Secure Template Protection In Touch-less Based Fingeprint Verification System
Hiew Bee Yan
Faculty of Information Science and Technology
Multimedia University,
Jalan Ayer Keroh Lama,
75450 Melaka, Malaysia.
e-mail: byhiew@mmu.edu.my
Andrew Teoh Beng Jin
School of Electrical and Electronic Engineering,
Yonsei University, Seoul,
South Korea.
e-mail: bjteoh@yonsei.ac.kr
Abstract— Cancellable biometrics has been a challenging and
essential approach to protect the privacy of biometric
templates. Multiple Random Projections (MRP) is our
formerly presented two-factor cancellable formulation. In that
method, the biometric data is changed in a revocable but non-
invertible manner by projecting every fixed length feature
vector (extracted from the raw biometrics) onto a user-specific
random subspace. In this paper, we propose a variant of MRP,
namely Multiple Random Projections-Support Vector Machine
(MRP-SVM). The MRP’s template protection characteristics
are inherited by MRP-SVM due to existence of the property of
dot product and non-linear kernel. Furthermore, the
verification performance is improved. This approach is
verified using the touch-less based acquired fingerprint images.
Touch-less based acquired images are free from latent
fingerprint issues that can lead to fraudulent use. Hence, the
security and privacy protection of fingerprint biometric
templates is consolidated by the cancellable biometrics
approach.
Keywords-touch-less based fingerprint verificatio; template
protection; Support Vector Machine
I. INTRODUCTION
It is commonly known that the biometrics trait of a
person cannot be easily replaced. Once a biometrics is ever
compromised, it would mean the loss of a user’s identity
forever [1]. Therefore, protecting the biometric templates is a
major concern and also a challenging task [2]. Cancellable
biometrics is a concept where the biometric template is
secured by incorporating protection and the replacement
features into biometrics. Fundamentally, cancellable
biometrics alters the biometric images or features before
matching. The variability in the distortion parameters offers
the cancellable nature of the scheme. A good cancellable
biometrics formulation must fulfil four requirements [3]: (i)
Diversity: The same cancellable template cannot be
employed in two different applications; (ii) Reusability:
Straightforward revocation and reissue in the occurrence of
compromise; (iii) One-way transformation: Non-invertibility
of template calculation to avoid recovery of biometric data;
(iv) Performance: The recognition performance should not
be deteriorated by the formulation.
A. Related Works
Ratha et al initiated the first notion of cancellable
biometric formulation. The basic idea is to generate
deformed biometric data by distorting the biometric image in
a repeatable but non-reversible manner [2]. They realized
this idea on fingerprint minutiae by introducing three non-
reversible transformation functions, namely Cartesian, polar
and surface folding transformations [4]. Though they
achieved diversity, reusability and one-way transformation,
the experimental results show that the performance using the
transformed template is degraded. Savvides et al. [5]
encrypted the training images which are used to synthesise
the correlation filter for biometrics authentication. They
showed that different templates can be obtained from the
same biometrics by varying the random convolution kernels.
Thus, those templates could be cancelled. But, a
deterministic deconvolution with a known random kernel
would jeopardise its security. Ang et al. [6] generated a
cancellable fingerprint template through a key-dependent
geometric transformation on a minutiae based fingerprint
representation. Nevertheless, the matching accuracy was
degraded notably in the distorted domain. In [7], the authors
brought in the concept of biometric-based tokens that support
robust distance computations, which offer cryptographic
security such that it can be revoked and replaced a new one.
Another cancellable biometrics approach was introduced by
Teoh et al. [8] using an iterative dot product between the
tokenised pseudo-random number (PRN) and the biometric
data. However, this formulation suffered from verification
performance degradation when the genuine token was stolen
and used by an impostor to claim as the genuine user (stolen-
token scenario). Teoh and Chong [9] introduced Multispace
Random Projection (MRP) as one of the cancellable
biometrics approaches in face recognition. MRP fulfils all of
the good cancellable biometrics requirements as stated
above. However, the Equal Error Rate (EER) result is only
30% due to the poor classification capability of the matching
metric. In order to protect the privacy of biometric templates
while keeping the ability to match the protected data against
a reference, Julien Bringer et al. [10] applied secure sketches
to cancellable biometrics. Most recently, the inferior
performance of BioHash in stolen-token scenario was
circumvented by a user-dependent multi-state
discretisation(Ud-MsD) method [11]. This method is an
extension of a BioHash method that incorporated a Ud-MsD
instead of utilising the original simple thresholding scheme.
The Ud-MsD provides a better accuracy and a more secured
biometric template than that of the original BioHash
technique. Besides, it can render a long bit string, which is a
useful feature to resist brute-force attacks.