IGP, SCRB, Madhya Pradesh
Automated Multi-Modal Biometrics Identification System (AMBIS) incorporates state-of-the-art biometric technologies to serve law enforcement applications beyond traditional Automated Fingerprint Identification System (AFIS) capabilities
National Crime Records Bureau (NCRB) is the central repository of total fingerprint biometrics being used for tracking criminals across the country. At present the Automated Fingerprint Identification System (AFIS) exists at the NCRB headquarter and 22 other states headquarters. Eleven more states are yet to install the system. These AFIS have been running standalone with least features to deliver the desired result for tracking criminals in states.
Shortcomings of AFIS
Most of AFIS are of outdated technology and have proprietary encoding and matching algorithms, which lack commonality and interoperability. Further, none of these AFIS has inter-state/inter AFIS connectivity module and functionality and therefore no data portability and interoperability is achieved even amongst various versions of same vendor and AFIS of other vendors. All AFIS have miserably poor capability to search latent print. Moreover, no AFIS has the capability to store and search palm print and is not complete package of all required core functionalities.
Because of the aforementioned reasons, these AFIS have virtually failed to track criminals and have lost their credibility and usefulness. Despite the huge database availability in the country, a fraction of it has been digitised and much lesser has reached the NCRB for tracking criminals.
Automated fingerprint identification systems (AFIS) have been widely used in forensics for the past two decades, and recently they have become relevant in civil applications, as well.
Keeping the above facts and anomalies in mind, a National Benchmarking Committee has been formed so that NCRB comes with a state-of-art system similar to the one, which the FBI has. It also removes all anomalies of the present system so that tracking of criminal becomes seamless as is imminently required for the success of CCTNS.
Automated fingerprint identification systems (AFIS) have been widely used in forensics for the past two decades, and recently they have become relevant in civil applications as well. Whereas, large-scale biometric applications require high identification speed and reliability and multi-biometric systems that incorporate fingerprint, iris and face.
Automated Multi-Modal Biometrics Identification System (AMBIS) incorporates state-of-the-art biometric technologies to serve law enforcement applications beyond traditional AFIS capabilities. The various modalities used today include finger (ten print flats and rolls, latent), face (mug shot and latent face) multi modal biometrics technology for CCTNS, iris (dual iris scans) and palm (print and latent). These offer a number of advantages for improving identification quality and usability.
Proposed national AFIS
AFIS is a system in which images of known fingerprints are encoded and stored in a computer database. This database of known fingerprints and other images of ten digit fingerprints is then utilised to search unidentified latent fingerprints through the system to determine identity. The system encodes the fingerprints that are being searched and finds fingerprints in the system that most closely resemble the fingerprint being searched. A qualified examiner compares the fingerprints reported by the AFIS and determines if identity of the searched fingerprint (inked or latent) can be established. It is observed that most of the latent prints found at the scene of crime are partial palm prints. Therefore an AFIS having palm print search and storage facilities are also required.
The main objective of National Automated Fingerprint Identification System (NAFIS) is to provide the national level fingerprint database of criminals and improve crime detection rate with the help of fingerprint identification. It is proposed to have a National AFIS system at NCRB, which will store finger print data of all states. All states should have their state AFIS. The state can deploy remote stations at district, sub-divisional or police station level as required. All these AFIS systems will be interconnected with automatic remote updating and query facility. AFIS with web-enabled updating and query processing facility will be appreciated.
The NAFIS will maintain the fingerprint data in standard ANSI/NIST format. All states AFIS will be connected to NAFIS with strong networking facility. NAFIS will follow the server-client architecture and also support web-based scenario.
The NAFIS will provide automated fingerprint search capabilities, latent searching capability, electronic image storage, and electronic exchange of fingerprints and responses 24 hours a day, 365 days a year.
As a result of submitting fingerprints for search, electronic responses to criminal ten-print fingerprint submissions would be received within hours.
Database Specification for AMBIS
During All India Directors Conference of Finger Prints Bureaux held at Bhopal on 6-7 January, 2011 all the finger print experts from different states proposed to have National AFIS at NCRB, New Delhi. AFIS data centers should be hosted on state data center, so that there is a seamless integration of state AFISs to national AFIS.
As in times to come, the average size of FP database will be nearly 10 lakhs or above for large states like Madhya Pradesh, Andhra Pradesh, Tamil Nadu, Uttar Pradesh, etc. In a similar manner, it is 5 lakhs for small states like Kerala, Chhattisgarh, Jharkhand, etc.
So the proposed Database of AMBIS at NCRB should be: –
10 Digit Database: 1crore with upgradibility up to 1.5 crore
Chance Prints Database: 5 lakh with upgradibility up to 10 lakh
Palm Print Database: 5 lakh with upgradibility up to 10 lakh
Iris Database: 10 lacs with upgradibility up to 20 lakh
Face Database: 10 lacs with upgradibility up to 20 lakh
Here we are considering these huge sizes of various databases in view of CCTNS project in which all districts units as well as police stations will be connected through a dedicated network. And as we know, fingerprint is an integral part of CCTNS project.
The system must perform reliable identification with large databases, as biometric identification systems tend to accumulate False Acceptance Rate (FAR) with database size increase and using a single fingerprint, face or iris image for identification becomes unreliable for a large-scale application. Several fingerprint images from a person’s different fingers or iris images from person’s two eyes may be taken to increase matching reliability. Also, multi-biometric technologies (i.e. collecting fingerprint, face and/or iris samples from the same person) can be employed for greater reliability.
Multi-biometric systems can solve a number of problems of uni-modal approaches. One source for such problems can be found in the lack of dynamic update of parameters, which does not allow current systems to adapt to changes in the working settings
The system must show high productivity and efficiency, which correspond to its scale. System scalability is important, as the system might be extended in the future, so a high productivity level should be kept by adding new units to the existing system. The daily number of identification requests could be very high. Identification requests should be processed in a very short time (ideally in real time), thus high computational power is required. Support for large databases (tens or hundreds of millions of records) is also required. Another important criteria is general system robustness. The system must be tolerant to hardware failures, as even temporary pauses in its work may cause big problems taking into account the application size.
The system must support major biometric standards. This should allow the use of system-generated templates or databases with systems from other vendors and vice-versa. The system may need to match flat (plain) fingerprints with rolled fingerprints, as our department collect rolled fingerprint databases. The system must be able to work in the network, as in most cases client workstations are remote from the server with the central database. Further, a forensic system must be able to edit latent fingerprint templates in order to submit latent fingerprints into the AFIS for the identification.
Multi-biometric systems can solve a number of problems of uni-modal approaches. One source for such problems can be found in the lack of dynamic update of parameters, which does not allow current systems to adapt to changes in the working settings. They are generally calibrated once and for all, so that they are tuned and optimised with respect to standard conditions. In this work, it is proposed that an architecture where, for each single-biometry subsystem, parameters are dynamically optimised according to the behaviour of all others. This is achieved by an additional component, the supervisor module, which analyses the responses from all subsystems and modifies the degree of reliability required from each of them to accept the respective responses.