Papillon Systems

PAPILLON: UP-TO-DATE BIOMETRIC SOLUTIONS
from Russia

  
INFORMATION PRESS-OFFICE PRODUCTS PERFORMANCE SITE MAP

Q     m     [x]  

    Sign up   
 РУС | ENG | ESP 
Scanner Selection Guidance
 
Scanner Selection Criteria
Functional Capabilities and Scanner Platen Size
Compliance with Image Quality Requirement
Nail-to-Nail Rolling and Compensation for Smear
Scanning of “Complicated” Exemplars and Moisture Suppression
User-Friendly Interface and Ergonomical Characteristics

 

One of the important conditions for increasing the AFIS hit ratio is the use of the most recent imaging technologies to oust the traditional ink method for obtaining finger and palm prints.

High technologies of image acquisition, based on livescan devices, enable the capture of most crisp and sharpest fingerprint images, thereby producing forensic-quality tenprint collections.

High quality of prints ensures higher accuracy of automatic recognition and coding of images in any AFIS, and as a consequence, the increase of its major search characteristics -reliability, selectivity and operating speed.

Besides the profound improvement of data quality, finger and palm scanners solve another range of tasks vitally important for the systems of automatic identification, namely:

  • High efficiency through real-time image capture and entry to the AFIS
  • Remote access to the AFIS for data transmission and database insertion
  • Minimization of human factor

Wide application of AFISs for the growing needs of law enforcement agencies and identification departments as well as increased interest in biometric identification for various civil applications are currently being considered by scanner vendors worldwide making them design and offer new models.


Today, a great number of various types of fingerprint and palmprint scanners are being marketed - optical, capacitor, thermal, electromagnetic, etc. Identification systems for criminal application (AFISs) mostly employ optical one. The optical scanners first appeared more than 50 years ago.


The pictures below show a general view and a cross-sectional drawing of the first optical device patented in the Soviet Union (1956) called “a papillascope”, designed for fingerprint examination..

Папилляроскоп Папилляроскоп


A bedrock principle used in all optical scanners is an effect of frustrated total internal reflection of light at the border of transparent environments, i.e. the beams, illuminating the scanner platen by falling on it from the inside, are completely reflected only from furrows on the finger skin but not from the places where friction ridges are.

In practice, advanced optical scanners generate images with the best showings of resolution and dynamic range as compared them with other types of livescan devices.

Criteria to Use When Selecting a Scanner for AFIS:

  1. Functional capabilities and scan area size
  2. Conformance of captured image data to specifications and requirements stated in performance standards or other regulatory documents
  3. Ability to compensate for smeared images (relevant to the scanners used for rolled fingerprint capture)
  4. Ability to generate forensic-quality images of complicated samples (too dry fingers, with worn-out or low-ridge patterns)
  5. Ability to ignore moisture
  6. User-friendly graphical interface and other ergonomic characteristics (such as customizability for instance


Relevance Analysis of Fingerprint Scanner Data (according to PAPILLON’s estimates)

       


Functional Capabilities and Scan Area Size

The choice of a functional type depends upon the tasks, which have to be solved with the scanner:
 

–  Input of tenprints into AFIS databases

Since searches in the AFIS involve not only tenprint files but also deal with latents lifted at crime scenes, as well as with pattern fragments (unknown dead, victims of accidents, disasters or hostilities), a live-scan device you choose is to be capable of capturing full - from nail to nail - rolled fingerprints.

If it is contemplated also to take plain impressions, the scanner must have an ample sensing area for scanning 4 flat fingers simultaneously.

If your AFIS operates with palmprints, including latent palmprints as well, you need a live-scan device of a palmprint type, which has a sensing area sufficient for taking palmprint images.
 

Identity checks against AFIS database by using touch prints
(as implemented in the PAPILLON AFIS)

To support this mode, it suffices to use a single finger livescan device with a sensing area dedicated to capturing a plain impression of a single finger taken without rolling.

The size of the scanner platen must provide an acquisition of complete, not truncated, fingerprint or palmprint images. Failing to meet the above-stated requirement might cause the loss of a part of ridge details, thereby decreasing the AFIS efficiency.

It should be noted also, the smaller than needed scan area makes it difficult to position correctly a finger or palm on the scanner platen.

According to up-to-date specifications, the scan area dimensions required as per types of impressions are as follows (W x H):

Rolled fingerprints Plain impressions of 4 fingers Half palm impressions

PIV single finger flat impression
(FAP levels)

minimum 1.6 x 1.5" (41 х 38 mm) minimum 3.2 x 3.0" (81 х 76 mm) minimum 5.0 x 5.0" (127 х 127 mm) minimum 0.8 x 0.8" (20 х 20 mm)

 


Compliance with Image Quality Requirements

For testing flatbed scanners and live-scan fingerprint devices to meet the USA FBI’s requirements, PAPILLON is guided by The Test Procedures for Verifying IAFIS Scanner Image Quality Requirements, CJIS­-TD-­0010 v.7.

The requirements set forth in the said document have been duplicated in some international and Russian national standards adopted in the sphere of biometric identification. Particularly, the Automatic Identification. Biometric Identification. Data Format for the Interchange of Biometric Information. Part 4. Fingerprint Image Data GOST R ISO/MEC 19794-4-2006 standard definitely requires all scanners used for law enforcement to be FBI-certified.

According to the above-mentioned standards, certified scanners must meet image quality requirements in each of the following areas:

  • Minimum resolution of 500±5 pixels per inch (in both the row and column directions)
  • Geometric image accuracy
  • Modulation transfer function
  • Signal-to-noise ratio
  • Gray-scale range of image data
  • Gray-scale linearity
  • Output gray level uniformity

No doubt, certification is essential, as it is a reliable proof of scanner designers’ proficiency. However, the FBI certificate granted to a fingerprint scanner does not ensure its high performance when it is integrated into an AFIS. This fact can be easily explained: fingerprint scanners have been imposed the same requirements as were flatbed scanners used for distortion-free conversion of paper tenprints, immovable plain originals, to electronic format.

Being a direct reader of ridge details, a live-scan fingerprint device features some other characteristics that must be also tested. These characteristics have not been measured or somehow worded yet in normative documents, but still they are critical to enrollment since they directly affect the quality of fingerprint files and AFIS performance. Among these characteristics are the following:

  • Ability to compensate for smeared images of rolled fingerprints
  • Ability to generate forensic quality images of “complicated” samples (dry or worn-out fingers, low-ridge patterns, etc.)
  • Ability to capture high-quality images that are not obscured by moisture (excessive sweating)

 
Nail-to-Nail Rolling and Compensation for Smears

How does the size of captured images affect the AFIS reliability?

PAPILLON conducted statistical investigations having explored the operation of an AFIS database in one of the biggest cities of Russia. The database contained 2 million tenprint and 130 thousand latent records. The following observations have been done:


AFIS Operation Mode Result

1. Standard, when searches involved registered rolled fingerprints

11,000 Latent-to-Tenprint (LT-TP) hits

2. Full roled fingerprints were replaced with slaps (constituting 55% of the rolled finger area).

35% reduction in the number of LT-TP hits

3. Full rolled fingerprints were replaced with square fragments of 1 inch on side (as adopted in some foreign AFISs). The area of truncated images is 9% less than of source images.

7% reduction in the number of LT-TP hits

4. Changing some image parameters that are tested at benchmarking of live fingerprint scanners (as mentioned above):

– Decreasing the resolution of fingerprint images from 500 to 400 pixels per inch

1% reduction in the number of LT-TP hits

– Decreasing the resolution of fingerprint images from 500 to 300 pixels per inch

3% reduction in the number of LT-TP hits

– Reducing the gray-scale range from 128 to 16 gray levels

2% reduction in the number of LT-TP hits

– Artificial geometric distortions of fingerprint images at the rate of ±5%

0.5% reduction in the number of LT-TP hits

 
These statistics affirm our belief that:

  • Among the data critical to AFIS performance (in addition to FBI-certified parameters) is the size of an image area captured at enrollment.
  • A 10% increase in image area versus slaps area earns an 8% increase in AFIS search reliability.

In summary, the above-stated findings highlight the value of full fingertip scanning – from nail to nail – for electronic submission to the AFIS. Otherwise, many latents may go unsolved due to the inability to match them with incomplete fileprints.

The latents below were identified against the AFIS databases of rolls, though unsolved when comparing with inch-per-inch images.


Latent Print Image
 
 


Rolled Fingerprint 
 
 


Only 4 minutiae from 11 coded on the latent image appear within the selected inch-per-inch area.
 


Latent Print Image  



Rolled Fingerprint   


Only 5 minutiae from 18 coded on the latent image appear within the selected inch-per-inch area.

 
 
Smeared images and methods of compensation

A rolled fingerprint is a linear scan of the ridge pattern of a finger reproduced on a paper record card or captured on a fingerprint scanner platen.

Nail (distal) phalanxes are not shaped to ideal cylinders, and that is why a finger inevitably slips over the scanner platen (or tenprint card) when being rotated from side to side, thereby causing appearance of smears obscuring ridge details. The larger the rolled print, the more likely the image will be smeared.

Depending also upon experience and skill of fingerprinting operators, the same imperfection may be also caused by anatomical peculiarities of the fingers.

Fingerprints taken in the traditional ink method contain areas which are of little use or completely useless for recognition due to illegibility of smeared ridge details. It should be noted again that the ink technology of fingerprinting is virtually unable to eliminate this sort of defects.

Смаз прокатанного изображения на бумажной дактилокарте
Ink fingerprint image obscured by smears  


Unlike traditional ink fingerprinting, live-scan fingerprinting allows the developers to devise and to realize ad hoc program methods to produce high quality images that are not affected by smears. By eliminating smears, this imaging technology delivers a 5-10% increase in usable area of a rolled fingerprint.

Among the hardware-based methods of compensation for smears are:

  • increasing a frame refresh rate when capturing images
  • transporting a CCD element

The disadvantage of the first from the above-mentioned hardware-based methods consists in that it requires the use of high-speed cameras. As the high-speed and high-resolution cameras are expensive, using them in the devices races up the cost of scanner manufacture, which means that the issue is solved per se at the consumers’ expense. Every effort to use cameras with lower resolution (and cheaper, respectively) provokes downsizing of the scan area and lessens the information value of captured images.

The drawback of the second hardware-based method is the presence of some moving mechanical parts and automation facilities within the scanner, which sophisticates the device, lowers its reliability and requires regular maintenance and replacement of worn-out parts (such as e.g. winding and moving parts of the step motor, carriage drive, end switches, etc.).

The software-based method of compensating for smears appearing on a rolled fingerprint image provides virtual splicing of sequentially captured pictures. Such approach to the problem is devoid of any disadvantages that feature the hardware-based compensation methods, though imposes high demands to the developers who create the scanner software.

So, creating a scanner, which would be capable of eliminating the problem of smeared pictures and generating high-resolution images, and for all that having a price acceptable to consumers, is a contradictory and intricate task.

Naturally, only a few developers have been able to find a proper solution to the said problem. Respectively, not all of the marketed live-scan fingerprint devices cope with task of compensating for smears.
 

 
How to test the scanner’s capability of compensating for smeared images at fingerprinting?

So far, no procedures have been elaborated yet for testing fingerprint scanners in compensating for smeared images captured at rolling. Consumers can be only recommended to evaluate visually the quality of images generated at rolling fingers over the scanner platen.

To do this, it suffices to roll thumbs (usually producing a maximum smearing) of several individuals, and then, having printed the captured images out, to examine them closely.


There are obvious shifts and discontinuities on the image.


Shifts and discontinuities are blurred over


All discontinuities are qualitatively compensated  

 

Scanning of “Complicated” Exemplars and Moisture Suppression

Too dry or vice-versa sweaty skin, worn-out or flat-relief ridges are factors distorting a true picture of fingerprint images.

Dry and worn-out human fingers are usually imaged fuzzy or faint, ridge lines being illegible, and a number of fine details and critical areas are “lost” through poor contact between the finger and the scanner platen.

Vice versa, images of excessively sweaty fingers are obtained too dark, with bright black spots, because of moisture or sweat filling the furrows between friction ridges.

A forensic-quality fingerprint scanner is required to lower a negative impact of the above-mentioned issues and to consistently deliver high quality prints by maintaining every friction ridge detail irrespective of skin condition.

The issue of skin dryness can be considerably tackled, for example, by using an elastic coating to cover the contact surface of the platen. Some manufacturers have discovered effective solutions how to ignore moisture on fingers and on the platen.

To test how the device operates with dry fingers, the skin should be dried out with some alcohol or chalk. To imitate sweaty skin, just dip the fingers into water or damp them with a wetted wipe.

By analyzing and comparing the obtained images, you will be able to evaluate the efficiency of scanners when dealing with “complicated” samples.



Impression of a dry finger


  


Impression of a sweaty finger

   

User-Friendly Interface and Ergonomical Characteristics

Among the quality performance criteria of a system for electronic fingerprinting are some very important functions enabling automation of the scanning process and minimizing the human factor that affects the quality of captured images, as well as operability and usability of the system overall. Described below are most critical of the mentioned functions:

  • Automatic slap-to-roll comparison
    The system checks that each rolled fingerprint has its correct location and belongs to the same person by correlating them with the plain impressions, preventing thereby operators from typical mistakes, when the right hand is scanned instead of the left one and vice versa, and a wrong finger in place of the specified.
  • Automatic image quality control
    Image size and quality are evaluated automatically. In case of unacceptable results, a message appears noting the problem. The ability to scan as many times as needed to achieve better results is provided.
  • Autosave
    The image just captured is automatically inserted into associated electronic tenprint card when proceeding to next finger.
  • Embedded database maintenance
    By maintaining an embedded (internal) database of electronic tenprints, the scanner delivers an ability to search for an individual’s data by his touch fingerprint or by alphanumeric descriptive information specified as search criteria.
  • Autonomous guidance through the scanning procedure
    The scanning process is guided without resort to the workstation monitor and keyboard, but through the display and keypad embedded in the scanner.
  • Standard hardware interface
    The scanner is interfaced with the workstation computer through standard USB and IEEE1394.

At the same time, it is impracticable to comprise a comprehensive list of all functionalities an ideal system of electronic fingerprinting should support. Well-engineered software generally imparts flexibility to the system allowing it to be customized and tailored as customers need.

Ergonomical requirements imposed on fingerprint scanners are standard and need no special explanations. These include characteristics of usability, easy mounting, setup and servicing, etc. Another essential requirement concerns adequacy of dimensions and modeling of the scanner to service and environmental conditions (such as, for instance, ruggedness to withstand extreme working conditions in law-enforcement environment).

To top

 

PAPILLON Fingerprint/Palmprint Livescan Devices

 

  Copyright 2001-2018 Papillon Systems. All rights reserved.