19-07-2014, 12:55 PM
Altered Fingerprints: Analysis and Detection
[attachment=66004]
Abstract
The widespread deployment of Automated Fingerprint Identification Systems (AFIS) in law enforcement and border control
applications has heightened the need for ensuring that these systems are not compromised. While several issues related to fingerprint
system security have been investigated, including the use of fake fingerprints for masquerading identity, the problem of fingerprint
alteration or obfuscation has received very little attention. Fingerprint obfuscation refers to the deliberate alteration of the fingerprint
pattern by an individual for the purpose of masking his identity. Several cases of fingerprint obfuscation have been reported in the
press. Fingerprint image quality assessment software (e.g., NFIQ) cannot always detect altered fingerprints since the implicit image
quality due to alteration may not change significantly. The main contributions of this paper are: 1) compiling case studies of incidents
where individuals were found to have altered their fingerprints for circumventing AFIS, 2) investigating the impact of fingerprint
alteration on the accuracy of a commercial fingerprint matcher, 3) classifying the alterations into three major categories and suggesting
possible countermeasures, 4) developing a technique to automatically detect altered fingerprints based on analyzing orientation field
and minutiae distribution, and 5) evaluating the proposed technique and the NFIQ algorithm on a large database of altered fingerprints
provided by a law enforcement agency. Experimental results show the feasibility of the proposed approach in detecting altered
fingerprints and highlight the need to further pursue this problem
INTRODUCTION
FINGERPRINT recognition has been successfully used by
law enforcement agencies to identify suspects and
victims for almost 100 years. Recent advances in automated
fingerprint identification technology, coupled with the
growing need for reliable person identification, have
resulted in an increased use of fingerprints in both
government and civilian applications such as border
control, employment background checks, and secure facility
access [2]. Examples of large-scale fingerprint systems in the
US government arena include the US-VISIT’s IDENT
program [3] and the FBI’s IAFIS service [4]
ANALYSIS OF ALTERED FINGERPRINTS
Based on a database of altered fingerprints made available
to us by a law enforcement agency, we first 1) determine
the impact of fingerprint alteration on the matching
performance, 2) categorize altered fingerprints into three
types2
: obliteration, distortion, and imitation (see Figs. 9, 10,
and 11), and 3) assess the utility of an existing fingerprint
quality measure in terms of altered fingerprint detection.
AUTOMATIC DETECTION OF ALTERED FINGERPRINTS
In the previous section, we showed that the NFIQ algorithm
is not suitable for detecting altered fingerprints, especially
the distortion and imitation types. In fact, the distorted and
imitated fingerprints are very hard to detect for any
fingerprint image quality assessment algorithm that is
based on analyzing local image quality. In this section, we
consider the problem of automatic detection of alterations
based on analyzing ridge orientation field and minutiae
distribution. The flowchart of the proposed alteration
detector is given in Fig. 13
Analysis of Orientation Field
Orientation field4 describes the ridge flow of fingerprints
and is defined as the local ridge orientation in the range
½0; Þ. Good quality fingerprints have a smooth orientation
field except near the singular points (e.g., core and delta).
Based on this fact, many orientation field models have been
developed by combining the global orientation field model
for the continuous flow field of the fingerprint with the local
orientation field model around the singular points [34], [35],
[36]. The global orientation field model represents either
arch-type fingerprints, which do not have any singularity,
or the overall ridge orientation field except singularity in
fingerprints. If the global orientation field model alone is
used for orientation field approximation, the difference
between the observed orientation field and the model will
ideally be nonzero only around the singular points. On the
other hand, for obfuscated fingerprints, the model fitting
error is observed in the altered region as well. Thus, we use
the difference between the observed orientation field
extracted from the fingerprint image and the orientation
field approximated by the model as a feature vector for
classifying a fingerprint as natural fingerprint or altered
one. The main steps of the proposed algorithm are
described below:
Analysis of Minutiae Distribution
A minutia in the fingerprint indicates ridge characteristics
such as ridge ending or ridge bifurcation. Almost all
fingerprint recognition systems use minutiae for matching.
In addition to the abnormality observed in orientation field,
we also noted that minutiae distribution of altered
fingerprints often differs from that of natural fingerprints.
Based on the minutiae extracted from a fingerprint by the
open source minutiae extractor in NBIS, a minutiae density
map is constructed by using the Parzen window method
with uniform kernel function. Let Sm be the set of minutiae
of the fingerprint
EXPERIMENTS
The proposed algorithm was evaluated at two levels:
finger level (one finger) and subject level (all 10 fingers).
At the finger level, we evaluate the performance of
distinguishing between natural and altered fingerprints.
At the subject level, we evaluate the performance of
distinguishing between subjects with natural fingerprints
and those with altered fingerprints. Since most AFIS used
[attachment=66004]
Abstract
The widespread deployment of Automated Fingerprint Identification Systems (AFIS) in law enforcement and border control
applications has heightened the need for ensuring that these systems are not compromised. While several issues related to fingerprint
system security have been investigated, including the use of fake fingerprints for masquerading identity, the problem of fingerprint
alteration or obfuscation has received very little attention. Fingerprint obfuscation refers to the deliberate alteration of the fingerprint
pattern by an individual for the purpose of masking his identity. Several cases of fingerprint obfuscation have been reported in the
press. Fingerprint image quality assessment software (e.g., NFIQ) cannot always detect altered fingerprints since the implicit image
quality due to alteration may not change significantly. The main contributions of this paper are: 1) compiling case studies of incidents
where individuals were found to have altered their fingerprints for circumventing AFIS, 2) investigating the impact of fingerprint
alteration on the accuracy of a commercial fingerprint matcher, 3) classifying the alterations into three major categories and suggesting
possible countermeasures, 4) developing a technique to automatically detect altered fingerprints based on analyzing orientation field
and minutiae distribution, and 5) evaluating the proposed technique and the NFIQ algorithm on a large database of altered fingerprints
provided by a law enforcement agency. Experimental results show the feasibility of the proposed approach in detecting altered
fingerprints and highlight the need to further pursue this problem
INTRODUCTION
FINGERPRINT recognition has been successfully used by
law enforcement agencies to identify suspects and
victims for almost 100 years. Recent advances in automated
fingerprint identification technology, coupled with the
growing need for reliable person identification, have
resulted in an increased use of fingerprints in both
government and civilian applications such as border
control, employment background checks, and secure facility
access [2]. Examples of large-scale fingerprint systems in the
US government arena include the US-VISIT’s IDENT
program [3] and the FBI’s IAFIS service [4]
ANALYSIS OF ALTERED FINGERPRINTS
Based on a database of altered fingerprints made available
to us by a law enforcement agency, we first 1) determine
the impact of fingerprint alteration on the matching
performance, 2) categorize altered fingerprints into three
types2
: obliteration, distortion, and imitation (see Figs. 9, 10,
and 11), and 3) assess the utility of an existing fingerprint
quality measure in terms of altered fingerprint detection.
AUTOMATIC DETECTION OF ALTERED FINGERPRINTS
In the previous section, we showed that the NFIQ algorithm
is not suitable for detecting altered fingerprints, especially
the distortion and imitation types. In fact, the distorted and
imitated fingerprints are very hard to detect for any
fingerprint image quality assessment algorithm that is
based on analyzing local image quality. In this section, we
consider the problem of automatic detection of alterations
based on analyzing ridge orientation field and minutiae
distribution. The flowchart of the proposed alteration
detector is given in Fig. 13
Analysis of Orientation Field
Orientation field4 describes the ridge flow of fingerprints
and is defined as the local ridge orientation in the range
½0; Þ. Good quality fingerprints have a smooth orientation
field except near the singular points (e.g., core and delta).
Based on this fact, many orientation field models have been
developed by combining the global orientation field model
for the continuous flow field of the fingerprint with the local
orientation field model around the singular points [34], [35],
[36]. The global orientation field model represents either
arch-type fingerprints, which do not have any singularity,
or the overall ridge orientation field except singularity in
fingerprints. If the global orientation field model alone is
used for orientation field approximation, the difference
between the observed orientation field and the model will
ideally be nonzero only around the singular points. On the
other hand, for obfuscated fingerprints, the model fitting
error is observed in the altered region as well. Thus, we use
the difference between the observed orientation field
extracted from the fingerprint image and the orientation
field approximated by the model as a feature vector for
classifying a fingerprint as natural fingerprint or altered
one. The main steps of the proposed algorithm are
described below:
Analysis of Minutiae Distribution
A minutia in the fingerprint indicates ridge characteristics
such as ridge ending or ridge bifurcation. Almost all
fingerprint recognition systems use minutiae for matching.
In addition to the abnormality observed in orientation field,
we also noted that minutiae distribution of altered
fingerprints often differs from that of natural fingerprints.
Based on the minutiae extracted from a fingerprint by the
open source minutiae extractor in NBIS, a minutiae density
map is constructed by using the Parzen window method
with uniform kernel function. Let Sm be the set of minutiae
of the fingerprint
EXPERIMENTS
The proposed algorithm was evaluated at two levels:
finger level (one finger) and subject level (all 10 fingers).
At the finger level, we evaluate the performance of
distinguishing between natural and altered fingerprints.
At the subject level, we evaluate the performance of
distinguishing between subjects with natural fingerprints
and those with altered fingerprints. Since most AFIS used