, unadulterated finger rather than an artificial or spoof sample. This is so because the raw data collected by these systems contain little or no information about the physical properties of the fingerprint ridges presented. For example, a conventional optical fingerprint reader based on total
internal reflectance (TIR) acquires images that represent the points of optical contact between the sensor platen and any material with a minimum index of refraction. Because materials have an appropriate refractive index and can be formed to contain a fingerprint pattern, such a system is susceptible to spoof attempts.

According to Sarrail, to address these shortcomings, an optical fingerprint sensor has been developed that is able to work across the range of common operational conditions while also providing strong spoof detection. The sensor is based on multispectral imaging (MSI) and is configured to image both the surface and subsurface characteristics of the finger under a variety of optical conditions. The combination of surface and subsurface imaging ensures that usable biometric data can be taken across a wide range of environmental and physiological conditions. Bright ambient lighting, wetness, poor contact between the finger and sensor, dry skin, and various topical contaminants present little impediment to collecting usable MSI data.

A customised algorithm is used to fuse multiple raw MSI images into a single high-quality composite fingerprint image. This single fingerprint image can be used to match other MSI fingerprint images as well as images collected using other methods. Thus, the MSI fingerprint is backward compatible and can be used with existing fingerprint databases collected with different imaging technologies, explained Sarrail. The surface and subsurface data collected by the MSI sensor provide rich information about the optical properties of the bulk sample. A classification methodology has been developed to operate on the MSI data and determine if the measured optical properties of the sample are consistent with those of living human skin. If so, the sample is deemed to be genuine; otherwise, the sample is identified as a possible spoof attempt. Sarrail says that this provides the means by which an MSI sensor can provide strong assurance of sample authenticity.

The banking industry have adopted MSI imaging in many of its interfaces with customers to provide reliable identification. The latest application of fingerprint technology is now being piloted at ATMs, in addition to the current pin number.

According to Sarrail, this is a real challenge for banks. The finger imaging technology stands between the customer and the dispensing of cash. It has to work every time irrespective of how clean the finger is. MSI is the only technology that will provide that assurance.

Send your comments to engineerit@ee.co.za

Related Articles

  • Drone registration fees to be amended
  • Advancing GIS in tertiary education in a time of budget cuts
  • Make your data actionable from edge to cloud
  • Light quality metrics explained
  • Transforming condition monitoring of rotating machines