Abstract— In recent year’s person identification using his physical or behavioral
trait has become one of the ground breaking formulas for authentication.
Authentication is the primary step in access control of any system. In accessing
ones elementary mobile device to ingressing huge industrial server
authentication plays a significant role. Authentication can be based on
something one knows, something one has or something one is. Biometrics is a
genre of something one is, it can either be a physical or a behavioral trait of
a person. This work presents a survey on secure biometric genres used in
authentication techniques with major outlook on comparison of passwords with gesture
Authentication, Password, Gesture, Trait
means life measurement, “bio” means “life” and “metric” means
“measurement”. Biometrics universally is the study
of automated identification of an individual, by utilizing one’s own behavioral or physical attributes. Application of Biometric dates back to ancient Egypt,
during the construction of the great pyramid of Khufu, it is a ground breaking concept
to apply biometrics in Encryption techniques and authentication process.
Biometrics is fascinating because they are instinctive possessions of an
individual and there is no necessity to remember them like passwords 1.
as stated earlier can be broadly classified into behavioral biometrics and
physical or physiological biometrics which can be either unimodal or multimodal.
Physical biometric represents physical characteristics of a person and this
list includes face 2, fingerprint, finger-knuckle print, ear, palm print,
iris, hand geometry, Dental X-ray and retina, to mention a few. Behavioral
biometric is an avant-garde technology which involves uniquely identifying
humans based on few measurable patterns in human activities. Few of its types
include voice of a person, signature, odor, keystroke, gait, Deoxyribonucleic Acid
(DNA), Electrocardiogram (ECG), Thermo gram and gesture limiting to only few of
the examples being mentioned 9.
This paper is organized into following
sections. Section II briefs about related work in the area of Biometrics. Section III details different Biometric genre.Section IV
throws light into the comparison of password based authentication and Gesture based authentication considering
various parameters into account for the study and also lists the areas
of biometric application. Section
V concludes the comparison between these authentication techniques.
Wang ,Stark C. Draper and Prakash Ishwar1 provided an insight into secure
biometrics or in a layman’s language known as protection of the generated biometric template from an attacker. They
also discussed about biometric signal model for processing biometric signals
and biometric performance measures.
Unsang Park and
Anil K. Jain, explained about face biometric recognition2. They initiated to use analytical information and soft facial biometric features for
improving face image matching and retrieval performance. They developed an automatic facial mark
detection method and provided experimental results based on FERET database and
forensic domain based mugshot databases. They proved that face-recognition
performance of state-of-art matcher improves by use of soft biometric traits.
Xi Zhao, Shi and Ioannis A. Kakadiaris, illustrated
the application of statistical touch dynamics image to retain discriminative
power for user authentication 3. The proposed method significantly reduced
computational time in online authentication processes. They also evaluated the
usability and effectiveness of this method and implemented as App in Android
Xiaoqian Jiang, Lucila, Ohno-Machado, Lijuan Cui, Samuel Cheng, Hongkai Xiong, proposed a method that involves
non-asymmetric SW codes4. This method involves following steps, biometric
feature extraction, non-asymmetric SW encryption/decryption and lastly Privacy
preserving authentication, which aims at fixing security holes within existing
P.N, Elyor Kodirov, Deok-Jai Choi,
Guee-Sang Lee , Mohd Fikri Azli A and
Shohel Sayeed, proposed a novel approach for implicitly authenticating smart
phone users using both shape and texture information of an ear image captured
during the call 5. This approach involves extracting and combining all Local
Binary Pattern to form a histogram and also involves extracting ear geometry
features by using ear location centre that can be adjusted very easily by
users. The authors concluded that is method is easy and simple to
Kresimir Delac and
provided an overview of biometric methods. They provided an insight into both
the unimodal and multimodal biometric and their advantages and disadvantages.
already stated Biometric is broadly classified into two genres with many types
falling under each category. A third genre of biometric can be formed by
amalgamation of physical and behavioral biometric and it is avowed as Hybrid
Biometric 13 15. This section
describes various biometrics employed in diverse spheres.
A. Finger print Biometric
Fingerprint technology is the most
widely accepted form of biometric authentication method. The fingerprint
specimen consists of minutiae, furrows, valleys and ridges at its tip. The
accuracy of matching the template with the input biometric is very high since
the input directly is in contact with the reader device. In real-time fingerprint
matching, matching can be performed using any of these methods, minutiae based,
ridge based or correlation based methods 4 16.
B. Finger-Knuckle Print Biometric
Finger-knuckles are the joints of
fingers that are visible prominently when fist is made by clenching the hand. Finger-Knuckle
Print (FKP) biometric is an upcoming technology which is gaining popularity in
recent years in authentication process. FKP biometric is relatively stable
since it remains unchanged throughout the lifetime of a person. In methods
using FKP biometric, Region of Interest (RoA) is first determined and then nearest
neighbor classifier is used for recognition and comparison. To obtain 100
percent accuracy using FKP in authentication process, multiple finger types
yield better result 8 16.
C. Palm print Biometric
Palmprint biometric is similar to
Fingerprint in having principle lines, valleys and ridges similar to a finger
print. Palm print can be characterized as having distinctive features like wrinkles,
width of fingers, Meta Carpo Phalangeal (MCP), Distal Inter Phalangeal (DIP),
Proximal Inter Phalangeal (PIP), overall palm area and circularity, the
perimeter and elongation fingers indices, which various for every person .Systems
involving this method for authentication uses Principle Component analysis strategy
and hand geometry features 10.
D. Hand Geometry Biometric
Hand Geometry features may involve
height of fingers and distance between them, width of the palm, angle between
the fingers and hand dimensions. Systems employing this form of biometric for
authentication may have error rates lesser than 10 percent in verification
process. Hand geometry can be reliable when combined with other forms of
biometric for identification 11.
E. Iris Biometric
Iris is complex human phenotype (externally visible
trait) and that portion of eye which surrounds the pupil and is colored. It is
also claimed that the combination of dark spots, filaments, pits, striations,
rings, and freckles add for the accuracy of iris biometric identification. It
is believed that iris texture patterns are so unique that both the eyes of same
individual are different. 6 15 22
F. Retina Biometric
Retina is that portion of the eye
embedded deep in its rear hemisphere with tissues and vessels, forming a person’s
unique identity. Retina recognition technology is based on the differentiation
of unique vascular patterns from eye to eye and person to person and nearly
impossible to acquire retina images of a person without his consent 14 15.
G. Face Biometric
Face recognition and retrieval system
uses facial images consisting of local and global texture features. The
prominent components of face include eyes, nose, mouth, ears and chin. Face
matching can also be performed using soft facial biometric traits unique to an
individual and they include moles, scares, pockmark which may be permanent and
the non permanent facial features may include wrinkle, freckle, scares,
abrasion, acne, dark skin, whitening and many more 2. Accuracy of face
matching may increase if expressions of subjects are not taken into account.
H. Ear Biometric
Ear biometric is one of the easily
employed passive forms of biometric. The merits of using ear biometrics in
authentication process over other forms of biometrics are, shape of the ear and
the cartilage tissue of the pinna are distinctive and does not change during
life time of a person, color distribution of ear remains uniform and unchanged
so much information is not lost during ear image processing, since ear size is
small, one can work efficiently and effectively with lower resolution involved 5.
I. Dental X-ray Biometerics
This form of biometric is an intelligent
approach which involves identifying an individual based on his naturally unique
dental features. As suggested by many researchers this method has lesser error
rate when compared to finger print or palm print methods 12.
J. Vein Biometrics
Veins are elastic blood vessels that
transport blood to heart from various regions of the body and are prominent
features of any person’s hand. These Vascular patterns have strong immunity to
forging since they are not visible to naked eyes and lie underneath the skin. To
capture subjects vein pattern infrared sensors are used and when the palm is
illuminated a dark pattern appears which is visible due to deoxidized
hemoglobin in the palm veins 23
K. Voice Biometric
Voice biometric has made the world of technology
more easily adoptable and is true since most of devices used by humans today
work on speech commands.
Voice of an individual is predominantly
dependent on his physical traits used in creation of sound like mouth, vocal
tracts, larynx, pharynx, nasal cavities and lips. The behavioral trait of an
individual may change according to his emotional state, medical conditions and
age. Recorded audio files of
varying types and their differing qualities are the primary challenge in voice
recognition systems. Obtaining all levels of training data is an added issue
and also the distance of the individual from the recording device can be a
L. Signature Biometric
is a unique expression by an individual’s hand geometry and is a traditional
style of metric for person authentication and identification. Though many
theories relating to signature based authentication systems are proposed, no
due attention is rendered in employing them.
some instances signatures belonging to different subjects may look similar and
two genuine signature instances of same user may have variations in texture and
other dynamic property. To declare a signature genuine an Elastic distance
matching algorithm may be made use of and the signature characteristics
M. Gait Biometric
Gait of a person is the way he walks and
is influenced by his physical or psychological status such as emotion, fatigue,
well-being, etc. In addition, external factors such as clothing, shoes, carried
loads, and ground characteristics also influence gait. Gait is a complex
spatio-temporal biometrics measured using the concept of Bio-kinematics and is
not a distinctive feature 21.
N. Keystroke Biometric
A prevalent belief in major class of
society is that every individuals typing style or pattern on keyboard is
unique. It provides discriminatory information even though not completely
distinctive. The merits of using keystroke biometrics is that since the person
using the system is keying information he can be easily monitored, but it
majorly does not contribute to privacy and in turn security. The application of
keystroke biometrics and databases consisting of keystroke templates majorly
affect the performance of keystroke biometrics system 20.
O. Odor Biometric
Odor is a predominant feature of every
individual and is caused generally at a very lesser concentration by one or
more volatilized chemical composition. They are spread around and can either be
pleasant or unpleasant. To employ this type of biometric in authentication
process an array of sensors are to be used, to identify the odor and it becomes
a complex process to synchronize the array of sensors. Though this is a promising
approach it is still in infancy and needed to be well established in the field
of Biometric authentication systems 19.
P. DNA Biometric
DNA technology is the most widely
accepted form of complex human genotypes biometric authentication means. DNA or Deoxyribonucleic acid of an
individual is very unique and probably most reliable feature which cannot be
forged. The DNA pattern infact is a one dimensional code which is unique to a
person and preset in every cell of an individual’s body. The limitations of DNA
biometrics are, since it is present in every cell of a person’s body, it is
easy to steal, but DNA matching involves complex process so it is difficult to
employ in real time by the masses 22.
Q. ECG Biometric
(ECG) is an emerging modality in the field of Biometric recognition systems
which are difficult to counterfeit. ECG depends on structural and functional
characteristics of the heart and contains information pertaining to the
electrical operations of the heart which are unique and highly personalized. In
short ECG is a record of heart’s electrical activity. Researchers have stated
several issues that need to be solved before using ECG in biometric systems for
authentication few of them are ECG signal changes over time and heart rate variability
due to ageing 24.
R. Thermogram Biometric
It is an emerging modality in the field
of Biometric Authentication Techniques. Thermo gram relates to capturing heat
radiation pattern emitted from human body parts like face, hand, veins and so
on, using infra-red cameras and since humans
maintain constant temperature and are homoeothermic When compared to
visual images, thermal imaging has better accuracy.
Facial Thermo gram allows one to passively
perform detection in different circumstances by recording facial temperature
variations caused by vein structure on facial surface 25.
S. Gesture Biometric
Gesture biometrics is one of the most
interesting and trendsetting innovation of previous few decades. Gesture refers
to non verbal communication or movement of body parts usually hand or fingers to
indicate a person’s intention. Gestures are revocable that is when a gesture is
no longer effective it can be replaced and since it does not involve touch on
any device, forging is difficult. Having these merits gestures are widely used
now a days in person identification and authentication process 3 7 26.
Analysis And Discussion
This section provides an insight
into the comparison of passwords with one of the biometrics based
authentication methods and also the areas of application of biometrics.
Comparision between Passwords and Gesture Biometric
comparative study between two classes of authentication is tabulated in TABLE
I. The two authentication classes that are gesture authentication represented
in table as “Ges” and the classical password authentication method represented
as “Pwd” are compared with respect to the parameters Interest, Ease of Use and
Ease of remembrance.
of Password and Gesture authentication Approaches
Password Authentication Vs Gesture Authentication
Ease of Use
Ease of Remembrance
‘I’ in above table
represent Interest of the user, ‘N’ represents Neutral, ‘NI’ represents Not
Interested, ‘D’ represents Difficult, ‘VD’ represents Very Difficult, ‘E’
represents Easy, ‘VE’ represents Very Easy, ‘A’ represents Average.
Comparision between Password and Gesture authentication
The above table was
formulated by giving same type of gesture and password to each of the subjects
under study. With respect to their Interest in using the password or gesture,
Ease of their Use and Ease of remembrance the results are correlated, compared and
tabulated. This study was consolidated for a period of nine months involving 63
subjects belonging to different age group.
For a period of three
months first study was conducted. Involving same subjects, for a period of next
six months, second study was conducted and the third study was done for a
period of successive nine months.
Fig.1 depicts the
tabulated Comparative study between password and gesture based authentication
X-axis represents Range
of Ages (RoA) i.e., different age group considered for study of the two methods
of authentication and Y-axis represents Average Percentage of Correlation
between the Parameters considered for Gesture and Password (APCPGP) for 9
months. The graph was derived by assuming values for all the variables
considered in Table I.
genre can also be evaluated based on the Biometric characteristics listed here
and they are universality, uniqueness, permanence, measurability, performance,
acceptability, circumvention. There can also be many other parameters not
listed here that can be used to evaluate various biometric genre.
When conducting this
study special case related to memory loss like Dementia, Alzheimer’s and mental
illness were not taken into account
B. Areas of Applications
The most revolutionary idea of using
what a person himself is, in various fields, is a cutting edge technology and
known as Biometric as already stated. Biometric concept can be employed in
various areas and few of them are listed here.
Biometric are used in access control
systems to provide various levels of security.
Biometric are used in person
Identification and authentication.
They are used in Encryption algorithms
in various Networks.
They are increasingly useful for security applications in the banking industry and various web applications.
Biometric are used in security
applications like video surveillance,
airport checking, computer or mobile devices systems, etc.
Used in the field of forensic medicine
to find culprits in the area of crime.
The study of biometric has a voluminous scope and
future. Biometric and its use in security solutions for networks and in
authentication process is a ground breaking technology. Biometrics can be
easily employed since the burden of remembering neither the passwords nor the
issues related to maintaining any smart cards is involved.
Biometric and its applications will rule the world of
encryption and authentication in mere future. Biometric applications not
limiting to only above said areas, currently their usage in home automation is
catching an eye. In mere future there will be no area spared where in biometric
technology is not made use of.
This paper is merely an effort to provide an insight
into biometrics and its classes. There are currently ample of ongoing
researches in this field to explore the unexplored within us and use them for
the benefit of mankind.