Automated Fingerprint Identification System commonly known as AFIS, is a system that stores and processes fingerprints using digital mode. AFIS is the advanced technology for fingerprint identification systems and a readymade solution for automatic data processing in a rapid manner. It provides secure and controlled access and doesn’t expose data to cybercriminals compared to the typing finger system which can easily be locked or unlocked quickly. AFIS is the computerized storage system to store millions of fingerprints. In earlier times before the computerization of the system, this job of comparison would require weeks and days for completion while the present-day demands the utilization of computer system making it a job of few minutes.
The AFIS system is a software-based computer system that has a database of known ten prints and palm-prints that may be added to at any time. Images of individual fingerprints or palm prints, maybe from a crime scene, can be entered into the system, which is then mapped by the system's algorithms. These maps are then automatically compared to the database, resulting in a list of possible matches that the fingerprint analyst can analyze to see if they match or not, i.e., do the comparison of whether the unknown prints belong to someone already existing in the database. So, Automated Fingerprint Identification System (AFIS) is considered as a system for storing, searching, and retrieving computerized photographs of finger and palm prints as well as the demographic data. It is a high-speed, high-capacity image processing technology that improves the ability of fingerprint examiners' to search for latent prints and identify crime scene evidence, and the ability to scan arrested prints against an ever-growing pool of fingerprint records.
How Does AFIS Operate?
AFIS mainly proceeds through the effective application of highly sophisticated algorithms. Many such algorithms have been regularly developed and improved based on real-world experience throughout the years. The three main steps of AFIS operation includes:-
- Image Enhancement
- Feature Extraction
- Feature Matching
1. Image Enhancement -
Image enhancement is done in several ways mostly by the use of Gabor filters to increase the brightness of the ridges and valley structure of the fingerprints images and for enhancing the image quality. There are some common examples for image enhancement software like Optimisation Image Enhancement Algorithms, as their name implies, deal with a variety of issues that affect the basic quality of latent print pictures.
2. Feature Extraction -
In regard to the feature extraction method, the most studied is to extract features from a thinned image using the crossing number concept. The minutiae points (typically ridge ends and ridge bifurcations) that distinguish one print from another are identified using feature extraction methods. Algorithms that can locate non-minutiae points, such as pores or textures, could help with this as well.
3. Feature Matching -
There are different algorithms used for feature matching systems. The probable match can be established by studying the features combining minutiae and non-minutiae techniques to find a match.
AFIS mainly uses two areas, one is fingerprint verification and the second is fingerprint identification. The fingerprint verification is done by utilizing high-security sectors which give only limited access to entry; the staff or certain people who try to enter or want access to the data, use of fingerprint scanner can solve the situation. The second area i.e. fingerprint identification means comparing the fingerprints with a stored one. An Automated Fingerprint Identification System uses biometrics to store the digital image of individual fingerprints for database comparisons.
As digital technology progresses fingerprints are increasingly being used as fraud prevention measures. AFIS is able to search the database for a complete or partial fingerprint and tell the probable matching candidates. Matches usually contains a score expressing the likelihood of being a correct match in the AFIS database. The accuracy of the search can be increased when more fingerprints from the same person are available. When analyzing a crime scene, it is of paramount importance to be able to separate fingerprints of the usual occupants and those of possible suspects. For this, AFIS is indispensable due to its quick response and matching time. It can group fingerprints from the same individuals, reducing search times and the complexity of the necessary searches. In simple cases, it can take only a few hours from finding a fingerprint to identify and apprehend a suspect.
In India, a database called as National Automated Fingerprint Identification System (NAFIS) has been prepared for keeping records of the finger impressions of a large number of criminals. The database collects data of around 80 lakh criminals from all over the country. National Crime Records Bureau (NCRB) will be working on a plan to allocate separate space for NAFIS to each state of India. Provisions will be made for each state which already has AFIS to share their data to NAFIS through bridge software. NAFIS will help AFIS by enhancing the ability to distinguish crimes depending on the crime pattern and the modus operandi of the criminal across the state. This is considered as one of the most secured applications used by law enforcement officers for the identification of individuals based on their fingerprints. The application software works on the pattern matching algorithm that involves the comparison of two different fingerprint images, in order to determine whether they represent the fingerprint impression of the same person or not.
AFIS has now been used tremendously in various domains mostly in ATM machines, police investigations, and many more. Its uniqueness, popularity, and permanence make it popular to use in a variety of sectors. AFIS is also being tried to propose a new algorithm using MATLAB, some modifications to the Gabor filter algorithm to do the minutiae-based matching by modifying the technique by employing two points as fingerprint features: minutiae and ridge points and using a weight table to assign each feature a similarity value based on its quality and increasing the accuracy of the system to reduce the error rate.