2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

Use of AADHAAR Biometric Database for Crime

Investigation - Opportunity and Challenges
Dr. Manish Kumar1 Dr. M. Hanumanthappa2
Asst. Professor, Dept. of Computer Applications, Professor, Dept. of Computer Science and Applications,
M. S. Ramaiah Institute of Technology, Jnana Bharathi Campus, Bangalore University,
Bangalore-560 054, INDIA Bangalore -560 056, INDIA
E-Mail:- [email protected] E-mail:- [email protected]

Dr. T. V. Suresh Kumar3

Professor & Head, Dept. of Computer Applications,
M. S. Ramaiah Institute of Technology, Bangalore-560 054, INDIA
E-Mail:[email protected]

Abstract—Now biometric technique is a front line security

measure for both identity verification and crime detection. It’s a II.NATIONAL IDENTITY PROJECTS IN OTHER JURISDICTIONS
century old techniques and become an integral part of almost all Today all most all the countries are providing some kind of
crime investigation. In recent day’s lots of advanced techniques,
Nation Identity to their citizens. Some country has very
Software and Hardware system has been developed to improve
the accuracy and performance of the Biometric Identification
advanced state-of-the art technologies where as some of them
System for Crime Investigation. In this paper we are discussing are still developing it. This section will discuss in short about
how Aadhaar Biometric Database can be used for tracing different kind of National Identity Projects implemented in few
criminal and crime investigation. Aadhaar is a Unique countries.
Identification Numbers (UID) provided by The Unique
Identification Authority of India to Indian Citizens. A. United States
Social Security number (SSN) is a nine-digit number issued
Index Terms—Aadhaar, Unique Identification Authority of to U.S. citizens, permanent residents, and temporary (working)
India (UIDAI), Biometric, Iris, Fingerprint. residents.
The Social Security number (SSN) was created in 1936 for
I.INTRODUCTION the sole purpose of tracking the earnings histories
Fingerprints analysis is one of the most vital and integral of U.S. workers, for use in determining Social Security benefit
process in crime investigation. Fingerprint databases are entitlement and computing benefit levels. Since then, use of
commonly used worldwide by almost all the police department the SSN has expanded substantially. Today the SSN may be
and law enforcement agencies for the purposes of forensics, the most commonly used numbering system in the United
crime investigation and identifying the criminal’s involvements States.
in various crimes. B. United Kingdom
In India, it’s a challenging task for investigators to fish out
The Identity Cards Act was introduced on 11th September,
fingerprints of criminals which often get lost or lay hidden in
2001, amidst rising concerns about identity theft and the
piles of files in police stations because no attempt has been
misuse of public services. The card was to be used to obtain
made to put them in one database.
social security services, but the ability to properly identify a
For the country like India, having world second largest
person to their true identity was central to the proposal, with
population, it’s really a very difficult to architect a separate
wider implications for prevention of crime and terrorism. The
biometric or identity database for different-different purposes
cards were linked to a central database (the National Identity
individually. As 90% of Indian population has enrolled in
Register), which would store information about all of the
Aadhaar, It become a standard biometric database which can
holders of the cards. The Act specified 50 different categories
be used for various purposes including the crime investigation.
of information that the NIR could hold, It includes, digitised
In this paper, we will examine how Aadhaar biometric
facial scan, iris scan and up to 10 fingerprints scan.
database can be used for crime investigation and how it can be
integrated with the existing system in India, the challenges of C. Estonia
integration, and possible solutions. Paper also highlights some Estonia has one of the most highly developed national ID
of the technical and legal challenges for its implementation, card system. The national ID card system in Estonia gives their
which is a policy matter. citizens secure digital access to various services. The card uses

‹,(((
 2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

2048-bit public key encryption and legally equivalent to a III.USE OF BIOMETRIC DATABASE FOR CRIME INVESTIGATION
manual signature. Today a massive computing power and well written pattern
D. France matching algorithms have made it possible to automatically
search and detect the patterns of similarity in many different
The biometric ID card was to include a compulsory chip
kind of digital database of patterns which may include
containing personal information, such as fingerprints, a
fingerprints, faces, and irises. Effectiveness and accuracy of
photograph, home address, eye colour etc. A second, optional
such system are very promising even for a very large
chip was to be implemented for online authentication and
databases. These kind of technologies are continuously
electronic signatures, to be used for e-government services and
evolving and offering a potential edge to the law-enforcement
e-commerce. The law was passed with the purpose of
agencies. Many countries has developed a separate biometric
combating “identity fraud”.
database for this purpose, where as many of them have linked it
E. Argentina with their National Identity database for effective analysis
Documento Nacional de Identidad or DNI (which means [2][13].
National Identity Document) is the main identity document for The Federal Bureau of Investigation, United States
Argentine citizens, as well as temporary or permanent resident. maintains the largest database of faces. The FBI has 548
The basic objective of DNIe is to digitally certify the million pictures catalog of searchable “face photos” which
identity of the person, allow the electronic signature of includes criminal mug shots, photos of terrorists taken by
documents and gives the possibility to use the electronic soldiers in Afghanistan and, more recently, state driver’s
identity in digital services. license and ID photos of ordinary citizens who have committed
The card is issued to the citizens at birth and updated at the no crime at all.
age of 8 and 14. The DNI is a valid travel document for Nearly a decade ago, feared by another 9/11- style attack,
entering Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, FBI signed a $1 billion contract with Lockheed Martin to
Paraguay, Peru, Uruguay, and Venezuela. develop a Next Generation Identification (NGI) system.
Initially a pilot program was launched by FBI in 2011. In late
F. India 2014, the facial recognition program finally became fully
The Unique Identification Authority of India (UIDAI) was operational. The technology essentially measures minute
constituted in 2009 with the objective to issue Unique distances in a person’s face and logs the information. While
Identification numbers (UID), named as “Aadhaar” to Indian these methods and techniques are still in their infancy, it could
Citizens. It’s a 12-digit unique identity number. With the help help law enforcement agencies to locate and identify a suspect
of 10 fingerprint scan and iris scan of the individuals, system using surveillance videos, mug shots or even photos taken from
ensures no duplication in generating UID. It is considered as social network sites like Facebook and Twitter.
one of the world largest national ID project. The whole system Similar to FBI, Fingerprint Analysis and Criminal Tracing
is designed in such a way that it can be easily verified by System (FACTS) is an advanced Automated Fingerprint
Government and Non-Government services. The system has Identification System used in India. The system supports
been promoted as a tool for reducing fraud in many private and storing of fingerprints of convicted, arrested, absconding
public distribution system and enabling the government to criminals in the database and supports performing automatic
better deliver public benefits. fingerprint searches against the existing criminal database. The
In this section we discussed about the National Identity system is designed with the objectives to provide efficient and
project in few countries. Some of these projects could not be effective tool to the investigative agency for faster and
implemented because of some legal complexity or lack of effective way of identifying the criminals.
support from citizens and law makers. However in many The National Crime Records Bureau (NCRB), Ministry of
countries citizen identity card system is implemented Home Affairs, Government of India has planned to develop a
successfully and it has given fruitful results to government as comprehensive online database of fingerprints of criminals and
well as citizens both. convicts from every state by 2017, a project that will help
Implementation of National Level Citizens Identity system improve investigation and prosecution all over the country. The
is huge task. It requires enormous effort and financial support. data will be connected to the Criminal and Criminal Tracking
Many times such national level identity systems are used for Network & Systems (CCTNS), a nationwide integrated online
multiple purposes and it’s quite obvious that such system system conceived after the 2008 Mumbai terrorist attacks to
cannot be developed for individual and specific purposes. track crimes and offenders.
These systems generally stores citizen’s basic information as
well as their biometric information and normally its built in IV.UNIQUE IDENTIFICATION AUTHORITY OF INDIA
such a way that it can be used for multiple purposes. In many The Unique Identification Authority of India (UIDAI) is a
countries such systems are used for identifying the forgery, tax statutory authority established under the provisions of
evasion and crime investigation. Our objective is to explore, the Aadhaar Act, 2016 by the Government of India. Prior to its
how such system can be used for law enforcement and crime establishment as a statutory authority, UIDAI was functioning
investigation. Next section discuss about the use of biometric as an attached office of the then Planning Commission.
system for crime investigation.
 2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

UIDAI was created with the objective to issue Unique V.PERFORMANCE MEASUREMENT OF BIOMETRIC
Identification numbers (UID), named as "Aadhaar", to all IDENTIFICATION SYSTEM
residents of India. The main aspect to evaluate a biometric system is its
Aadhaar system is built purely as an “Identity Platform”. accuracy. From the user’s point of view, an error of accuracy
The main purpose of Aadhaar is to provide a biometric occurs when the system fails to authenticate the identity of a
attached identity that is verifiable online. The whole system is registered person or when the system erroneously authenticates
designed in such a way that it can be easily verified by the identity of an intruder[8][9][11].
Government and Non-Government services. The system has A biometric matching system’s response is typically a
been promoted as a tool for reducing fraud in many private and matching score ‘s’ that quantifies the similarity between the
public distribution system and enabling the government to input and the database template representations. The higher the
better deliver public benefits. score, the more certain the system is that the two biometric
A. Biometrics Database in Aadhaar measurements come from the same person. A threshold t
regulates the system decision. The system infers that pairs of
The Aadhaar biometric system design has followed global
biometric samples generating scores higher than or equal to t
best practices and currently considered to be the state-of-the-
are mate pairs (that is, they belong to the same person)[10].
art. Aadhaar system has used the iris biometric in addition to
Consequently, pairs of biometric samples generating scores
fingerprint and face biometrics to achieve de-duplication
lower than t are nonmate pairs (that is, they belong to different
accuracy in excess of 99% and also ensure total inclusion.
persons)[4][5]. The distribution of scores generated from pairs
UIDAI capture the following, biometrics for the enrollment:
of samples from different persons is called an impostor
distribution; the score distribution generated from pairs of
• 10 finger prints samples from the same person is called a genuine distribution
• 2 iris scans (see Figure 1).
• Photograph of face

UIDAI has used three ABIS (Automatic Biometric

Identification System) software solution to perform biometric
de-duplication as a risk mitigation strategy. De-duplication at
this scale (1.2 billion residents) had not been previously
attempted anywhere in the world. Aadhaar is the first ever
multi-ABIS system implemented in the world, and brings
significant advantages. Aadhaar system not only assure highest
accuracy, it’s also provide a very reliable and scalable service.
B. Scalable Architecture
Aadhaar is developed to provide unique identity to more
than a billion people. With an aggressive target of reaching 600
million Aadhaars in a short span of 4 years meant that about 1
million Aadhaars need to be generated every day. Processing
of every enrolment requires matching 10 fingerprints, both
irises, and demographics with every existing record in the
Figure 1:- Biometric system error rates
database. Currently, with the Aadhaar database at 600 million,
processing 1 million enrolments every day roughly translates to
There are several factors that can provoke accuracy errors.
about 600 trillion biometric matches every day. On the other
Biometrical Error rates are quantitative metrics of the accuracy
hand, authentication is expected to perform within sub-seconds
of biometrical systems. They help biometric researchers and
even when authentication volume is few 100 million requests
industry providers to improve their solutions and help system
every day.
users to evaluate and choose the appropriate solutions for their
All this requires entire system and supporting processes to
applications. There are two types of error metrics: (a) to
scale to massive levels. Aadhaar system has already
evaluate capture devices. (b) to evaluate biometric algorithms.
demonstrated the scale by reaching 600+ million Aadhaars,
The Failure to Capture Rate (FCR) evaluates the capacity
processes 1 million enrolments every day, manages about 4000
of a capture device to properly read a biometrical record in
terabytes of data, and has deployed authentication services to
several environmental and human conditions. The failure of
handle 100 million authentications every day.
capture device is known as "Capture Error". For biometric
As we are proposing the use of Aadhaar biometric database
algorithms evaluation there are metrics associated to each
for criminal identification and crime investigation, it’s very
phase of the recognizing process[14].
important to review the accuracy of such system. Next section
The Failure to Enroll Rate (FER) measures the capacity of
discusses a very common technique for biometric performance
Enroll Component of the algorithm to detect, extract and
evaluation [12].
compose the template from the biometric record read by the
 2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

capture device, i.e, the FER is computed only for biometric TABLE I. FEASIBILITY OF IRIS BASED ON-LINE AUTHENTICATION
record properly read.
% of Residents
In the case of matching algorithm we have two important
Findings Single-Eye Dual-Eye
error rates: The False Matching Error Rate (FMR) measures
Cameras Cameras
the rate of actual impostor comparisons that are erroneously
Authenticated in 95.89 99.29
considered genuines. On the contrary the False Non matching
Error Rate (FNMR) measures the rate of actual genuine first try
comparisons that are erroneously considered impostors. Authenticated in 99.21 99.40
In the case of ranking algorithm the main ranking error rate multiple tries
is the Rank (n) that measures the rate of genuine comparisons Failed 0.79 0.60
be included among the n most similar comparisons from a authentication
database[1][3][7]. (FRR+FTC)
An operational biometric system makes a trade-off VII.CRIMINAL TRACING SYSTEM USING AADHAAR
between false match rate (FMR) and false nonmatch rate
(FNMR). In fact, both FMR and FNMR are functions of the First of all it’s very important to discuss, why we proposed
system threshold t: If the system’s designers decrease t to make the use of Aadhaar biometric database for criminal tracing?
the system more tolerant to input variations and noise, FMR One of the reasons behind use of Aadhaar biometric database
increases [6]. On the other hand, if they raise t to make the is, it has highest number of citizen enrollment with a great
system more secure, then FNMR increases accordingly. accuracy. Building a similar system for law and order will
System performance can be depicted at all operating points require a huge effort, time and money which are already
(thresholds t) in the form of a receiver operating characteristic invested to build the system like Aadhaar and it’s a great
(ROC) curve. An ROC curve plots FMR against (1 – FNMR) success.
or FNMR for various values of threshold t (see Figure 2). In India all most all the state government police department
do have some kind of their own local database of convicted
criminals and history sheeter.
The National Crime Records Bureau (NCRB) has planned
to develop a comprehensive online database of fingerprints of
criminals and convicts from every state by 2017, a project that
will help improve investigation and prosecution all over the
country. The data will be connected to the Criminal and
Criminal Tracking Network & Systems (CCTNS), a
nationwide integrated online system conceived after the 2008
Mumbai terrorist attacks to track crimes and offenders.
Normally whenever any crime occurs, police department
take the help of forensics experts to scan and preserve the
fingerprints founds at the crime scene. These fingerprints play
a vital role for the conviction of criminal as it is one of crucial
evidence. State police department normally maintain a local
Figure 2:- Receiver operating characteristic curve database of these fingerprints collected during the investigation
VI.PERFORMANCE OF AADHAAR BIOMETRIC SYSTEM of various cases. Sometime this database helps the police to
crack many other cases, if the same criminal is involved in
As fingerprints alone will not be sufficient for achieving multiple cases. However this biometric database maintained by
higher accuracy in biometric identification, Aadhaar has used police department is useful for investigating those cases where
multi-model biometric fusion techniques. UIDAI has achieved the criminals are repeated offender. However, if the people
accuracy as high as 99% de-duplication due to the fusion of involved in the crime are first time offender or a migrated
fingerprint and iris scores. offender, who has moved from some other state or place, the
A Proof of Concept (PoC) study was carried out by UIDAI local police biometrics database may be of no use for
to assess the feasibility of using iris for on-line biometric investigation. In such case, where the criminals are first time
authentication in the Indian context (Table 1). The findings of offender or a migrated criminals, centralize biometric database
this study are presented in the report. This report covers the can be a great resource for investigation.
design framework, field implementation, data collection and Building a separate biometric database only for such
analysis. It interprets empirical results to assess effectiveness of purpose and enrolling such a huge population in this system
iris technology, authentication processes and devices [12]. will be a very costly affair and tedious process and may not be
a feasible solution. As an alternate Aadhaar biometric database
can be the best suited option for this purpose. The proposed
system architecture (Figure 3) show the overall process flow of
the system.
 2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

In normal process, whenever any criminal incidents are IX.USE OF AADHAAR BIOMETRIC DATABASE- OPPORTUNITY
reported, police department with the help of forensic experts AND CHALLENGES
collect fingerprints from the crime scene. Police department No doubt such kind of huge centralize biometric database like
analyses these fingerprints and if it is matches with any Aadhaar gives lot more opportunity in application point of
fingerprints in the existing local biometric database of police view but it also raises some challenges and its worth
department, then it’s easy for the police to identify the mentioning here. It looks very simple to use Aadhaar for
criminals as most likely the records of such criminal may be criminal tracing purpose but it has some technical and legal
available with the police. If the fingerprints do not match with hurdles.
the local database then it can be forwarded to central finger
print database, which is a repository database of all the states A. Opportunity and Advantages
fingerprint database. The central biometric database can be Aadhaar has highest number of citizen enrollment till
useful, if the criminals are repeated offenders and migrated. today. It’s one of the largest systems in the world storing such
However, if the criminals are first time offenders then central a huge number of Biometric information. The system is very
database may be of no use and hence it can be analyzed with reliable and scalable. Building similar type of system only for
the help of Aadhaar database. forensic and criminal investigation purpose may not be feasible
and viable solution. Since Aadhaar is already well established
VIII.ERROR TOLERANCE AND CASE OF MISTAKEN IDENTITY system, its uses for the other purposes can be explored.
What happens when the match turns out to be wrong?
Though, the Aadhaar biometric performance is more than 99%. B. Legal Hurdle
In Indian context error rate between 0.25% - 1% is a huge Aadhaar was never intended for Forensic purpose and thus
number. It will falsely identify 10-11 Million people as a it was not designed in that way. The citizens who enrolled for
suspect. Aadhaar never gave his consent of using his information for
One of the advantage of Aadhaar database is, it also store such investigations.
the demographic information of the enrolled citizens. In the C. Technical Challenge
case of false identification, demographic information can be
used as a filter to reduce the false identification. If we study the UIDAI’s own technical report, the system
has approximately a chance of 0.5 per cent occurrence of false
identification. In Indian context this fractional values could
possibly make Lakhs of people as Suspect. We can narrow

Figure 3:- Architecture of Aadhaar Based Criminal Tracing System

 2017 International Conference on Innovations in Information, Embedded and Communication Systems (ICIIECS)

down this number but still chances are there that innocent may [13]V. Nagaraju and L. Fiondella, "A survey of homeland security
suffer, if the investigative agency solely depends on the system biometrics and forensics research," 2016 IEEE Symposium on
identification process. Technologies for Homeland Security (HST), Waltham, MA,
Now though the system has some legal hurdles and 2016, pp. 1-7.
technical challenges, it seems there is no other alternatives [14]Z. u. H. Usmani, S. Irum, H. Afzal and S. Mahmud, "How to
better than the Aadhaar. Build an Automated Fingerprint Identification System," 2013
International Symposium on Biometrics and Security
ACKNOWLEDGMENT Technologies, Chengdu, 2013, pp. 40-47.

I would like to thank MSRIT Management and my colleagues

for their valuable suggestion, constant support and
encouragement.
REFERENCES
[1] A. Sankaran, M. Vatsa and R. Singh, "Latent Fingerprint
Matching: A Survey," in IEEE Access, vol. 2, no. , pp. 982-
1004, 2014.
[2] Anil K Jain, Arun Ross, "Bridging the gap: From biometrics to
forensics", Philosophical Transactions of The Royal Society B
Biological Sciences 370(1674),August 2015.
[3] Anil K. Jain, Salil Prabhakar, and Sharath Pankanti. 2001. Twin
Test: On Discriminability of Fingerprints. In Proceedings of the
Third International Conference on Audio- and Video-Based
Biometric Person Authentication (AVBPA '01), Josef Bigün
and Fabrizio Smeraldi (Eds.). Springer-Verlag, London, UK,
UK, 211-216.
[4] Bouchrika, I., Goffredo, M., Carter, J. and Nixon, M. (2011), On
Using Gait in Forensic Biometrics. Journal of Forensic Sciences,
56: 882–889.
[5] D. Brown and K. Bradshaw, "A multi-biometric feature-fusion
framework for improved uni-modal and multi-modal human
identification," 2016 IEEE Symposium on Technologies for
Homeland Security (HST), Waltham, MA, 2016, pp. 1-6.
[6] D. Meuwly and R. Veldhuis, "Forensic biometrics: From two
communities to one discipline," 2012 BIOSIG - Proceedings of
the International Conference of Biometrics Special Interest
Group (BIOSIG), Darmstadt, 2012, pp. 1-12.
[7] http://www.griaulebiometrics.com/en-us/book/understanding-
biometrics/evaluation/accuracy
[8] Q. Gao and D. Pinto, "Some challenges in forensic fingerprint
classification and interpretation," 2016 IEEE Long Island
Systems, Applications and Technology Conference (LISAT),
Farmingdale, NY, 2016, pp. 1-7.
[9] Rainer Kärgel, Mario Hildebrandt, and Jana Dittmann. 2012.
An evaluation of biometric fingerprint matchers in a forensic
context using latent impressions. In Proceedings of the on
Multimedia and security (MM&Sec '12). ACM, New York, NY,
USA, 133-138.
[10]Salil Prabhakar, Sharath Pankanti, and Anil K. Jain. 2003.
Biometric Recognition: Security and Privacy Concerns. IEEE
Security and Privacy 1, 2 (March 2003), 33-42.
[11]Tobias Kiertscher, Robert Fischer, and Claus Vielhauer. 2011.
Latent fingerprint detection using a spectral texture feature. In
Proceedings of the thirteenth ACM multimedia workshop on
Multimedia and security (MM&Sec '11). ACM, New York, NY,
USA, 27-32.
[12]Unique Identification Authority of India, Role of Biometric
Technology in Aadhaar Authentication IRIS Authentication
Accuracy – PoC Report Date: 14 September 2012
(https://uidai.gov.in/images/iris_poc_report_14092012.pdf).