 |
| |
 |
|
 |
| |
 |
|
| From the users' point of view, Biometrics is gaining wide
popularity for two main reasons: |
|
 Security and
Management |
The biometrics market arose
from the need to prevent any risks or damages in various
fields and it uses identity authentication to answer that
need. Its uses range from physical control and management
including entrance control, diligence management, and machine
access control to security and management in the information
industry, including computer security, distance education,
e-commerce, information security, and so on. These days,
computer processing of so much important information fosters
information exchange in the cyberspace, and expands e-commerce
fields to include online banking.
Against this backdrop,
the demand for valid identity authentication is soaring along
with the growth of the related businesses. |
|
| Convenience |
Convenience is one of the
greatest advantages of biometrics compared to existing methods
of personal authentication such as keys, identification
numbers (ID) and passwords. In other words, everyone can be
uniquely identified without the need for an ID, a magnetic
card, a smart card, a key or a personal identification number
(PIN).
A user can verify each individual by using only
physical traits such as fingerprints and voice.
Also,
using biometrics, a machine can electronically recognize a
user, thereby enabling its system to allow for the automatic
response to that user's request. In short, biometrics is able
to deliver both safety and convenience in the identity
verification field, thereby reaping huge economic benefits. |
|
 |
|
A biometrics system should use
personal traits developed with the following "ideal" criteria:
universal (everyone has it), unique (no two people have it
alike), permanent (does not change and cannot be changed),
collectable (easy to obtain and quantify with a sensor).
Concerns during the system design should include
Performance (precision, speed, tenacity, resource request,
operational/environmental components), Acceptability (the
degree of acceptance by people), and Circumvention (how easy
it is to outsmart the system). The traits also become a
yardstick for the system evaluation.
All biometric
technologies invariably have their own advantages and
disadvantages. Table 1-1 shows the characteristics index of
various kinds of biometric technologies. |
|
| Biometrics |
Universal |
Unique |
Perman-ence |
Collectable |
Perfor-mance |
Accepta-bility |
Circum-
vention |
| Face |
High |
Low |
Medium |
High |
Low |
Low |
High |
| Fingerprint |
Medium |
High |
High |
Medium |
High |
Medium |
High |
| Iris |
High |
High |
High |
Medium |
High |
Low |
High |
| Signature |
Low |
Low |
Low |
High |
Low |
High |
Low |
| Voice |
Medium |
Low |
Low |
Medium |
Low |
High |
Low |
| Vein |
Medium |
Medium |
Medium |
Medium |
Medium |
Medium |
High |
| DNA |
High |
High |
High |
Low |
High |
Low |
Low |
|
|
| 1.1 Index of Biometrics
Types |
 |
 |
|
| Biometrics using Physical
Characteristics |
| Fingerprints, face, iris, vein,
cornea, hand, DNA pattern, ear, etc |
|
 Relatively
stable |
| Does not
change much in a lifetime |
| Huge,
expensive equipment needed. Intrusive method |
|
| Biometrics using Behavioral
Traits |
| Signature,
voice recognition, walking style |
| Change
according to psychological condition |
| Influenced by
physical traits (men/women, build) |
| Change a lot |
| Simple,
inexpensive equipment. Non-.offensive method |
|
 |
|
 |
|
| Figure 1.1 Fingerprint
Enrollment and Authentication Processes in a General
Fingerprint Recognition System |
|
Figure 1.1 is a concept map to
show the process of fingerprint enrollment and authentication
in the system. All biometrics systems identify an individual
according to the process. The first step in the biometrics
system application is the enrollment of a user in the system.
The registered biological information is raw data. By
processing the data with signal algorithms such as image
process, feature data to represent vital information of that
individual can be obtained. The feature data will play the
role of the Template. Storing the Template in a database
finishes the user enrollment process. After enrollment, the
user should provide his or her bio-information such as
fingerprints or facial planes to use the system.
The
biometrics system processes the information and gets his or
her feature data and compares the data with the Template,
thereby calculating similarity or matching score. After that,
it matches the score with the critical value set in the
initial setting to finally verify the identity. Figure 1.2
shows the biometrics system in logical blocks. |
|
 |
|
| Figure 1.2 Biometrics System
Logical Blocks |
|
 |
|
| Verification |
|
| By providing a user's ID and
bio-information to the system at the same time, the user
activates the system to compare his or her feature data
obtained from the entered bio-information with the stored
Template of the corresponding ID. This process of 1:1 matching
is called verification or authentication. |
|
| This process is being applied
to various business areas such as access control and payment
authentication where a PIN is needed. For example, this can be
used for a door-lock so that a user enters his or her ID and
has the door reading the fingerprint. Or it can also be used
for PC Log-in |
| Identification |
|
| A user provides his or her
bio-information only to the system. Next, the system matches
the information with a Template with the highest similarity,
and finds the enrollment information of that user. This is the
1:N matching process called identification or
recognition. |
|
| It is being used for AFIS
(Automatic Fingerprint Identification System) to find a
criminal's identity through fingerprints on the crime scene.
Or a camera uses the face recognition technology to find a
wanted man in a crowd. |
|
 |
|
The most striking difference
between the biometrics system and a traditional authentication
system based on an ID/password is that the new system cannot
generate 100 percent 'Yes' or 'No' answers. On the other hand,
other existing systems can do so according to letters or
numbers entered. In the case of biometrics, biological
information could change in terms of its shape or angle when
it is read, so matching scores against Templates could change
accordingly.
As a result, even a valid person may be
rejected or a wrong person may be accepted. The ratios
developed to evaluate the probabilities of the two cases are
called FRR (False Rejection Ratio) and FAR (False Acceptance
Ratio). |
|
|
| Figure 1.3 Concept Map on FRR
& FAR |
|
| In Figure 1.3 shown above, the
shaded portion with oblique lines on the left part of a
critical value is FRR, and the other portion on the right
part, is FAR. Certainly, a system with less shaded portion is
better. |
|
|
| Figure 1.4 Changes in FRR and
FAR according to changes in the critical value |
|
The above Figure 1.4 shows the
change patterns of FAR and FRR according to changes in the
critical value. The point where FAR and FRR are the same is
called EER (Equal Error Rate), and is used as a yardstick for
evaluating the system's performance along with the two other
ratios.
The smaller the shaded portion in Figure 1.3 is,
the closer the graph in Figure 1.4 will be to the X and Y
axis. A user can set a different critical value according to
his or her purposes. For example, in the case of a very
important security system, the user can reduce the possibility
that an unauthorized person is accepted to near zero by
setting a critical value high enough. Instead, even an
authorized person can fail to get access at times.
On the
other hand, if police is looking for a criminal with
fingerprints obtained from the crime scene, it will be
necessary to search for and analyze all possible fingerprints.
In areas such as these, it is useful to set the critical value
low enough to find all possible matches in the fingerprint
database. In short, the higher the critical value is, the less
convenient the system can be but, at the same time, its
security is higher. Conversely, the lower the value is, the
more convenient the system is, but its security is lax.
Therefore, it is very important for a user to set an
appropriate critical value according to needs or areas of
application. |
|
