SlideShare ist ein Scribd-Unternehmen logo
1 von 147
Downloaden Sie, um offline zu lesen
CC-9013 Biometrics
Dr. C. Saravanan
NIT Durgapur
dr.cs1973@gmail.com
Introduction
• Biometrics are automated methods of
recognizing a person based on a
• Physiological or
• Behavioral characteristic.
• The features measured are face,
fingerprints, hand geometry, handwriting,
iris, retinal, vein, and voice.
Dr. C. Saravanan, NIT Durgapur,
India
Fingerprint
Dr. C. Saravanan, NIT Durgapur,
India
Iris and Retina
Dr. C. Saravanan, NIT Durgapur,
India
Hand Geometry
Dr. C. Saravanan, NIT Durgapur,
India
Handwriting
Dr. C. Saravanan, NIT Durgapur,
India
Facial Recognition
Dr. C. Saravanan, NIT Durgapur,
India
Vein
Dr. C. Saravanan, NIT Durgapur,
India
Need
• Security breaches and transaction fraud
increases.
• Biometric technologies are becoming the
foundation of an extensive array of highly
secure identification and personal
verification solutions.
• Confidential financial transactions and
Personal data privacy.
Dr. C. Saravanan, NIT Durgapur,
India
Applications
• Biometric-based authentication
applications include
• workstation, network, and domain access,
• single sign-on, application logon,
• data protection,
• remote access to resources,
• transaction security and Web security.
Dr. C. Saravanan, NIT Durgapur,
India
Basic Image Operations
• Enhancement
• Filter
• Edge Detection
• Localisation
• Smoothning
• Sharpning
• Thresholding
Dr. C. Saravanan, NIT Durgapur,
India
Enhancement
• A process of enhancing the visual quality
of images due to nonideal image
acquisition process (e.g., poor illumination,
coarse quantization etc.)
• No reference (original) image is available
for comparison
• Human vision system (HVS) is the
JUDGE.
Dr. C. Saravanan, NIT Durgapur,
India
Enhanced Image
Dr. C. Saravanan, NIT Durgapur,
India
Technique Types
• Point operations
• Histogram Equalization
• Unsharp masking
• Homomorphic filtering
Dr. C. Saravanan, NIT Durgapur,
India
Point operations
• Point operations are zero-memory
operations where a given gray level
x∈[0,L] is mapped to another gray level
y∈[0,L] according to a transformation.
• Based only on the intensity of single
pixels.
• Linear
• Applications - Contrast Enhancement /
Feature Enhancement
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Important point operations
• Image negatives
• Contrast stretching
• Gray-level slicing
• Bit-plane slicing
Dr. C. Saravanan, NIT Durgapur,
India
Image negatives
• The transformation is very simple :
s=T(r)
T(r)=(L-1)-r
L is the number of gray levels.
The result of this transformation is that low
intensities are made high and vice versa.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Contrast stretching
• Contrast stretching (often called
normalization) attempts to improve the
contrast in an image by “stretching” the
range of intensity values to a desired
range of values.
Pout=(Pin-c)((b-a)/(d-c)) + a
8-bit graylevel images the lower and upper
limits might be 0 and 255, a & b, lowest
and highest pixel values currently present
in the image c & d.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Gray-level slicing
• Give a high value for all the gray-levels in
the specified range and a very low value
for all the other gray-levels.
Bit-plane slicing
• The intensity of each pixel of an image is
defined by several bits - highest order bits
are dominant.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Histogram Equalisation
• A histogram with a small spread has low
contrast
• A histogram with a wide spread has high
contrast
• An image with its histogram clustered at
the low end of the range corresponds to a
dark image
Dr. C. Saravanan, NIT Durgapur,
India
Histogram Equalisation steps
1. Find the running sum of the histogram values
2. Normalize the values from step 1 by dividing by
the total number of pixels
3. Multiply the values from step 2 by the maximum
gray level value and round to the closest integer
4. Map the gray level values to the results from
step 3 using a one-to-one correspondence
Dr. C. Saravanan, NIT Durgapur,
India
Histogram Example
Step 1:
Gray level Number of pixels Running Sum
value (Histogram values)
0 10 10
1 8 10+8=18
2 9 10+8+9=27
3 2 29
4 14 43
5 1 44
6 5 49
7 2 51
Dr. C. Saravanan, NIT Durgapur,
India
Example continues ...
Step 2: Normalizing by dividing by the total
number of pixels (51) we get 10/51, 18/51,
27/51, 29/51, 43/51, 44/51, 49/51, 51/51
Step 3: Multiply by the maximum gray level
value (7) and round we obtain 1, 2, 4, 4, 6,
6, 7, 7
Step 4: Map the original value to the results
from step 3
Dr. C. Saravanan, NIT Durgapur,
India
Histogram equalisation
Dr. C. Saravanan, NIT Durgapur,
India
Unsharp masking
• Combine histogram modification and
filtering operations.
Input Image -> filter -> Histogram
modification -> output image
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Homomorphic filter
• Simultaneously normalizes the brightness
across an image and increases contrast.
• To make the illumination of an image more even,
the high-frequency components are increased
and low-frequency components are decreased,
because the high-frequency components are
assumed to represent mostly the reflectance in
the scene (the amount of light reflected off the
object in the scene), whereas the low-frequency
components are assumed to represent mostly
the illumination in the scene.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Frequency Domain Methods
• Compute Fourier Transform of the image
to be enhanced.
• Multiply the result by a filter
• Take the inverse transform to produce the
enhanced image
• Will diminish camera noise, spurious pixel
values, missing pixel values etc.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Neighbourhood Averaging
• Smooth Image F(x,y) = Average pixel
value in a neighbourhood of I(x,y)
• For example, 3 x 3 neighbourhood
• Each pixel value is multiplied by 1/9
• Sum of 9 pixel value is the output
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Edge Preserving
• Also called Median Filtering
• Median of the neighbourhood pixel values
• More like neighbours
• Edges are preserved
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Image Sharpening
• Human perception is highly sensitive to
edges and fine details of an image, and
since they are composed primarily by high
frequency components, the visual quality
of an image can be enormously degraded
if the high frequencies are attenuated or
completed removed.
Dr. C. Saravanan, NIT Durgapur,
India
Sharpening
• In contrast, enhancing the high-frequency
components of an image leads to an
improvement in the visual quality.
• Image sharpening is widely used in
printing and photographic industries for
increasing the local contrast and
sharpening the images.
• Enhance detail that has been blurred.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Edge Detection
• Identifying and Locating sharp
discontinuities in an image.
• Discontinuities are abrupt changes in pixel
intensity.
• Sobel operator / filter
• Canny edge operator
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Sobel Operator
• 3 x 3 convolution kernels convolved with
original image.
• * denotes convolution operation
AGx *
101
202
101













 AGy *
121
000
121













Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Canny Edge Operator
1. Smoothing: Blurring of the image to remove noise.
2. Finding gradients: The edges should be marked where
the gradients of the image has large magnitudes.
3. Non-maximum suppression: Only local maxima should
be marked as edges.
4. Thresholding: Potential edges are determined by
thresholding.
5. Edge tracking by hysteresis: Final edges are
determined by suppressing all edges that are not
connected to a very certain (strong) edge.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Smoothing

















24542
491294
51215125
491294
24542
159
1
B
The image is first smoothed by applying a Gaussian filter.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Finding gradients
• Finds edges where the grayscale intensity
of the image changes the most.
• These areas are found by determining
gradients of the image
• Gradients at each pixel in the smoothed
image are determined by applying Sobel-
operator
Dr. C. Saravanan, NIT Durgapur,
India
Non-maximum suppression
• Convert the blurred edges in the image of the
gradient magnitudes to “sharp” edges.
• This is done by preserving all local maxima in
the gradient image, and deleting everything else.
1343
3465
6754
4532
Dr. C. Saravanan, NIT Durgapur,
India
Thresholding
• The edge-pixels remaining will probably be
true edges in the image.
• But some may be caused by noise or color
variations.
• Discern between these would be to use a
threshold.
• The Canny edge detection algorithm uses
double thresholding, Strong and Weak.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Edge Tracking by Hysteresis
• Strong edges are included in the final
edge image.
• Weak edges are included if and only if
they are connected to strong edges.
Dr. C. Saravanan, NIT Durgapur,
India
Biometric Identification
• Biometric identification compares a
biometric "signature" to all the records
stored in a database to determine if there
is a match (1 : N).
• Because it requires comparing each
existing record in the database against the
new biometric characteristic, it can be slow
and is usually not suitable for real-time
applications such as access control or
time and attendance.
Dr. C. Saravanan, NIT Durgapur,
India
Biometric Identification
• Biometric identification used most
frequently in such applications as law
enforcement — for instance, the
comparison of a fingerprint from a crime
scene to a database of prints collected
from convicted criminals.
Dr. C. Saravanan, NIT Durgapur,
India
Biometric Verification
• Biometric verification compares a newly-
scanned biometric characteristic to a
measurement previously collected from
that same person to verify that individual's
identity (1 : 1).
• For instance, when an employee is hired,
that employee's fingerprint will be enrolled
into the company's biometric time and
attendance system.
Dr. C. Saravanan, NIT Durgapur,
India
FAR
False Acceptance Rate (FAR)
• is the measure of the likelihood that the
biometric security system will incorrectly
accept an access attempt by an
unauthorized user.
• is stated as the ratio of the number of false
acceptances divided by the number of
identification attempts.
Dr. C. Saravanan, NIT Durgapur,
India
FRR
False Recognition Rate (FRR)
• is the measure of the likelihood that the
biometric security system will incorrectly
reject an access attempt by an authorized
user.
• is stated as the ratio of the number of false
rejections divided by the number of
identification attempts.
Dr. C. Saravanan, NIT Durgapur,
India
Positive / Negative Identification
• False Non-Match Rate (FNMR)
• False Match Rate (FMR)
• False Reject Rate (FRR)
• False Accept Rate (FAR)
• False Positive Identification Rate (FPIR)
• False Negative Identification Rate (FNIR)
Dr. C. Saravanan, NIT Durgapur,
India
Matching Score Distribution
• A matching algorithm was defined as an
algorithm that make a decision about genuine or
impostor nature of a comparison between two
templates.
• In the first step an Evaluation Algorithm assigns
a similarity score to the comparison. That
similarity score is a value on the range (0..1) and
as higher be the score value more similar the
images.
• The second step decides if the comparison is
genuine or impostor using a frontier threshold or
decision threshold (DT).
Dr. C. Saravanan, NIT Durgapur,
India
Receiver Operating Characteristic
(ROC)
• accepted method for summarizing the
performance of imperfect pattern matching
systems.
• parametrically as a function of the decision
threshold,
• the rate of “false positives” on the x-axis
• the rate of “true positives” on the y-axis
• ROC curves are threshold independent
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Detection Error Trade-off (DET)
• modified ROC curve
• plots error rates on both axes
• giving uniform treatment to both types of
error
• distinguishes different wellperforming
systems more clearly
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Expected Overall Error
gi TFNMRTFMRTE   )()()(
• Expected overall error takes into account the
possibility of different FM and FNM and is given
as
where,
T=threshold
pi=probability of a random user being an imposter.
pg=probability of a random user being genuine.
Dr. C. Saravanan, NIT Durgapur,
India
Equal Error Rate (EER)
• EER is the value where FMR and FNMR
are equal.
• Lower be the EER the lower error rate of
the algorithm.
• Select the EER score value as Decision
Threshold (DT) is frequently a good
decision for a regular biometric
application.
Dr. C. Saravanan, NIT Durgapur,
India
Myths & Misrepresentations
• Biometric “X” is best for all applications
• Biometric “X” is unique for each individual
• A single number quantifies system
accuracy
• System is “plug and play”
• Real accuracy performance can be
predicted
Dr. C. Saravanan, NIT Durgapur,
India
• The vendors reporting best FAR and FRR
has the “most accurate system”
• Multiple biometrics outperform single
biometrics
• “Our biometric system does not use a
decision threshold”
• “Our feature extractor can be used with
any match engine”
• Large templates mean better accuracy
• Face recognition prevents terrorism
Dr. C. Saravanan, NIT Durgapur,
India
• Biometrics means 100 percent security
• Biometric systems invade our privacy
• Biometric sensors are unhygienic or
otherwise harmful
Dr. C. Saravanan, NIT Durgapur,
India
Biometric System Design Issues
• System Architecture
– Centralised / Distributed Server
– Client Computer
– Device at User End
• Hardware & Software Implementation
– Sample Acquitision
– User Interface
– Biometric Processing Components
Dr. C. Saravanan, NIT Durgapur,
India
BSD Issues (continues...)
– Communication Channels
– Database Design
– Interoperability
• Administration Policy
– Integrity of Enrolment
– Quality of Enrolment Samples
– System Configuration
– Exception Handling
– Privacy Measures
Dr. C. Saravanan, NIT Durgapur,
India
Biometric System Security
• Includes IT Security
• Earlier related financial
• At present, Passports, Visas, etc.
– Biometric Security Evaluation
– Biometric Transaction Security
– Protection of Biometric Data
Dr. C. Saravanan, NIT Durgapur,
India
Authentication Protocol
• is a type of cryptographic protocol with the purpose of
authenticating entities wishing to communicate securely.
• AKA, CAVE-based_authentication, Challenge-
handshake authentication protocol (CHAP), CRAM-MD5
Diameter, Digest, Extensible Authentication Protocol
(EAP), Host Identity Protocol (HIP), Kerberos, MS-CHAP,
LAN Manager, NTLM, Password-authenticated key
agreement protocols, Password Authentication Protocol
(PAP), Protected Extensible Authentication Protocol
(PEAP), Protocol for Carrying Authentication for Network
Access (PANA), RADIUS, Secure Remote Password
protocol (SRP), TACACS and TACACS+, RFID-
Authentication Protocols, Woo Lam 92 (protocol)
Dr. C. Saravanan, NIT Durgapur,
India
Kerberos
• MIT developed Kerberos
• Kerberos (or Cerberus - Greek monstrous
three-headed guard dog)
• Works on the basis of 'tickets' to allow
nodes communicating over a non-secure
network to prove their identity to one
another in a secure manner.
• Protects against eavesdropping and replay
attacks.
Dr. C. Saravanan, NIT Durgapur,
India
Selection of Suitable Biometric
• Ethnic Background
– who will be the users of the biometric system
– education levels and a variety of attitudes
• Employee Education
– technical background of the users.
– higher the technological background less
training required
Dr. C. Saravanan, NIT Durgapur,
India
• Frequency of Use
– some biometric systems are more suitable for
high frequency of usage
• User characteristics
– will the users be in a hurry and possibly be a
little bit impatient (public restroom)
Dr. C. Saravanan, NIT Durgapur,
India
Biometric Attributes
1. Universal:
Every person must possess the characteristic / attribute.
The attribute must be one that is universal and seldom lost to
accident or disease.
2. Invariance of properties:
They should be constant over a long period of time.
The attribute should not be subject to significant differences based
on age either episodic or chronic disease.
3. Measurability:
The properties should be suitable for capture without waiting time
and must be easy to gather the attribute data passively.
4. Singularity:
Each expression of the attribute must be unique to the individual.
The characteristics should have sufficient unique properties to
distinguish one person from any other.
Dr. C. Saravanan, NIT Durgapur,
India
Height, weight, hair and eye color are all attributes that
are unique assuming a particularly precise measure, but
do not offer enough points of differentiation to be useful
for more than categorizing.
5. Acceptance:
The capturing should be possible in a way acceptable to
a large percentage of the population. Excluded are
particularly invasive technologies, i.e. technologies which
require a part of the human body to be taken or which
impair the human body.
6. Reducibility:
The captured data should be capable of being reduced
to a file which is easy to handle.
7. Reliability and tamper-resistance:
The attribute should be impractical to mask or
manipulate. The process should ensure high reliability
and reproducibility.
Dr. C. Saravanan, NIT Durgapur,
India
8. Privacy: The process should not violate
the privacy of the person.
9. Comparable: Should be able to reduce
the attribute to a state that makes it
digitally comparable to others. The less
probabilistic the matching involved, the
more authoritative the identification.
10. Inimitable: The attribute must be
irreproducible by other means. The less
reproducible the attribute, the more likely it
will be authoritative.
Dr. C. Saravanan, NIT Durgapur,
India
Zephyr Chart
Dr. C. Saravanan, NIT Durgapur,
India
Zephyr Chart is a navigation aid for pilots designed to display
aviation charts in a format easy to use in the cockpit.
Multi Biometrics (MB)
• MB - Reconciliation of Evidence enhances
the recognition accuracy of human
authentication systems.
• Multibiometric systems combine the
information presented by multiple
biometric sensors, algorithms, samples,
units, or traits.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
MB
• Besides enhancing matching performance,
these systems are expected to
– improve population coverage,
– deter spoofing and
– impart fault tolerance to biometric applications
• Establishes an identity based on who you
are,
– rather than by what you possess, such as an ID card,
– or what you remember, such as a password.
Dr. C. Saravanan, NIT Durgapur,
India
Risks with Single Biometric Trait
• Noise in sensed data
• Non-universality
• Upper bound on identification accuracy
• Spoof attacks (voice and signature -
mimic)
• fingerprints can also be spoofed by
inscribing ridge-like structures on synthetic
material such as gelatine and play-doh.
Dr. C. Saravanan, NIT Durgapur,
India
Advantages of MB
• If a biometric trait (eg. fingerprint) fails
other trait (eg. voice) aid to authenticate
the person.
• Facilitates the filtering or indexing of
largescale biometric databases. Among
the traits, the simple one may be used for
indexing.
Dr. C. Saravanan, NIT Durgapur,
India
• Difficulties has been increased the
spoofing effort. Randomly traits can be
acquired at the time of authentication.
• In case of noise with one traint other trait
could help in authentication.
• Certain traits may not be applicable all the
time, the other traits may be used.
• Highly fault tolerant, because even a trait
fails other traits are available.
Dr. C. Saravanan, NIT Durgapur,
India
MB Classifications
• multi-sensor, multi-algorithm,
• multi -instance (left and right index fingers
and left and right irises),
• multi-sample (pose of a face - front, left,
right),
• multi-modal (face and voice) and hybrid.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
MB Fusion
• Fusion before matching
• Fusion after matching
• Sensor level - multiple sensors - face.
• Feature level - multi features are
consolidated into single feature.
• Score level - generating a single score.
• Rank level - ranking of enrolled identities
• Decision level - access only to the final
recognition decision.
Dr. C. Saravanan, NIT Durgapur,
India
Fusion
• Combining different sources of information
for intelligent systems.
• The information are signals delivered by
different sensors and images from various
modalities.
• The fusion concepts and methods gather
tools like weighted average, neural
networks, sub-band filtering, and rules
based knowledge.
Dr. C. Saravanan, NIT Durgapur,
India
Fusion Categories
• Multiview - same modality and time -
different places and conditions.
• Multimodal - CT, MRI
• Multitemporal - same scene - different time
• Multifocus - regions based
Dr. C. Saravanan, NIT Durgapur,
India
Types of Multisensor Fusion
• Signal-level fusion
• Image-level fusion
• Feature-level fusion
• Symbol-level fusion
Dr. C. Saravanan, NIT Durgapur,
India
Fusion Applications
• Intelligent robots
– visual, tactile, force/torque, and other types of sensors
– Stereo camera fusion
• Medical image
– Fusing X-ray computed tomography (CT) and magnetic
resonance (MR) images
• Manufacturing
– Complex machine/device diagnostics
• Military and law enforcement
– Detection, tracking, identification of ocean (air, ground)
target / event
• Remote sensing
Image Fusion
• Improves reliability (redundant information)
• Improves capability (complementary info.)
Dr. C. Saravanan, NIT Durgapur,
India
IF Techniques
• Pixel values are manipulated to achieve
desired result (SPATIAL Domain).
• Fusion operations are performed on
frequency (TRANSFORM domain).
• Primitive fusion (Averaging Method, Select
Maximum, and Select Minimum),
• Discrete Wavelet transform based fusion,
• Principal component analysis (PCA) based
fusion.
Dr. C. Saravanan, NIT Durgapur,
India
• The fusion methods such as averaging, Brovey
method, principal component analysis (PCA)
and Intensity-Hue-Saturation (IHS) based
methods fall under spatial domain approaches.
• Another important spatial domain fusion method
is the high pass filtering based technique.
• The disadvantage of spatial domain approaches
is that they produce spatial distortion in the
fused image.
• Spatial distortion becomes a negative factor
while we go for further processing such as
classification problem
Dr. C. Saravanan, NIT Durgapur,
India
IF Methods
• Intensity-hue-saturation (IHS) transform based fusion
• Principal component analysis (PCA) based fusion
• Multi scale transform based fusion:-
– High-pass filtering method
– Pyramid method:-(i) Gaussian pyramid (ii) Laplacian
• Pyramid (iii) Gradient pyramid (iv) Morphological pyramid
– Ratio of low pass pyramid
– Wavelet transforms:- (i) Discrete wavelet transforms (DWT)
(ii) Stationary wavelet transforms (iii) Multiwavelet
transforms
– Curvelet transforms
Dr. C. Saravanan, NIT Durgapur,
India
Simple Average
• K (i, j) = {X (i, j) + Y (i, j)}/2
Where X (i , j) and Y ( i, j) are two input
Images
regions of images that are in focus tend to
be of higher pixel intensity
Dr. C. Saravanan, NIT Durgapur,
India
Select Maximum
• The value of the pixel P (i, j) of each image
is taken and compared to each other.
• The greatest pixel value is assigned to the
corresponding pixel of the output image.
• This algorithm chooses the in-focus
regions from each input image by
choosing the greatest value for each pixel,
resulting in highly focused output.
Dr. C. Saravanan, NIT Durgapur,
India
DWT
• Wavelets can be described by using two
functions viz.
• the scaling function f (t), also known as
“father wavelet” and
• the wavelet function or “mother wavelet”.
Mother wavelet (t)
• undergoes translation and scaling operations to
give self similar wavelet families.
Dr. C. Saravanan, NIT Durgapur,
India
• The wavelet transform decomposes the
image into low-high, high-low, high-high
spatial frequency bands at different scales
and the low-low band at the coarsest scale
• The L-L band contains the average image
information whereas the other bands
contain directional information due to
spatial orientation.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Higher absolute values of wavelet coefficients in the
high bands correspond to salient features such as
edges or lines.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
PRINCIPAL COMPONENT
ANALYSIS (PCA)
• Statistical Procedure
• Transforms a number of correlated variables into
a number of linearly uncorrelated variables
called principal components.
• The first principal component accounts for as
much of the variance in the data as possible and
• Each succeeding component accounts for as
much of the remaining variance as possible.
Dr. C. Saravanan, NIT Durgapur,
India
Different Names of PCA
• discrete Karhunen–Loève transform (KLT) in signal
processing,
• the Hotelling transform in multivariate quality control,
• Proper Orthogonal Decomposition (POD) in mechanical
engineering,
• Singular Value Decomposition (SVD) of X (Golub and
Van Loan, 1983),
• Eigenvalue Decomposition (EVD) of XTX in linear
algebra,
• Factor Analysis (for a discussion of the differences
between PCA and factor analysis,
Dr. C. Saravanan, NIT Durgapur,
India
• Eckart–Young theorem (Harman, 1960), or
Schmidt–Mirsky theorem in sychometrics,
• Empirical Orthogonal Functions (EOF) in
meteorological science,
• Empirical Eigenfunction decomposition (Sirovich,
1987),
• Empirical Component Analysis (Lorenz, 1956),
• Quasiharmonic modes (Brooks et al., 1988),
• Spectral Decomposition in noise and vibration,
and empirical modal analysis in structural
dynamics.
Dr. C. Saravanan, NIT Durgapur,
India
Multimodal Biometrics
• Multimodal Biometrics are systems that are
capable of using more than one physiological or
behavioural characteristic for enrolment,
verification or identification.
• will use a combination of the following
recognition technologies, up to three of them, to
compare the identity of a person.
• If one of the technologies fails for any reason,
your system can still use another one or two.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Forms of Biometric Identification
• Some forms of behavioural biometric
identification include the following:
– Keystroke or Typing Recognition
– Speaker identification or Recognition
• Some forms of physical biometric
identification include the following:
– Fingerprint – Iris
– Retina – Finger Geometry
– Signature/Handwriting – Voice
– Facial Proportions – Hand Geometry
Dr. C. Saravanan, NIT Durgapur,
India
Benefits
• The benefits of multimodal biometrics is
that by using more than one means of
identification, your system can retain a
high threshold recognition setting and your
system administrator can decide the level
of security that is needed.
Dr. C. Saravanan, NIT Durgapur,
India
Need
• There is a great need for multimodal biometrics as most
biometric systems used in real applications are
unimodal, which means they rely on only one area of
identification.
• Some examples of these are fingerprints, faces and
voices and these systems are quite vulnerable to many
problems such as noisy data, non-universality and
spoofing.
• This leads to a high false acceptance rate and false
rejection rate, limited discrimination capability, and lack
of permanence.
Dr. C. Saravanan, NIT Durgapur,
India
Applications
• Strong Potential:
– Physical access
– Civil ID
– Criminal ID
• Moderate Potential:
– Network/PC access
– Kiosk/ATM
• Modest Potential:
– Retail/POS
– Surveillance
– eCommerce
– telephony
Dr. C. Saravanan, NIT Durgapur,
India
Reliability
• Multimodal systems are more reliable because
using many independent biometrics that meet
very high performance requirements and they
counteract the problems listed above.
• They also effectively deter spoofing because it is
near impossible to spoof multiple biometric traits
and the system can request the user to present
random traits that only a live person can do.
Dr. C. Saravanan, NIT Durgapur,
India
Biometric System Security
• Most biometric systems will advertise a False
Accept Rate (FAR) and False Reject Rate (FRR)
to characterize the security provided by the
system.
• The FAR tells you how often someone will be
recognized successfully when he/she should not
have been recognized, and the FRR tells you
how often someone who should have been
recognized successfully is not recognized.
Dr. C. Saravanan, NIT Durgapur,
India
• The FAR is an upper bound on security.
• The FAR assumes random and real
samples of the biometric characteristic are
presented to the system.
• Attackers don’t have to use random or real
samples; they are much more likely to pick
samples intelligently and achieve a much
higher FAR.
Dr. C. Saravanan, NIT Durgapur,
India
• The attacker can do this in two ways.
• First the attacker can use a copy of the biometric
characteristic of a valid user collected by the system to
produce a fake biometric characteristic that will allow
access.
• This attack is called a physical spoof attack and will
allow the attacker to bypass the biometric system.
• This attack is like finding a password written down or
watching a valid user enter a password, copying it, and
using it to gain access to the system.
Dr. C. Saravanan, NIT Durgapur,
India
• Second, if the attacker does not have a copy of the
biometric characteristic of a valid user, the attacker can
attempt to create a fake biometric characteristic by
guessing “intelligently” or the attacker can use a
database of real samples of invalid users “intelligently”
instead of randomly.
• As the guesses get better, the attacker will likely bypass
the biometric system in fewer attempts than presenting
random samples.
Dr. C. Saravanan, NIT Durgapur,
India
• This is loosely comparable to an attacker
using a dictionary to attack a password
based system rather than randomly going
through all possible passwords of the
proper lengths.
Dr. C. Saravanan, NIT Durgapur,
India
Liveness
• For biometric systems, liveness detection
makes both attacks more difficult.
• However, while liveness detection is
improving in biometric systems overall,
many systems employing liveness
detection are still susceptible to physical
spoof attacks.
Dr. C. Saravanan, NIT Durgapur,
India
• Rather than attacking using samples of the
biometric characteristic, other attacks on a
biometric system may be possible, such
as cryptographic attacks, network attacks,
operating system attacks, etc.
• All these potential vulnerabilities must be
considered when implementing a biometric
system with the intent of enhancing
system security.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Eight Types of Attacks
• Type 1 attack involves presenting a fake
Biometric
– (e.g., finger made from silicon, face mask, lens
including fake iris texture) to the sensor.
• The second type of attack is called a replay
attack,
– because an intercepted biometric (with or without
the cooperation of the genuine user) data is
submitted to the feature extractor, bypassing the
sensor.
Dr. C. Saravanan, NIT Durgapur,
India
• In the third type of attack, the feature
extractor module is replaced with a Trojan
horse program
– that functions according to its designer's
specifications
Dr. C. Saravanan, NIT Durgapur,
India
• In the fourth type of attack, genuine
feature values are replaced with values
(synthetic or real) selected by the attacker.
• In the fifth type of attack, the matcher is
replaced with a Trojan horse program.
• The attacks on the template database
(e.g., addition, modification, or removal of
templates) constitute the sixth type of
attack.
Dr. C. Saravanan, NIT Durgapur,
India
• In the seventh type of attack, the
templates are tampered (stolen, replaced,
or altered)
– in the transmission medium between the
template database and matcher.
• Lastly, the matcher result (accept or reject)
can be overridden by the attacker.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Attacks depend on
(i) biometric modality (e.g., it is harder to physically
replicate a retina scan than it is to forge a
signature),
(ii) type of sensor (2D vs. 3D face sensors),
(iii) type of matcher operating on the same
biometric (face matchers based on texture
vs. geometry), and
(iv) the security settings (reflected via False Accept
Rate) of the biometric system.
Dr. C. Saravanan, NIT Durgapur,
India
Dr. C. Saravanan, NIT Durgapur,
India
Face Synthesis
Dr. C. Saravanan, NIT Durgapur,
India
Fingerprint Watermarked in
Face
Dr. C. Saravanan, NIT Durgapur,
India
Secure Sytems Properties
(i) accuracy
(ii) rapid
(iii) reliable
(iv) protecting privacy rights
(v) cost effectively
(vi) user-friendly
(vii) minimal changes to the existing
infrastructures
Dr. C. Saravanan, NIT Durgapur,
India
Liveness Detection Methods
• Fingerprints
– periodicity of sweat and the sweat diffusion
pattern were used to detect fake fingerprints.
– liveness detection method based on skin
elasticity properties
– The user is asked to deliberately rotate his finger
when removing it from the sensor surface thus
producing considerable skin distortion which is
later used as a fingerprint liveness measure.
Dr. C. Saravanan, NIT Durgapur,
India
– An odour sensor (electronic nose) is used to discriminate
the skin odour from that of other materials such as latex,
silicone or gelatine.
• Iris
– the spectrographic properties of different parts of the eye
(tissue, fat, blood, melanin pigment), the coaxial retinal
back reflection (the red eye effect) and
– the four Purkinje reflections caused by each of the four
optical surfaces comprised inside the eye.
– very high quality cameras required.
Dr. C. Saravanan, NIT Durgapur,
India
– permanent oscillation that the eye pupil
presents even under uniform lighting
conditions
– measure the pupil response to a sudden
lighting event
– infrared light reflections from the moist
cornea
Dr. C. Saravanan, NIT Durgapur,
India
Vulnerabilities of Biometric
Systems
• Administration: Intentional or unintentional
administrative mistakes.
• User: A legitimate user wants to upgrade
his privileges to the administrative level.
• Enrolment: Breaking registration
procedures.
• Spoofing: A fake biometric is used for
authentication as a legitimate user.
Dr. C. Saravanan, NIT Durgapur,
India
• Mimicry: Attacker mimics the biometric characteristics of
the legitimate user.
• Undetect: Attacks undetected by the system may
encourage new attacks.
• Fail secure: Result of abnormal utilization conditions of
biometric system or IT environment.
• Bypass: Bypassing biometric system for access. This can
be achieved by surpassing physical barriers, forcing a
legitimate user to present his biometric to the sensor, or by
cooperation of legitimate user.
Dr. C. Saravanan, NIT Durgapur,
India
• Corrupt attack – Weakening the system by
making changes in the IT environment or
biometric system. Modification or
replacement of system parameters is an
example.
• Degrade: Certain software in the IT
environment decreases the system’s
security level.
Dr. C. Saravanan, NIT Durgapur,
India
• Tamper: Counterfeiting the hardware of
the system.
• Residual: Latent fingerprints may be used
to make artificial fingerprints or accepted
directly by the sensor.
Dr. C. Saravanan, NIT Durgapur,
India
• Cryptological attack: Encryption can be broken in data
transmission and this biometric data can be used for
another type of attack (e.g. replay attack).
• Brute force attack: Attacker presents the biometric
characteristic to the system repeatedly in order to be
authenticated.
• This type of attack depends on FAR parameter.
• Evil twin attacks: Biometric characteristic of imposter is
very similar to the enrolled user’s biometric.
Dr. C. Saravanan, NIT Durgapur,
India
• Fake template: Introducing fake biometric
template into the database or onto smart cards.
• Noise: Access can be gained by the attacker
when noise is applied to the system.
• Poor image: Quality supervision may be utilized.
If low quality images are accepted for
registration then attacker may hope to deceive
the system as in the case of noisy images.
• Weak ID: Similar to “poor image” weakness, and
tries to fake the system by weak templates.
Dr. C. Saravanan, NIT Durgapur,
India
Any Questions ?
Dr. C. Saravanan, NIT Durgapur,
India
References
 Biometrics: Identity Assurance in the Information Age, John D.
Woodward Jr.
 Biometrics: Advanced Identity Verification: The Complete Guide,
Julian Ashbourn
 Biometrics: Identity Verification in a Networked World, Samir Nanavati
 Digital Image Processing, Gonzalez and Woods
 Wikipedia
Dr. C. Saravanan, NIT Durgapur,
India

Weitere ähnliche Inhalte

Was ist angesagt?

7 Forensic Science Powerpoint Chapter 07 Forensic Anthropology
7  Forensic Science Powerpoint Chapter 07 Forensic Anthropology7  Forensic Science Powerpoint Chapter 07 Forensic Anthropology
7 Forensic Science Powerpoint Chapter 07 Forensic Anthropology
Grossmont College
 

Was ist angesagt? (20)

Fingerprint presentation
Fingerprint presentationFingerprint presentation
Fingerprint presentation
 
Biometric
BiometricBiometric
Biometric
 
Fingerprints
FingerprintsFingerprints
Fingerprints
 
Pattern recognition Hand Geometry
Pattern recognition Hand GeometryPattern recognition Hand Geometry
Pattern recognition Hand Geometry
 
Portrait Parle via Bertillon System By G S Shaktawat
Portrait Parle via Bertillon System By G S ShaktawatPortrait Parle via Bertillon System By G S Shaktawat
Portrait Parle via Bertillon System By G S Shaktawat
 
7 Forensic Science Powerpoint Chapter 07 Forensic Anthropology
7  Forensic Science Powerpoint Chapter 07 Forensic Anthropology7  Forensic Science Powerpoint Chapter 07 Forensic Anthropology
7 Forensic Science Powerpoint Chapter 07 Forensic Anthropology
 
Toolmarks
ToolmarksToolmarks
Toolmarks
 
Multi modal biometric system
Multi modal biometric systemMulti modal biometric system
Multi modal biometric system
 
Signature verification in biometrics
Signature verification in biometricsSignature verification in biometrics
Signature verification in biometrics
 
Biometrics
BiometricsBiometrics
Biometrics
 
Automated Fingerprint Identification Systems
Automated Fingerprint Identification SystemsAutomated Fingerprint Identification Systems
Automated Fingerprint Identification Systems
 
Forensic nanotechnology
Forensic nanotechnologyForensic nanotechnology
Forensic nanotechnology
 
Mobile Forensics
Mobile Forensics Mobile Forensics
Mobile Forensics
 
Biometrics overview ppt
Biometrics overview pptBiometrics overview ppt
Biometrics overview ppt
 
Forensic anthropology
Forensic anthropologyForensic anthropology
Forensic anthropology
 
Digital Forensic
Digital ForensicDigital Forensic
Digital Forensic
 
Forensic analysis of foot wear impression
Forensic analysis of foot wear impressionForensic analysis of foot wear impression
Forensic analysis of foot wear impression
 
Forensic serology
Forensic serologyForensic serology
Forensic serology
 
Examination of mechanical impression
Examination of mechanical impressionExamination of mechanical impression
Examination of mechanical impression
 
PALATAL RUGAE IN FORENSIC ODONTOLOGY
PALATAL RUGAE IN FORENSIC ODONTOLOGYPALATAL RUGAE IN FORENSIC ODONTOLOGY
PALATAL RUGAE IN FORENSIC ODONTOLOGY
 

Ähnlich wie Biometrics

12-Image enhancement and filtering.ppt
12-Image enhancement and filtering.ppt12-Image enhancement and filtering.ppt
12-Image enhancement and filtering.ppt
AJAYMALIK97
 

Ähnlich wie Biometrics (20)

Digital image processing - OLD
Digital image processing - OLDDigital image processing - OLD
Digital image processing - OLD
 
Computer vision
Computer visionComputer vision
Computer vision
 
Image enhancement
Image enhancementImage enhancement
Image enhancement
 
Digital Image Fundamentals - II
Digital Image Fundamentals - IIDigital Image Fundamentals - II
Digital Image Fundamentals - II
 
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
Rethinking Data Augmentation for Image Super-resolution: A Comprehensive Anal...
 
Digital image processing 2
Digital image processing   2Digital image processing   2
Digital image processing 2
 
denoising.pptx
denoising.pptxdenoising.pptx
denoising.pptx
 
Processing_of_Satellite_Image_using_Digi.pptx
Processing_of_Satellite_Image_using_Digi.pptxProcessing_of_Satellite_Image_using_Digi.pptx
Processing_of_Satellite_Image_using_Digi.pptx
 
Processing of satellite_image_using_digi
Processing of satellite_image_using_digiProcessing of satellite_image_using_digi
Processing of satellite_image_using_digi
 
Simultaneous Smoothing and Sharpening of Color Images
Simultaneous Smoothing and Sharpening of Color ImagesSimultaneous Smoothing and Sharpening of Color Images
Simultaneous Smoothing and Sharpening of Color Images
 
IRJET- Skin Cancer Detection using Local and Global Contrast Stretching
IRJET- Skin Cancer Detection using Local and Global Contrast StretchingIRJET- Skin Cancer Detection using Local and Global Contrast Stretching
IRJET- Skin Cancer Detection using Local and Global Contrast Stretching
 
Plastic surgeryinvariantfacedetection.pptx
Plastic surgeryinvariantfacedetection.pptxPlastic surgeryinvariantfacedetection.pptx
Plastic surgeryinvariantfacedetection.pptx
 
Study on Data Augmentation Methods for Sonar Image Analysis
Study on Data Augmentation Methods for Sonar Image AnalysisStudy on Data Augmentation Methods for Sonar Image Analysis
Study on Data Augmentation Methods for Sonar Image Analysis
 
Image Resolution Enhancement using DWT and Spatial Domain Interpolation Techn...
Image Resolution Enhancement using DWT and Spatial Domain Interpolation Techn...Image Resolution Enhancement using DWT and Spatial Domain Interpolation Techn...
Image Resolution Enhancement using DWT and Spatial Domain Interpolation Techn...
 
A Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement TechniquesA Review of Image Contrast Enhancement Techniques
A Review of Image Contrast Enhancement Techniques
 
Algorithm for the Comparison of Different Types of First Order Edge Detection...
Algorithm for the Comparison of Different Types of First Order Edge Detection...Algorithm for the Comparison of Different Types of First Order Edge Detection...
Algorithm for the Comparison of Different Types of First Order Edge Detection...
 
A010110104
A010110104A010110104
A010110104
 
12-Image enhancement and filtering.ppt
12-Image enhancement and filtering.ppt12-Image enhancement and filtering.ppt
12-Image enhancement and filtering.ppt
 
A Review Paper on Fingerprint Image Enhancement with Different Methods
A Review Paper on Fingerprint Image Enhancement with Different MethodsA Review Paper on Fingerprint Image Enhancement with Different Methods
A Review Paper on Fingerprint Image Enhancement with Different Methods
 
aip.pptx
aip.pptxaip.pptx
aip.pptx
 

Mehr von National Institute of Technology Durgapur (8)

Morphological operations
Morphological operationsMorphological operations
Morphological operations
 
Cs 331 Data Structures
Cs 331 Data StructuresCs 331 Data Structures
Cs 331 Data Structures
 
Prog lang-c
Prog lang-cProg lang-c
Prog lang-c
 
Computer and Information Technology
Computer and Information TechnologyComputer and Information Technology
Computer and Information Technology
 
Punctuations
PunctuationsPunctuations
Punctuations
 
Database design for HPC
Database design for HPCDatabase design for HPC
Database design for HPC
 
Human computer interaction
Human computer interactionHuman computer interaction
Human computer interaction
 
Natural language processing
Natural language processingNatural language processing
Natural language processing
 

Kürzlich hochgeladen

1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi  6.pdf1029-Danh muc Sach Giao Khoa khoi  6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
QucHHunhnh
 
1029 - Danh muc Sach Giao Khoa 10 . pdf
1029 -  Danh muc Sach Giao Khoa 10 . pdf1029 -  Danh muc Sach Giao Khoa 10 . pdf
1029 - Danh muc Sach Giao Khoa 10 . pdf
QucHHunhnh
 
The basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptxThe basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptx
heathfieldcps1
 
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
ZurliaSoop
 

Kürzlich hochgeladen (20)

SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptxSKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
 
Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)Accessible Digital Futures project (20/03/2024)
Accessible Digital Futures project (20/03/2024)
 
Understanding Accommodations and Modifications
Understanding  Accommodations and ModificationsUnderstanding  Accommodations and Modifications
Understanding Accommodations and Modifications
 
1029-Danh muc Sach Giao Khoa khoi 6.pdf
1029-Danh muc Sach Giao Khoa khoi  6.pdf1029-Danh muc Sach Giao Khoa khoi  6.pdf
1029-Danh muc Sach Giao Khoa khoi 6.pdf
 
How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17How to Give a Domain for a Field in Odoo 17
How to Give a Domain for a Field in Odoo 17
 
Introduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The BasicsIntroduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The Basics
 
Unit-V; Pricing (Pharma Marketing Management).pptx
Unit-V; Pricing (Pharma Marketing Management).pptxUnit-V; Pricing (Pharma Marketing Management).pptx
Unit-V; Pricing (Pharma Marketing Management).pptx
 
ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.ICT role in 21st century education and it's challenges.
ICT role in 21st century education and it's challenges.
 
FSB Advising Checklist - Orientation 2024
FSB Advising Checklist - Orientation 2024FSB Advising Checklist - Orientation 2024
FSB Advising Checklist - Orientation 2024
 
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
 
Kodo Millet PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
Kodo Millet  PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...Kodo Millet  PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
Kodo Millet PPT made by Ghanshyam bairwa college of Agriculture kumher bhara...
 
1029 - Danh muc Sach Giao Khoa 10 . pdf
1029 -  Danh muc Sach Giao Khoa 10 . pdf1029 -  Danh muc Sach Giao Khoa 10 . pdf
1029 - Danh muc Sach Giao Khoa 10 . pdf
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
 
Food safety_Challenges food safety laboratories_.pdf
Food safety_Challenges food safety laboratories_.pdfFood safety_Challenges food safety laboratories_.pdf
Food safety_Challenges food safety laboratories_.pdf
 
ComPTIA Overview | Comptia Security+ Book SY0-701
ComPTIA Overview | Comptia Security+ Book SY0-701ComPTIA Overview | Comptia Security+ Book SY0-701
ComPTIA Overview | Comptia Security+ Book SY0-701
 
Towards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptxTowards a code of practice for AI in AT.pptx
Towards a code of practice for AI in AT.pptx
 
The basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptxThe basics of sentences session 3pptx.pptx
The basics of sentences session 3pptx.pptx
 
Python Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docxPython Notes for mca i year students osmania university.docx
Python Notes for mca i year students osmania university.docx
 
Single or Multiple melodic lines structure
Single or Multiple melodic lines structureSingle or Multiple melodic lines structure
Single or Multiple melodic lines structure
 
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
 

Biometrics

  • 1. CC-9013 Biometrics Dr. C. Saravanan NIT Durgapur dr.cs1973@gmail.com
  • 2. Introduction • Biometrics are automated methods of recognizing a person based on a • Physiological or • Behavioral characteristic. • The features measured are face, fingerprints, hand geometry, handwriting, iris, retinal, vein, and voice. Dr. C. Saravanan, NIT Durgapur, India
  • 3. Fingerprint Dr. C. Saravanan, NIT Durgapur, India
  • 4. Iris and Retina Dr. C. Saravanan, NIT Durgapur, India
  • 5. Hand Geometry Dr. C. Saravanan, NIT Durgapur, India
  • 6. Handwriting Dr. C. Saravanan, NIT Durgapur, India
  • 7. Facial Recognition Dr. C. Saravanan, NIT Durgapur, India
  • 8. Vein Dr. C. Saravanan, NIT Durgapur, India
  • 9. Need • Security breaches and transaction fraud increases. • Biometric technologies are becoming the foundation of an extensive array of highly secure identification and personal verification solutions. • Confidential financial transactions and Personal data privacy. Dr. C. Saravanan, NIT Durgapur, India
  • 10. Applications • Biometric-based authentication applications include • workstation, network, and domain access, • single sign-on, application logon, • data protection, • remote access to resources, • transaction security and Web security. Dr. C. Saravanan, NIT Durgapur, India
  • 11. Basic Image Operations • Enhancement • Filter • Edge Detection • Localisation • Smoothning • Sharpning • Thresholding Dr. C. Saravanan, NIT Durgapur, India
  • 12. Enhancement • A process of enhancing the visual quality of images due to nonideal image acquisition process (e.g., poor illumination, coarse quantization etc.) • No reference (original) image is available for comparison • Human vision system (HVS) is the JUDGE. Dr. C. Saravanan, NIT Durgapur, India
  • 13. Enhanced Image Dr. C. Saravanan, NIT Durgapur, India
  • 14. Technique Types • Point operations • Histogram Equalization • Unsharp masking • Homomorphic filtering Dr. C. Saravanan, NIT Durgapur, India
  • 15. Point operations • Point operations are zero-memory operations where a given gray level x∈[0,L] is mapped to another gray level y∈[0,L] according to a transformation. • Based only on the intensity of single pixels. • Linear • Applications - Contrast Enhancement / Feature Enhancement Dr. C. Saravanan, NIT Durgapur, India
  • 16. Dr. C. Saravanan, NIT Durgapur, India
  • 17. Important point operations • Image negatives • Contrast stretching • Gray-level slicing • Bit-plane slicing Dr. C. Saravanan, NIT Durgapur, India
  • 18. Image negatives • The transformation is very simple : s=T(r) T(r)=(L-1)-r L is the number of gray levels. The result of this transformation is that low intensities are made high and vice versa. Dr. C. Saravanan, NIT Durgapur, India
  • 19. Dr. C. Saravanan, NIT Durgapur, India
  • 20. Contrast stretching • Contrast stretching (often called normalization) attempts to improve the contrast in an image by “stretching” the range of intensity values to a desired range of values. Pout=(Pin-c)((b-a)/(d-c)) + a 8-bit graylevel images the lower and upper limits might be 0 and 255, a & b, lowest and highest pixel values currently present in the image c & d. Dr. C. Saravanan, NIT Durgapur, India
  • 21. Dr. C. Saravanan, NIT Durgapur, India
  • 22. Gray-level slicing • Give a high value for all the gray-levels in the specified range and a very low value for all the other gray-levels. Bit-plane slicing • The intensity of each pixel of an image is defined by several bits - highest order bits are dominant. Dr. C. Saravanan, NIT Durgapur, India
  • 23. Dr. C. Saravanan, NIT Durgapur, India
  • 24. Histogram Equalisation • A histogram with a small spread has low contrast • A histogram with a wide spread has high contrast • An image with its histogram clustered at the low end of the range corresponds to a dark image Dr. C. Saravanan, NIT Durgapur, India
  • 25. Histogram Equalisation steps 1. Find the running sum of the histogram values 2. Normalize the values from step 1 by dividing by the total number of pixels 3. Multiply the values from step 2 by the maximum gray level value and round to the closest integer 4. Map the gray level values to the results from step 3 using a one-to-one correspondence Dr. C. Saravanan, NIT Durgapur, India
  • 26. Histogram Example Step 1: Gray level Number of pixels Running Sum value (Histogram values) 0 10 10 1 8 10+8=18 2 9 10+8+9=27 3 2 29 4 14 43 5 1 44 6 5 49 7 2 51 Dr. C. Saravanan, NIT Durgapur, India
  • 27. Example continues ... Step 2: Normalizing by dividing by the total number of pixels (51) we get 10/51, 18/51, 27/51, 29/51, 43/51, 44/51, 49/51, 51/51 Step 3: Multiply by the maximum gray level value (7) and round we obtain 1, 2, 4, 4, 6, 6, 7, 7 Step 4: Map the original value to the results from step 3 Dr. C. Saravanan, NIT Durgapur, India
  • 28. Histogram equalisation Dr. C. Saravanan, NIT Durgapur, India
  • 29. Unsharp masking • Combine histogram modification and filtering operations. Input Image -> filter -> Histogram modification -> output image Dr. C. Saravanan, NIT Durgapur, India
  • 30. Dr. C. Saravanan, NIT Durgapur, India
  • 31. Homomorphic filter • Simultaneously normalizes the brightness across an image and increases contrast. • To make the illumination of an image more even, the high-frequency components are increased and low-frequency components are decreased, because the high-frequency components are assumed to represent mostly the reflectance in the scene (the amount of light reflected off the object in the scene), whereas the low-frequency components are assumed to represent mostly the illumination in the scene. Dr. C. Saravanan, NIT Durgapur, India
  • 32. Dr. C. Saravanan, NIT Durgapur, India
  • 33. Frequency Domain Methods • Compute Fourier Transform of the image to be enhanced. • Multiply the result by a filter • Take the inverse transform to produce the enhanced image • Will diminish camera noise, spurious pixel values, missing pixel values etc. Dr. C. Saravanan, NIT Durgapur, India
  • 34. Dr. C. Saravanan, NIT Durgapur, India
  • 35. Dr. C. Saravanan, NIT Durgapur, India
  • 36. Dr. C. Saravanan, NIT Durgapur, India
  • 37. Neighbourhood Averaging • Smooth Image F(x,y) = Average pixel value in a neighbourhood of I(x,y) • For example, 3 x 3 neighbourhood • Each pixel value is multiplied by 1/9 • Sum of 9 pixel value is the output Dr. C. Saravanan, NIT Durgapur, India
  • 38. Dr. C. Saravanan, NIT Durgapur, India
  • 39. Edge Preserving • Also called Median Filtering • Median of the neighbourhood pixel values • More like neighbours • Edges are preserved Dr. C. Saravanan, NIT Durgapur, India
  • 40. Dr. C. Saravanan, NIT Durgapur, India
  • 41. Image Sharpening • Human perception is highly sensitive to edges and fine details of an image, and since they are composed primarily by high frequency components, the visual quality of an image can be enormously degraded if the high frequencies are attenuated or completed removed. Dr. C. Saravanan, NIT Durgapur, India
  • 42. Sharpening • In contrast, enhancing the high-frequency components of an image leads to an improvement in the visual quality. • Image sharpening is widely used in printing and photographic industries for increasing the local contrast and sharpening the images. • Enhance detail that has been blurred. Dr. C. Saravanan, NIT Durgapur, India
  • 43. Dr. C. Saravanan, NIT Durgapur, India
  • 44. Edge Detection • Identifying and Locating sharp discontinuities in an image. • Discontinuities are abrupt changes in pixel intensity. • Sobel operator / filter • Canny edge operator Dr. C. Saravanan, NIT Durgapur, India
  • 45. Dr. C. Saravanan, NIT Durgapur, India
  • 46. Sobel Operator • 3 x 3 convolution kernels convolved with original image. • * denotes convolution operation AGx * 101 202 101               AGy * 121 000 121              Dr. C. Saravanan, NIT Durgapur, India
  • 47. Dr. C. Saravanan, NIT Durgapur, India
  • 48. Canny Edge Operator 1. Smoothing: Blurring of the image to remove noise. 2. Finding gradients: The edges should be marked where the gradients of the image has large magnitudes. 3. Non-maximum suppression: Only local maxima should be marked as edges. 4. Thresholding: Potential edges are determined by thresholding. 5. Edge tracking by hysteresis: Final edges are determined by suppressing all edges that are not connected to a very certain (strong) edge. Dr. C. Saravanan, NIT Durgapur, India
  • 49. Dr. C. Saravanan, NIT Durgapur, India
  • 50. Smoothing                  24542 491294 51215125 491294 24542 159 1 B The image is first smoothed by applying a Gaussian filter. Dr. C. Saravanan, NIT Durgapur, India
  • 51. Dr. C. Saravanan, NIT Durgapur, India
  • 52. Finding gradients • Finds edges where the grayscale intensity of the image changes the most. • These areas are found by determining gradients of the image • Gradients at each pixel in the smoothed image are determined by applying Sobel- operator Dr. C. Saravanan, NIT Durgapur, India
  • 53. Non-maximum suppression • Convert the blurred edges in the image of the gradient magnitudes to “sharp” edges. • This is done by preserving all local maxima in the gradient image, and deleting everything else. 1343 3465 6754 4532 Dr. C. Saravanan, NIT Durgapur, India
  • 54. Thresholding • The edge-pixels remaining will probably be true edges in the image. • But some may be caused by noise or color variations. • Discern between these would be to use a threshold. • The Canny edge detection algorithm uses double thresholding, Strong and Weak. Dr. C. Saravanan, NIT Durgapur, India
  • 55. Dr. C. Saravanan, NIT Durgapur, India
  • 56. Edge Tracking by Hysteresis • Strong edges are included in the final edge image. • Weak edges are included if and only if they are connected to strong edges. Dr. C. Saravanan, NIT Durgapur, India
  • 57. Biometric Identification • Biometric identification compares a biometric "signature" to all the records stored in a database to determine if there is a match (1 : N). • Because it requires comparing each existing record in the database against the new biometric characteristic, it can be slow and is usually not suitable for real-time applications such as access control or time and attendance. Dr. C. Saravanan, NIT Durgapur, India
  • 58. Biometric Identification • Biometric identification used most frequently in such applications as law enforcement — for instance, the comparison of a fingerprint from a crime scene to a database of prints collected from convicted criminals. Dr. C. Saravanan, NIT Durgapur, India
  • 59. Biometric Verification • Biometric verification compares a newly- scanned biometric characteristic to a measurement previously collected from that same person to verify that individual's identity (1 : 1). • For instance, when an employee is hired, that employee's fingerprint will be enrolled into the company's biometric time and attendance system. Dr. C. Saravanan, NIT Durgapur, India
  • 60. FAR False Acceptance Rate (FAR) • is the measure of the likelihood that the biometric security system will incorrectly accept an access attempt by an unauthorized user. • is stated as the ratio of the number of false acceptances divided by the number of identification attempts. Dr. C. Saravanan, NIT Durgapur, India
  • 61. FRR False Recognition Rate (FRR) • is the measure of the likelihood that the biometric security system will incorrectly reject an access attempt by an authorized user. • is stated as the ratio of the number of false rejections divided by the number of identification attempts. Dr. C. Saravanan, NIT Durgapur, India
  • 62. Positive / Negative Identification • False Non-Match Rate (FNMR) • False Match Rate (FMR) • False Reject Rate (FRR) • False Accept Rate (FAR) • False Positive Identification Rate (FPIR) • False Negative Identification Rate (FNIR) Dr. C. Saravanan, NIT Durgapur, India
  • 63. Matching Score Distribution • A matching algorithm was defined as an algorithm that make a decision about genuine or impostor nature of a comparison between two templates. • In the first step an Evaluation Algorithm assigns a similarity score to the comparison. That similarity score is a value on the range (0..1) and as higher be the score value more similar the images. • The second step decides if the comparison is genuine or impostor using a frontier threshold or decision threshold (DT). Dr. C. Saravanan, NIT Durgapur, India
  • 64. Receiver Operating Characteristic (ROC) • accepted method for summarizing the performance of imperfect pattern matching systems. • parametrically as a function of the decision threshold, • the rate of “false positives” on the x-axis • the rate of “true positives” on the y-axis • ROC curves are threshold independent Dr. C. Saravanan, NIT Durgapur, India
  • 65. Dr. C. Saravanan, NIT Durgapur, India
  • 66. Detection Error Trade-off (DET) • modified ROC curve • plots error rates on both axes • giving uniform treatment to both types of error • distinguishes different wellperforming systems more clearly Dr. C. Saravanan, NIT Durgapur, India
  • 67. Dr. C. Saravanan, NIT Durgapur, India
  • 68. Expected Overall Error gi TFNMRTFMRTE   )()()( • Expected overall error takes into account the possibility of different FM and FNM and is given as where, T=threshold pi=probability of a random user being an imposter. pg=probability of a random user being genuine. Dr. C. Saravanan, NIT Durgapur, India
  • 69. Equal Error Rate (EER) • EER is the value where FMR and FNMR are equal. • Lower be the EER the lower error rate of the algorithm. • Select the EER score value as Decision Threshold (DT) is frequently a good decision for a regular biometric application. Dr. C. Saravanan, NIT Durgapur, India
  • 70. Myths & Misrepresentations • Biometric “X” is best for all applications • Biometric “X” is unique for each individual • A single number quantifies system accuracy • System is “plug and play” • Real accuracy performance can be predicted Dr. C. Saravanan, NIT Durgapur, India
  • 71. • The vendors reporting best FAR and FRR has the “most accurate system” • Multiple biometrics outperform single biometrics • “Our biometric system does not use a decision threshold” • “Our feature extractor can be used with any match engine” • Large templates mean better accuracy • Face recognition prevents terrorism Dr. C. Saravanan, NIT Durgapur, India
  • 72. • Biometrics means 100 percent security • Biometric systems invade our privacy • Biometric sensors are unhygienic or otherwise harmful Dr. C. Saravanan, NIT Durgapur, India
  • 73. Biometric System Design Issues • System Architecture – Centralised / Distributed Server – Client Computer – Device at User End • Hardware & Software Implementation – Sample Acquitision – User Interface – Biometric Processing Components Dr. C. Saravanan, NIT Durgapur, India
  • 74. BSD Issues (continues...) – Communication Channels – Database Design – Interoperability • Administration Policy – Integrity of Enrolment – Quality of Enrolment Samples – System Configuration – Exception Handling – Privacy Measures Dr. C. Saravanan, NIT Durgapur, India
  • 75. Biometric System Security • Includes IT Security • Earlier related financial • At present, Passports, Visas, etc. – Biometric Security Evaluation – Biometric Transaction Security – Protection of Biometric Data Dr. C. Saravanan, NIT Durgapur, India
  • 76. Authentication Protocol • is a type of cryptographic protocol with the purpose of authenticating entities wishing to communicate securely. • AKA, CAVE-based_authentication, Challenge- handshake authentication protocol (CHAP), CRAM-MD5 Diameter, Digest, Extensible Authentication Protocol (EAP), Host Identity Protocol (HIP), Kerberos, MS-CHAP, LAN Manager, NTLM, Password-authenticated key agreement protocols, Password Authentication Protocol (PAP), Protected Extensible Authentication Protocol (PEAP), Protocol for Carrying Authentication for Network Access (PANA), RADIUS, Secure Remote Password protocol (SRP), TACACS and TACACS+, RFID- Authentication Protocols, Woo Lam 92 (protocol) Dr. C. Saravanan, NIT Durgapur, India
  • 77. Kerberos • MIT developed Kerberos • Kerberos (or Cerberus - Greek monstrous three-headed guard dog) • Works on the basis of 'tickets' to allow nodes communicating over a non-secure network to prove their identity to one another in a secure manner. • Protects against eavesdropping and replay attacks. Dr. C. Saravanan, NIT Durgapur, India
  • 78. Selection of Suitable Biometric • Ethnic Background – who will be the users of the biometric system – education levels and a variety of attitudes • Employee Education – technical background of the users. – higher the technological background less training required Dr. C. Saravanan, NIT Durgapur, India
  • 79. • Frequency of Use – some biometric systems are more suitable for high frequency of usage • User characteristics – will the users be in a hurry and possibly be a little bit impatient (public restroom) Dr. C. Saravanan, NIT Durgapur, India
  • 80. Biometric Attributes 1. Universal: Every person must possess the characteristic / attribute. The attribute must be one that is universal and seldom lost to accident or disease. 2. Invariance of properties: They should be constant over a long period of time. The attribute should not be subject to significant differences based on age either episodic or chronic disease. 3. Measurability: The properties should be suitable for capture without waiting time and must be easy to gather the attribute data passively. 4. Singularity: Each expression of the attribute must be unique to the individual. The characteristics should have sufficient unique properties to distinguish one person from any other. Dr. C. Saravanan, NIT Durgapur, India
  • 81. Height, weight, hair and eye color are all attributes that are unique assuming a particularly precise measure, but do not offer enough points of differentiation to be useful for more than categorizing. 5. Acceptance: The capturing should be possible in a way acceptable to a large percentage of the population. Excluded are particularly invasive technologies, i.e. technologies which require a part of the human body to be taken or which impair the human body. 6. Reducibility: The captured data should be capable of being reduced to a file which is easy to handle. 7. Reliability and tamper-resistance: The attribute should be impractical to mask or manipulate. The process should ensure high reliability and reproducibility. Dr. C. Saravanan, NIT Durgapur, India
  • 82. 8. Privacy: The process should not violate the privacy of the person. 9. Comparable: Should be able to reduce the attribute to a state that makes it digitally comparable to others. The less probabilistic the matching involved, the more authoritative the identification. 10. Inimitable: The attribute must be irreproducible by other means. The less reproducible the attribute, the more likely it will be authoritative. Dr. C. Saravanan, NIT Durgapur, India
  • 83. Zephyr Chart Dr. C. Saravanan, NIT Durgapur, India Zephyr Chart is a navigation aid for pilots designed to display aviation charts in a format easy to use in the cockpit.
  • 84. Multi Biometrics (MB) • MB - Reconciliation of Evidence enhances the recognition accuracy of human authentication systems. • Multibiometric systems combine the information presented by multiple biometric sensors, algorithms, samples, units, or traits. Dr. C. Saravanan, NIT Durgapur, India
  • 85. Dr. C. Saravanan, NIT Durgapur, India
  • 86. MB • Besides enhancing matching performance, these systems are expected to – improve population coverage, – deter spoofing and – impart fault tolerance to biometric applications • Establishes an identity based on who you are, – rather than by what you possess, such as an ID card, – or what you remember, such as a password. Dr. C. Saravanan, NIT Durgapur, India
  • 87. Risks with Single Biometric Trait • Noise in sensed data • Non-universality • Upper bound on identification accuracy • Spoof attacks (voice and signature - mimic) • fingerprints can also be spoofed by inscribing ridge-like structures on synthetic material such as gelatine and play-doh. Dr. C. Saravanan, NIT Durgapur, India
  • 88. Advantages of MB • If a biometric trait (eg. fingerprint) fails other trait (eg. voice) aid to authenticate the person. • Facilitates the filtering or indexing of largescale biometric databases. Among the traits, the simple one may be used for indexing. Dr. C. Saravanan, NIT Durgapur, India
  • 89. • Difficulties has been increased the spoofing effort. Randomly traits can be acquired at the time of authentication. • In case of noise with one traint other trait could help in authentication. • Certain traits may not be applicable all the time, the other traits may be used. • Highly fault tolerant, because even a trait fails other traits are available. Dr. C. Saravanan, NIT Durgapur, India
  • 90. MB Classifications • multi-sensor, multi-algorithm, • multi -instance (left and right index fingers and left and right irises), • multi-sample (pose of a face - front, left, right), • multi-modal (face and voice) and hybrid. Dr. C. Saravanan, NIT Durgapur, India
  • 91. Dr. C. Saravanan, NIT Durgapur, India
  • 92. MB Fusion • Fusion before matching • Fusion after matching • Sensor level - multiple sensors - face. • Feature level - multi features are consolidated into single feature. • Score level - generating a single score. • Rank level - ranking of enrolled identities • Decision level - access only to the final recognition decision. Dr. C. Saravanan, NIT Durgapur, India
  • 93. Fusion • Combining different sources of information for intelligent systems. • The information are signals delivered by different sensors and images from various modalities. • The fusion concepts and methods gather tools like weighted average, neural networks, sub-band filtering, and rules based knowledge. Dr. C. Saravanan, NIT Durgapur, India
  • 94. Fusion Categories • Multiview - same modality and time - different places and conditions. • Multimodal - CT, MRI • Multitemporal - same scene - different time • Multifocus - regions based Dr. C. Saravanan, NIT Durgapur, India
  • 95. Types of Multisensor Fusion • Signal-level fusion • Image-level fusion • Feature-level fusion • Symbol-level fusion Dr. C. Saravanan, NIT Durgapur, India
  • 96. Fusion Applications • Intelligent robots – visual, tactile, force/torque, and other types of sensors – Stereo camera fusion • Medical image – Fusing X-ray computed tomography (CT) and magnetic resonance (MR) images • Manufacturing – Complex machine/device diagnostics • Military and law enforcement – Detection, tracking, identification of ocean (air, ground) target / event • Remote sensing
  • 97. Image Fusion • Improves reliability (redundant information) • Improves capability (complementary info.) Dr. C. Saravanan, NIT Durgapur, India
  • 98. IF Techniques • Pixel values are manipulated to achieve desired result (SPATIAL Domain). • Fusion operations are performed on frequency (TRANSFORM domain). • Primitive fusion (Averaging Method, Select Maximum, and Select Minimum), • Discrete Wavelet transform based fusion, • Principal component analysis (PCA) based fusion. Dr. C. Saravanan, NIT Durgapur, India
  • 99. • The fusion methods such as averaging, Brovey method, principal component analysis (PCA) and Intensity-Hue-Saturation (IHS) based methods fall under spatial domain approaches. • Another important spatial domain fusion method is the high pass filtering based technique. • The disadvantage of spatial domain approaches is that they produce spatial distortion in the fused image. • Spatial distortion becomes a negative factor while we go for further processing such as classification problem Dr. C. Saravanan, NIT Durgapur, India
  • 100. IF Methods • Intensity-hue-saturation (IHS) transform based fusion • Principal component analysis (PCA) based fusion • Multi scale transform based fusion:- – High-pass filtering method – Pyramid method:-(i) Gaussian pyramid (ii) Laplacian • Pyramid (iii) Gradient pyramid (iv) Morphological pyramid – Ratio of low pass pyramid – Wavelet transforms:- (i) Discrete wavelet transforms (DWT) (ii) Stationary wavelet transforms (iii) Multiwavelet transforms – Curvelet transforms Dr. C. Saravanan, NIT Durgapur, India
  • 101. Simple Average • K (i, j) = {X (i, j) + Y (i, j)}/2 Where X (i , j) and Y ( i, j) are two input Images regions of images that are in focus tend to be of higher pixel intensity Dr. C. Saravanan, NIT Durgapur, India
  • 102. Select Maximum • The value of the pixel P (i, j) of each image is taken and compared to each other. • The greatest pixel value is assigned to the corresponding pixel of the output image. • This algorithm chooses the in-focus regions from each input image by choosing the greatest value for each pixel, resulting in highly focused output. Dr. C. Saravanan, NIT Durgapur, India
  • 103. DWT • Wavelets can be described by using two functions viz. • the scaling function f (t), also known as “father wavelet” and • the wavelet function or “mother wavelet”. Mother wavelet (t) • undergoes translation and scaling operations to give self similar wavelet families. Dr. C. Saravanan, NIT Durgapur, India
  • 104. • The wavelet transform decomposes the image into low-high, high-low, high-high spatial frequency bands at different scales and the low-low band at the coarsest scale • The L-L band contains the average image information whereas the other bands contain directional information due to spatial orientation. Dr. C. Saravanan, NIT Durgapur, India
  • 105. Dr. C. Saravanan, NIT Durgapur, India Higher absolute values of wavelet coefficients in the high bands correspond to salient features such as edges or lines.
  • 106. Dr. C. Saravanan, NIT Durgapur, India
  • 107. Dr. C. Saravanan, NIT Durgapur, India
  • 108. PRINCIPAL COMPONENT ANALYSIS (PCA) • Statistical Procedure • Transforms a number of correlated variables into a number of linearly uncorrelated variables called principal components. • The first principal component accounts for as much of the variance in the data as possible and • Each succeeding component accounts for as much of the remaining variance as possible. Dr. C. Saravanan, NIT Durgapur, India
  • 109. Different Names of PCA • discrete Karhunen–Loève transform (KLT) in signal processing, • the Hotelling transform in multivariate quality control, • Proper Orthogonal Decomposition (POD) in mechanical engineering, • Singular Value Decomposition (SVD) of X (Golub and Van Loan, 1983), • Eigenvalue Decomposition (EVD) of XTX in linear algebra, • Factor Analysis (for a discussion of the differences between PCA and factor analysis, Dr. C. Saravanan, NIT Durgapur, India
  • 110. • Eckart–Young theorem (Harman, 1960), or Schmidt–Mirsky theorem in sychometrics, • Empirical Orthogonal Functions (EOF) in meteorological science, • Empirical Eigenfunction decomposition (Sirovich, 1987), • Empirical Component Analysis (Lorenz, 1956), • Quasiharmonic modes (Brooks et al., 1988), • Spectral Decomposition in noise and vibration, and empirical modal analysis in structural dynamics. Dr. C. Saravanan, NIT Durgapur, India
  • 111. Multimodal Biometrics • Multimodal Biometrics are systems that are capable of using more than one physiological or behavioural characteristic for enrolment, verification or identification. • will use a combination of the following recognition technologies, up to three of them, to compare the identity of a person. • If one of the technologies fails for any reason, your system can still use another one or two. Dr. C. Saravanan, NIT Durgapur, India
  • 112. Dr. C. Saravanan, NIT Durgapur, India
  • 113. Dr. C. Saravanan, NIT Durgapur, India
  • 114. Forms of Biometric Identification • Some forms of behavioural biometric identification include the following: – Keystroke or Typing Recognition – Speaker identification or Recognition • Some forms of physical biometric identification include the following: – Fingerprint – Iris – Retina – Finger Geometry – Signature/Handwriting – Voice – Facial Proportions – Hand Geometry Dr. C. Saravanan, NIT Durgapur, India
  • 115. Benefits • The benefits of multimodal biometrics is that by using more than one means of identification, your system can retain a high threshold recognition setting and your system administrator can decide the level of security that is needed. Dr. C. Saravanan, NIT Durgapur, India
  • 116. Need • There is a great need for multimodal biometrics as most biometric systems used in real applications are unimodal, which means they rely on only one area of identification. • Some examples of these are fingerprints, faces and voices and these systems are quite vulnerable to many problems such as noisy data, non-universality and spoofing. • This leads to a high false acceptance rate and false rejection rate, limited discrimination capability, and lack of permanence. Dr. C. Saravanan, NIT Durgapur, India
  • 117. Applications • Strong Potential: – Physical access – Civil ID – Criminal ID • Moderate Potential: – Network/PC access – Kiosk/ATM • Modest Potential: – Retail/POS – Surveillance – eCommerce – telephony Dr. C. Saravanan, NIT Durgapur, India
  • 118. Reliability • Multimodal systems are more reliable because using many independent biometrics that meet very high performance requirements and they counteract the problems listed above. • They also effectively deter spoofing because it is near impossible to spoof multiple biometric traits and the system can request the user to present random traits that only a live person can do. Dr. C. Saravanan, NIT Durgapur, India
  • 119. Biometric System Security • Most biometric systems will advertise a False Accept Rate (FAR) and False Reject Rate (FRR) to characterize the security provided by the system. • The FAR tells you how often someone will be recognized successfully when he/she should not have been recognized, and the FRR tells you how often someone who should have been recognized successfully is not recognized. Dr. C. Saravanan, NIT Durgapur, India
  • 120. • The FAR is an upper bound on security. • The FAR assumes random and real samples of the biometric characteristic are presented to the system. • Attackers don’t have to use random or real samples; they are much more likely to pick samples intelligently and achieve a much higher FAR. Dr. C. Saravanan, NIT Durgapur, India
  • 121. • The attacker can do this in two ways. • First the attacker can use a copy of the biometric characteristic of a valid user collected by the system to produce a fake biometric characteristic that will allow access. • This attack is called a physical spoof attack and will allow the attacker to bypass the biometric system. • This attack is like finding a password written down or watching a valid user enter a password, copying it, and using it to gain access to the system. Dr. C. Saravanan, NIT Durgapur, India
  • 122. • Second, if the attacker does not have a copy of the biometric characteristic of a valid user, the attacker can attempt to create a fake biometric characteristic by guessing “intelligently” or the attacker can use a database of real samples of invalid users “intelligently” instead of randomly. • As the guesses get better, the attacker will likely bypass the biometric system in fewer attempts than presenting random samples. Dr. C. Saravanan, NIT Durgapur, India
  • 123. • This is loosely comparable to an attacker using a dictionary to attack a password based system rather than randomly going through all possible passwords of the proper lengths. Dr. C. Saravanan, NIT Durgapur, India
  • 124. Liveness • For biometric systems, liveness detection makes both attacks more difficult. • However, while liveness detection is improving in biometric systems overall, many systems employing liveness detection are still susceptible to physical spoof attacks. Dr. C. Saravanan, NIT Durgapur, India
  • 125. • Rather than attacking using samples of the biometric characteristic, other attacks on a biometric system may be possible, such as cryptographic attacks, network attacks, operating system attacks, etc. • All these potential vulnerabilities must be considered when implementing a biometric system with the intent of enhancing system security. Dr. C. Saravanan, NIT Durgapur, India
  • 126. Dr. C. Saravanan, NIT Durgapur, India
  • 127. Eight Types of Attacks • Type 1 attack involves presenting a fake Biometric – (e.g., finger made from silicon, face mask, lens including fake iris texture) to the sensor. • The second type of attack is called a replay attack, – because an intercepted biometric (with or without the cooperation of the genuine user) data is submitted to the feature extractor, bypassing the sensor. Dr. C. Saravanan, NIT Durgapur, India
  • 128. • In the third type of attack, the feature extractor module is replaced with a Trojan horse program – that functions according to its designer's specifications Dr. C. Saravanan, NIT Durgapur, India
  • 129. • In the fourth type of attack, genuine feature values are replaced with values (synthetic or real) selected by the attacker. • In the fifth type of attack, the matcher is replaced with a Trojan horse program. • The attacks on the template database (e.g., addition, modification, or removal of templates) constitute the sixth type of attack. Dr. C. Saravanan, NIT Durgapur, India
  • 130. • In the seventh type of attack, the templates are tampered (stolen, replaced, or altered) – in the transmission medium between the template database and matcher. • Lastly, the matcher result (accept or reject) can be overridden by the attacker. Dr. C. Saravanan, NIT Durgapur, India
  • 131. Dr. C. Saravanan, NIT Durgapur, India
  • 132. Attacks depend on (i) biometric modality (e.g., it is harder to physically replicate a retina scan than it is to forge a signature), (ii) type of sensor (2D vs. 3D face sensors), (iii) type of matcher operating on the same biometric (face matchers based on texture vs. geometry), and (iv) the security settings (reflected via False Accept Rate) of the biometric system. Dr. C. Saravanan, NIT Durgapur, India
  • 133. Dr. C. Saravanan, NIT Durgapur, India
  • 134. Face Synthesis Dr. C. Saravanan, NIT Durgapur, India
  • 135. Fingerprint Watermarked in Face Dr. C. Saravanan, NIT Durgapur, India
  • 136. Secure Sytems Properties (i) accuracy (ii) rapid (iii) reliable (iv) protecting privacy rights (v) cost effectively (vi) user-friendly (vii) minimal changes to the existing infrastructures Dr. C. Saravanan, NIT Durgapur, India
  • 137. Liveness Detection Methods • Fingerprints – periodicity of sweat and the sweat diffusion pattern were used to detect fake fingerprints. – liveness detection method based on skin elasticity properties – The user is asked to deliberately rotate his finger when removing it from the sensor surface thus producing considerable skin distortion which is later used as a fingerprint liveness measure. Dr. C. Saravanan, NIT Durgapur, India
  • 138. – An odour sensor (electronic nose) is used to discriminate the skin odour from that of other materials such as latex, silicone or gelatine. • Iris – the spectrographic properties of different parts of the eye (tissue, fat, blood, melanin pigment), the coaxial retinal back reflection (the red eye effect) and – the four Purkinje reflections caused by each of the four optical surfaces comprised inside the eye. – very high quality cameras required. Dr. C. Saravanan, NIT Durgapur, India
  • 139. – permanent oscillation that the eye pupil presents even under uniform lighting conditions – measure the pupil response to a sudden lighting event – infrared light reflections from the moist cornea Dr. C. Saravanan, NIT Durgapur, India
  • 140. Vulnerabilities of Biometric Systems • Administration: Intentional or unintentional administrative mistakes. • User: A legitimate user wants to upgrade his privileges to the administrative level. • Enrolment: Breaking registration procedures. • Spoofing: A fake biometric is used for authentication as a legitimate user. Dr. C. Saravanan, NIT Durgapur, India
  • 141. • Mimicry: Attacker mimics the biometric characteristics of the legitimate user. • Undetect: Attacks undetected by the system may encourage new attacks. • Fail secure: Result of abnormal utilization conditions of biometric system or IT environment. • Bypass: Bypassing biometric system for access. This can be achieved by surpassing physical barriers, forcing a legitimate user to present his biometric to the sensor, or by cooperation of legitimate user. Dr. C. Saravanan, NIT Durgapur, India
  • 142. • Corrupt attack – Weakening the system by making changes in the IT environment or biometric system. Modification or replacement of system parameters is an example. • Degrade: Certain software in the IT environment decreases the system’s security level. Dr. C. Saravanan, NIT Durgapur, India
  • 143. • Tamper: Counterfeiting the hardware of the system. • Residual: Latent fingerprints may be used to make artificial fingerprints or accepted directly by the sensor. Dr. C. Saravanan, NIT Durgapur, India
  • 144. • Cryptological attack: Encryption can be broken in data transmission and this biometric data can be used for another type of attack (e.g. replay attack). • Brute force attack: Attacker presents the biometric characteristic to the system repeatedly in order to be authenticated. • This type of attack depends on FAR parameter. • Evil twin attacks: Biometric characteristic of imposter is very similar to the enrolled user’s biometric. Dr. C. Saravanan, NIT Durgapur, India
  • 145. • Fake template: Introducing fake biometric template into the database or onto smart cards. • Noise: Access can be gained by the attacker when noise is applied to the system. • Poor image: Quality supervision may be utilized. If low quality images are accepted for registration then attacker may hope to deceive the system as in the case of noisy images. • Weak ID: Similar to “poor image” weakness, and tries to fake the system by weak templates. Dr. C. Saravanan, NIT Durgapur, India
  • 146. Any Questions ? Dr. C. Saravanan, NIT Durgapur, India
  • 147. References  Biometrics: Identity Assurance in the Information Age, John D. Woodward Jr.  Biometrics: Advanced Identity Verification: The Complete Guide, Julian Ashbourn  Biometrics: Identity Verification in a Networked World, Samir Nanavati  Digital Image Processing, Gonzalez and Woods  Wikipedia Dr. C. Saravanan, NIT Durgapur, India