An Automated Fingerprint Identification System (AFIS) is effectively a storage, search and retrieval system for finger and palm print electronic images and demographic data. AFIS is a high speed, high capacity image processing system that enhances the ability of the latent fingerprint examiners to search and identify crime scene evidence against ever increasing pools of fingerprint records.
AFIS systems have and continue to replace outdated manual methods of fingerprint classification employed by law enforcement agencies over the past century.
AFIS systems utilise specialised software and powerful computer hardware configurations to create unique mathematical ‘maps’ (algorithms) based upon relationships between the characteristics present within the finger or palm friction ridge skin structures. Modern AFIS systems rapidly extract information from the fingerprint to establish the pattern type, minutiae points and the axis of the image. The use of mathematical algorithms enables a fingerprint to be compared with millions of file prints within a matter of seconds.
The latest AFIS systems may also incorporate palm print matching capabilities. In the majority of operating systems palm print images are divided up into a number of small segments so that the software can effectively and efficiently code, store and search the palm data within a reasonable time frame (similar size to a rolled fingerprint impression).
AFIS software utilise the impressions obtained from the rolled index fingers or thumbs to search and ‘match’ against existing tenprint records within the database. This matching process is used in proving identity in the recording of criminal convictions (criminal history).
The AFIS system may use the rolled impressions or a combination of the rolled and flat impressions of all fingers to compare against the unsolved crime latent database. This process is used when comparing new arrest finger and palm images against the unsolved crime latent database or when comparing new crime scene latent evidence against the existing arrest records within the tenprint database.
Different systems offer “binary” or “grey-scale” images. Pure black and white images depicted as binary images frequently allow important data to “drop out” of the image, i.e. the scanner or reader recognised black and white but not the shades of grey. Modern systems offer grey scale images and, depending on the system, these images may be captured using up to 256 grey levels which gives the examiner more detailed data to compare and identify.
Latent images can be scanned from physical ‘lifts’, negatives, photographs or uploaded from digital cameras, all in high quality detail (up to 1000Dpi). Tenprint images can also be captured using Livescan technology (see CrimTrac page).
When searching an unknown print against the AFIS database, the system provides a “candidate list” of the closest matching fingerprint images from the tenprint database. The fingerprint examiner verifies the results and indicates whether an identification has been made against any one of the nominated candidates.
While the list provided by the AFIS is given in order of decreasing match value (as calculated by the search algorithm), the final identification, as established by the fingerprint expert, may not necessarily be among the first few candidates on the list. Despite the progress made in computer hardware and software, AFIS technology has not yet eliminated the need for human verification of AFIS match results.
The AFIS database contains several separate databases, each with its own specific purpose and storage scheme, to facilitate efficient overall system performance. These include the tenprint, palm print and unsolved crime case image databases (finger and palm).
The current Australian National Automated Fingerprint Identification System was commissioned in April 2001. The new AFIS utilises the SAGEM “Morpho” operating system incorporating the use of 256 grey scale and 500 Dpi images.
The Australian AFIS database currently contains 2,600,000 tenprint records that are uploaded by the individual State and Territory Police Agencies. The database contains tenprint records from all arrested persons (since 1941) and other tenprint records including police applicants and other occupations as required by State and Territory legislation (probity checks).
The new AFIS can be accessed from any State or Territory in Australia from 39 metropolitan and remote locations. Each site can upload both tenprint and latent data and review AFIS search results at a local level. The new NAFIS supports Livescan technology and will eventually integrate over 150 Livescan devices directly with the AFIS search system.
All image comparisons (tenprint and latent) can now be completed on screen, using high quality images from the database. Latent examiners are able to enhance crime scene latent evidence using a variety of digital ‘tools’ available within operating system software. These tools include fast Fourier transformation and 3 dimensional imaging that can be applied to the image prior to searching against the tenprint record database.
The SAGEM AFIS system replaced its predecessor the NEC AFIS system that had operated as the National AFIS in Australia since the early 1980’s. The NEC operating system was based upon the use of binary images. It was one of the first automated fingerprint systems in the world utilised on a national level by law enforcement agencies