Download as pdf or txt
Download as pdf or txt
You are on page 1of 102

UNIT-2

FINGERPRINT BIOMETRICS
Finger print using vein pattern of palm

Palm vein authentication uses the blood vessel patterns of the


palm vein in the subcutaneous tissue of the human body to
discriminate between individuals.

Palm vein patterns are captured by the camera with near-infrared


light. When a hypodermic vein is irradiated with near-infrared light,
the reduced haemoglobin contained in the vein absorbs
near-infrared light and the hypodermic vein creates a shadow on an
image.

22OBM103- BIOMETRICS AND ITS APPLICATION


1
UNIT-2

Finger print using vein pattern of palm


The shadow pattern is then extracted from the captured image of
the palm vein pattern using image processing technology. The
resulting vein patterns are compared using vessel structure features
such as directions and bifurcations, or by using the patterns
themselves.

In practical terms, palm vein patterns of a hand are used for


authentication because such parts of the hand are easy to expose
to a sensor

22OBM103- BIOMETRICS AND ITS APPLICATION

2
UNIT-2

Finger print using vein pattern of palm


The palm vein authentication technology has been deployed for its
ease-of-use and assurance that it has given users through its robust
security.

It has been widely adopted worldwide for personal identification at


financial institutions, as a computer login and room entrance control
method at corporations.

22OBM103- BIOMETRICS AND ITS APPLICATION

3
UNIT-2

Types of Biometric Traits

22OBM103- BIOMETRICS AND ITS APPLICATION


4
UNIT-2

Palm vein authentication


Palm vein authentication systems generally use optical palm
vein sensors for enrolment of the palm vein image or palm
vein pattern. In enrolment, users put their palm above the
sensor (figure 1).

The sensor detects a palm, and emits near-infrared light as


figure 1.

When near-infrared light is illuminated to the palm, the rays


will be scattered in the hand after penetrating the surface of
the hand. The illuminated light of a near infrared light enters
inside the palm as incident light. The incident light is
absorbed and scattered in the palm.
22OBM103- BIOMETRICS AND ITS APPLICATION
5
UNIT-2

Finger print using vein pattern of palm


A part of incident light is also absorbed within the vein. Deoxidized
hemoglobin absorbs near-infrared rays at a wavelength of
approximately 760 (nm). The palm vein sensor device emits
near-infrared ray and photographs the light that is scattered back
from the palm.

22OBM103- BIOMETRICS AND ITS APPLICATION

6
UNIT-2

Types of Biometric Traits


Image processing is used to extract the vein pattern properties and
to compare patterns from different palms. Figure 2 is a sample of a
captured palm vein pattern image by using an experimental
scanner.
In order to capture a palm vein pattern image with a camera, the
scattering light from a palm is used. A specular reflection light from a
palm obstructs the acquisition of the palm vein pattern.
By using a scattering light from a palm, the palm vein sensor can
obtain a palm vein pattern image. To collect only the scattered light,
the palm vein sensor has particular imaging optics and illumination
optics. As a result, when capturing an image of a palm using
near-infrared rays, veins appear as a shadow, darker than the
surrounding area.
22OBM103- BIOMETRICS AND ITS APPLICATION
7
UNIT-2

Types of Biometric Traits

22OBM103- BIOMETRICS AND ITS APPLICATION


8
UNIT-2

Types of Biometric Traits

Then the palm vein extraction algorithm extracts the vein


impression from the image taken by the sensor. This extracted vein
impression is then stored in the database as a template. In the
matching process, the registration data and the reference data are
compared with each other (figure 3).

A palm has a large number of veins, hence the accuracy rate is


much higher than other options available. Furthermore, exposing
the palm is a more intuitive and natural movement than showing the
wrist or back-of-hand. Palm veins are thicker than finger veins, so
they are more stable even when exposed to temperature
fluctuations. Also, the palm has the absence of hair.
22OBM103- BIOMETRICS AND ITS APPLICATION
9
UNIT-2

Finger print using vein pattern of palm

There are four properties of vein patterns:


1. Vein patterns exist under the skin and are normally not visible.
2. Vein patterns are unique to individuals (even identical twins have
different patterns).
3. Vein patterns do not change significantly with age.
4. Vein patterns can be captured from almost everyone.
The system’s recognition accuracy is extremely high, with a
false-negative rate of 0.01% and a false-positive rate of 0.00008%.
Furthermore, the fact that the system uses an individual’s thicker
veins allows for stable operations over time. Lastly, the system is
contact-free, helping to reduce any existing psychological burdens
placed on users.
figure 4. 22OBM103- BIOMETRICS AND ITS APPLICATION
10
UNIT-2

Fingerprint Recognition System:


Fingerprint recognition is one of the most widely used biometric
technologies, employed for various applications such as access
control, mobile device security, and forensic identification.

The system captures and analyzes the unique patterns and ridges
present on an individual's fingertips.

Here is an overview of the components and working of a fingerprint


recognition system:

22OBM103- BIOMETRICS AND ITS APPLICATION


11
UNIT-2

Fingerprint Recognition System:


Fingerprint Sensor:
A fingerprint sensor is a hardware device that captures the
fingerprint impressions. It may use different technologies like
optical, capacitive, or ultrasonic to acquire the unique features
of the fingerprint.
Image Preprocessing:
The captured fingerprint image often undergoes preprocessing
to enhance the quality and clarity of the image. This may
involve noise reduction, contrast adjustment, and image
enhancement techniques.

22OBM103- BIOMETRICS AND ITS APPLICATION


12
UNIT-2

Types of Biometric Traits


Fingerprint Recognition System:
Feature Extraction:
Feature extraction involves identifying and extracting distinctive
features from the fingerprint image. These features may include
ridge endings, bifurcations, and minutiae points, which are
unique to each fingerprint.
Template Creation:
A template is created based on the extracted features. This
template is a digital representation of the unique characteristics
of the fingerprint and serves as a reference for future
comparisons.

22OBM103- BIOMETRICS AND ITS APPLICATION


13
UNIT-2

Fingerprint Recognition System:


Storage:
The template is securely stored in a database. For security
purposes, the actual fingerprint image is usually not stored;
instead, only the template is retained.
Matching Algorithm:
When a user attempts to access a system or device, their
fingerprint is captured, and a new template is created. The
matching algorithm compares this newly created template with
the stored reference template to determine if there is a match.

22OBM103- BIOMETRICS AND ITS APPLICATION


14
UNIT-2

Types of Biometric Traits


Fingerprint Recognition System:
Decision Making:
The system makes a decision based on the similarity score
obtained from the matching algorithm. If the score exceeds a
predefined threshold, the system declares a match, and access
is granted.
Authentication/Identification:
In an authentication scenario, the system verifies whether the
presented fingerprint matches the stored template for the
claimed identity. In identification, the system searches the
database for a match among multiple templates.

22OBM103- BIOMETRICS AND ITS APPLICATION


15
UNIT-2

Types of Biometric Traits


Fingerprint Recognition System:
Advantages:
Uniqueness: Fingerprint patterns are highly unique to individuals.
Accuracy: Fingerprint recognition systems provide high accuracy
when properly implemented.
Non-intrusive: Capturing fingerprints is non-intrusive and does not
require physical contact.

22OBM103- BIOMETRICS AND ITS APPLICATION


16
UNIT-2
minutiae extraction

Fingerprint identification has a great utility in forensic science and


aids criminal investigations etc. Most of the automatic fingerprint
recognition systems are based on local ridge features known as
minutiae. Hence it is extremely important to mark these minutiae
accurately and reject the false ones.

However, fingerprint images are prone to degradation and


corruption due to factors such as skin variations and impression
conditions such as scars, dirt, humidity and non-uniform contact with
the scanning device. Thus it is necessary to apply some type of
image enhancement techniques before minutiae extraction.

22OBM103- BIOMETRICS AND ITS APPLICATION


17
UNIT-2

The most important step in automatic fingerprint matching is to


reliably extract the minutiae from the captured fingerprint images.
There exists a variety of techniques for extracting fingerprint
minutiae and are broadly classified into two types – techniques that
work on binarized images and those techniques that work on gray
scale images.

Fingerprints and minutiae


A fingerprint is a distinct pattern of ridges and valleys on the finger
surface of an individual. A ridge is defined to be a single curved
segment whereas a valley is the area between two adjacent ridges.
So the dark areas of the fingerprint are called ridges and white area
that exists between
22OBM103-them is known as
BIOMETRICS ANDvalleys.
ITS APPLICATION
18
UNIT-2

In case of a fingerprint identification system, the captured fingerprint


image needs to be matched against the stored fingerprint templates
of every user in the database. This involves a lot of computation and
search overhead and thus we need a fingerprint classification
system that will help us to severely restrict the size of the templates
database.

To accomplish this, we extract the minutiae features and match


against the incoming fingerprint. The template size of minutiae
based fingerprint representation is small and most of the fingerprint
identification systems are based on minutiae.

22OBM103- BIOMETRICS AND ITS APPLICATION


19
UNIT-2

Minutiae points
Minutiae points are the major features of a fingerprint image and are
used in the matching of fingerprints. These minutiae points are used
to determine the uniqueness of a fingerprint image. A good quality
fingerprint image can have 25 to 80 minutiae depending on the
fingerprint scanner resolution and the placement of finger on the
sensor.

22OBM103- BIOMETRICS AND ITS APPLICATION


20
UNIT-2

What is the definition of minutiae? Minutiae can be defined as the


points where the ridge lines end or fork. So the minutiae points are
the local ridge discontinuities and can be of many types.
These types are –
Ridge ending is the point where the ridge ends suddenly.
Ridge bifurcation is the point where a single ridge branches out
into two or more ridges.
Ridge dots are very small ridges.
Ridge islands are slightly longer than dots and occupy a middle
space between two diverging ridges.
Ponds or Lakes are the empty space between two diverging
ridges.
22OBM103- BIOMETRICS AND ITS APPLICATION
21
UNIT-2

Types of Biometric Traits


Spurs is a notch protruding from a ridge.
Bridges are the small ridges that join two longer adjacent ridges.
Crossovers are formed when two ridges cross each other.
Ridge endings and ridge bifurcations are the most commonly used
minutia types since all other types of minutiae are based on a
combination of these two types. Figure below shows some of the
common minutiae patterns.

22OBM103- BIOMETRICS AND ITS APPLICATION


22
UNIT-2

Types of Biometric Traits

22OBM103- BIOMETRICS AND ITS APPLICATION


23
UNIT-2

Minutiae based fingerprint recognition


It is the most widely used technique of fingerprint representation and
its configuration is highly distinctive. It is more accurate compared to
other correlation based systems and the template size is smaller in
minutiae based fingerprint representation. In this system, two
fingerprints match if their minutiae points match.

22OBM103- BIOMETRICS AND ITS APPLICATION


24
UNIT-2

Minutiae based fingerprint recognition


Minutiae based fingerprint technique is the backbone of most
currently available fingerprint recognition products.
Compared to other fingerprint features, the minutia point features
having corresponding orientation maps are distinct enough to
distinguish between fingerprints robustly. Fingerprint representation
using minutiae feature reduces the complex issue of fingerprint
recognition to an issue of point pattern matching.

22OBM103- BIOMETRICS AND ITS APPLICATION


25
UNIT-2

Types of Biometric Traits


Minutiae based fingerprint recognition
Minutiae extraction techniques
A good quality image is absolutely essential for minutiae extraction.
However, sometimes the image quality might be poor due to various
reasons and hence it becomes necessary to enhance the fingerprint
images before minutiae matching of fingerprints.

The minutiae extraction methods are classified into two broad


categories.
Methods that work on binarized fingerprint images:
Methods that work directly on gray-scale fingerprint images.
Given below is a diagram showing the different categories of
minutiae extraction techniques.
22OBM103- BIOMETRICS AND ITS APPLICATION
26
UNIT-2

Types of Biometric Traits


Minutiae based fingerprint recognition
Minutiae extraction techniques
A good quality image is absolutely essential for minutiae extraction.
However, sometimes the image quality might be poor due to various
reasons and hence it becomes necessary to enhance the fingerprint
images before minutiae matching of fingerprints.

The minutiae extraction methods are classified into two broad


categories.
Methods that work on binarized fingerprint images:
Methods that work directly on gray-scale fingerprint images.
Given below is a diagram showing the different categories of
minutiae extraction techniques.
22OBM103- BIOMETRICS AND ITS APPLICATION
27
UNIT-2

22OBM103- BIOMETRICS AND ITS APPLICATION


28
UNIT-2

The following subsections will discuss the above methods


elaborately:
Unthinned binarized images
We will look at the three methods of minutiae extraction from
un-thinned binarized images. These techniques are: Chaincode
processing, Run based methods and Ridge flow and local pixel
analysis based methods.

22OBM103- BIOMETRICS AND ITS APPLICATION


29
UNIT-2

Chaincode processing
This technique is based on the chaincode representation of object
contours and the pixel image can be recovered fully from the
chaincode of its contour. In this method, the transitions from white
background to black foreground are identified by scanning the
image from top to bottom and right to left.

It is then expressed as an array of contour elements by tracing the


contour counter clockwise and each element represents a pixel on
the contour. By tracing a ridge line along its boundary counter
clockwise, a minutiae ending is located when the ridge line makes a
significant left turn. Similarly, if the trace makes a right turn a
bifurcation minutia
22OBM103-is detected.
BIOMETRICS AND ITS APPLICATION
30
UNIT-2

Types of Biometric Traits


Run representation based methods
This technique is based on the horizontal and vertical run length
encoding from binary images. This process results in fast minutiae
extraction without requiring a computationally expensive thinning
process.

After the run-length encoding, the fingerprint images are depicted by


a cascade of runs and characteristic images are found out by
checking the runs adjacency of the runs. Not all characteristic runs
are true minutiae and their validity needs to be checked by some
geometric constraints.

22OBM103- BIOMETRICS AND ITS APPLICATION


31
UNIT-2

Types of Biometric Traits


Minutiae extraction from gray scale fingerprint images
Although it is still being researched, there are a number of
techniques to extract minutiae directly from gray scale fingerprint
images without binarization and thinning.
This extraction technique has a lot of relevance due to the following
reasons:
A great deal of information might be lost during the binarization
process.
Binarization and thinning are time consuming processes.
Binarization thinning operations introduce a large number of
spurious minutiae.
The binarization techniques do not prove to be particularly useful
when applied to low quality
22OBM103- images.AND ITS APPLICATION
BIOMETRICS
32
UNIT-2

Types of Biometric Traits


Minutiae extraction from gray scale fingerprint images
Although it is still being researched, there are a number of
techniques to extract minutiae directly from gray scale fingerprint
images without binarization and thinning.
This extraction technique has a lot of relevance due to the following
reasons:
A great deal of information might be lost during the binarization
process.
Binarization and thinning are time consuming processes.
Binarization thinning operations introduce a large number of
spurious minutiae.
The binarization techniques do not prove to be particularly useful
when applied to low quality
22OBM103- images.AND ITS APPLICATION
BIOMETRICS
33
UNIT-2

Ridge line following based minutiae extraction technique


One of the proposed methods in this technique is to directly extract
the minutiae from the gray scale image by following the ridge flow
lines with the help of local orientation field.

This method works on finding a local maximum relative to the


cross-section which is orthogonal to the ridge direction.

22OBM103- BIOMETRICS AND ITS APPLICATION


34
UNIT-2

Advantages of Fingerprint Biometrics:


Uniqueness: Fingerprint patterns are highly unique to individuals,
making them an excellent biometric trait for accurate identification.
Accuracy: Fingerprint recognition systems offer high accuracy
when properly implemented. The unique ridges and minutiae points
contribute to precise matching.
Non-intrusive: Capturing fingerprints is non-intrusive and does not
require physical contact, making it more user-friendly compared to
some other biometric modalities.
Widespread Acceptance: Fingerprint recognition has gained
widespread acceptance and is commonly used in various
applications, including access control, mobile devices, and law
enforcement.
22OBM103- BIOMETRICS AND ITS APPLICATION
35
UNIT-2

Ease of Use: Fingerprint scanning is easy and quick for users. It


doesn't involve complex actions, and individuals can be enrolled
and verified efficiently.
Reliability: Fingerprint patterns remain relatively stable over time,
providing a reliable means of identification.
Cost-Effectiveness: Fingerprint recognition systems can be
cost-effective to implement, especially in scenarios with a large user
base, such as access control for organizations.
Integration: Fingerprint biometrics can be integrated into existing
systems and applications, adding an additional layer of security.

22OBM103- BIOMETRICS AND ITS APPLICATION


36
UNIT-2

Disadvantages of Fingerprint Biometrics:


Environmental Factors: Fingerprints can be affected by
environmental factors such as dirt, moisture, and damage. This can
impact the accuracy of the recognition system.
Privacy Concerns: Some individuals may have concerns about the
storage and use of their biometric data, leading to privacy
considerations and potential resistance to adoption.
Sensor Quality: The accuracy of the system is dependent on the
quality of the fingerprint sensor. Low-quality sensors may result in
false positives or negatives.

22OBM103- BIOMETRICS AND ITS APPLICATION


37
UNIT-2

Hygiene Concerns: In public places, users may have hygiene


concerns about touching fingerprint sensors, especially during times
of health-related crises.
Potential for Spoofing: Although modern systems incorporate
anti-spoofing measures, there is still a potential risk of spoofing
through various means, such as high-quality replicas or latent
fingerprints left on surfaces.
Cultural and Religious Concerns: In some cultures or religions,
individuals may be uncomfortable with the physical contact required
for fingerprint scanning, leading to cultural or religious objections.

22OBM103- BIOMETRICS AND ITS APPLICATION


38
UNIT-2

Lack of Revocability: Unlike passwords or access cards,


fingerprints cannot be easily changed or revoked. If compromised, it
could pose a long-term security risk.
Legal and Ethical Issues: The use and storage of biometric data,
including fingerprints, may be subject to legal and ethical
regulations, requiring careful handling and compliance.

22OBM103- BIOMETRICS AND ITS APPLICATION


39
UNIT-3

FACE RECOGNITION
Background of face recognition:-
The dawn of Facial Recognition – 1960s
The earliest pioneers of facial recognition were Woody Bledsoe,
Helen Chan Wolf and Charles Bisson. In 1964 and 1965, Bledsoe,
along with Wolf and Bisson began work using computers to
recognise the human face.

Due to the funding of the project originating from an unnamed


intelligence agency, much of their work was never published.

22OBM103- BIOMETRICS AND ITS APPLICATION


40
UNIT-3

FACE RECOGNITION
However, it was later revealed that their initial work involved the
manual marking of various “landmarks” on the face such as eye
centres, mouth etc.

These were then mathematically rotated by a computer to


compensate for pose variation. The distances between landmarks
were also automatically computed and compared between images
to determine identity.

These earliest steps into Facial Recognition by Bledsoe, Wolf and


Bisson were severely hampered by the technology of the era, but it
remains an important first step in proving that Facial Recognition
was a viable22OBM103-
biometric. BIOMETRICS AND ITS APPLICATION
41
UNIT-3

FACE RECOGNITION
Advancing the accuracy of Facial Recognition – 1970s
Carrying on from the initial work of Bledsoe, the baton was picked
up in the 1970s by Goldstein, Harmon and Lesk who extended the
work to include 21 specific subjective markers including hair colour
and lip thickness in order to automate the recognition.

While the accuracy advanced, the measurements and locations still


needed to be manually computed which proved to be extremely
labour intensive yet still represents an advancement on Bledsoe’s
RAND Tablet technology.

22OBM103- BIOMETRICS AND ITS APPLICATION


42
UNIT-3

FACE RECOGNITION
Using linear algebra for Facial Recognition – 1980s/90s
It wasn’t until the late 1980s that we saw further progress with the
development of Facial Recognition software as a viable biometric for
businesses. In 1988, Sirovich and Kirby began applying linear
algebra to the problem of facial recognition.

A system that came to be known as Eigenface showed that feature


analysis on a collection of facial images could form a set of basic
features. They were also able to show that less than one hundred
values were required in order to accurately code a normalized facial
image.

22OBM103- BIOMETRICS AND ITS APPLICATION


43
UNIT-3

FACE RECOGNITION
In 1991, Turk and Pentland carried on the work of Sirovich and Kirby
by discovering how to detect faces within an image which led to the
earliest instances of automatic facial recognition. This significant
breakthrough was hindered by technological and environmental
factors, however, it paved the way for future developments in Facial
Recognition technology.

22OBM103- BIOMETRICS AND ITS APPLICATION


44
UNIT-3

FACE RECOGNITION
FERET Programme – 1990s/2000s
The Defence Advanced Research Projects Agency (DARPA) and
the National Institute of Standards and Technology (NIST) rolled out
the Face Recognition Technology (FERET) programme in the early
1990s in order to encourage the commercial facial recognition
market.
The project involved creating a database of facial images. Included
in the test set were 2,413 still facial images representing 856
people.
The hope was that a large database of test images for facial
recognition would inspire innovation and may result in more
powerful facial recognition technology.
22OBM103- BIOMETRICS AND ITS APPLICATION
45
UNIT-3

FACE RECOGNITION
Face Recognition Vendor Tests – 2000s
The National Institute of Standards and Technology (NIST) began
Face Recognition Vendor Tests (FRVT) in the early 2000s.

Building on FERET, FRVTs were designed to provide independent


government evaluations of facial recognition systems that were
commercially available, as well as prototype technologies.

These evaluations were designed to provide law enforcement


agencies and the U.S. government with the information necessary
to determine the best ways to deploy facial recognition technology.
22OBM103- BIOMETRICS AND ITS APPLICATION
46
UNIT-3

FACE RECOGNITION
Face Recognition Grand Challenge – 2006
Launched in 2006, the primary goal of the Face Recognition Grand
Challenge (FRGC) was to promote and advance face recognition
technology designed to support existing face recognition efforts in
the U.S. Government.

The FRGC evaluated the latest face recognition algorithms


available. High-resolution face images, 3D face scans, and iris
images were used in the tests.

The results indicated that the new algorithms were 10 times more
accurate than the face recognition algorithms of 2002 and 100 times
more accurate than those
22OBM103- of 1995, showing
BIOMETRICS AND ITS the advancements in
APPLICATION
facial recognition technology over the past decade. 47
UNIT-3

FACE RECOGNITION
Social Media – 2010-Current
Back in 2010, Facebook began implementing facial recognition
functionality that helped identify people whose faces may feature in
the photos that Facebook users update daily.

The feature was instantly controversial with the news media,


sparking a slew of privacy-related articles. However, Facebook
users by and large did not seem to mind.

Having no apparent negative impact on the website’s usage or


popularity, more than 350 million photos are uploaded and tagged
using face recognition each day.
22OBM103- BIOMETRICS AND ITS APPLICATION
48
UNIT-3

FACE RECOGNITION
iPhone X – 2017
Facial Recognition technology advanced rapidly from 2010 onwards
and September 12, 2017, was another significant breakthrough for
the integration of facial recognition into our day to day lives. This
was the date that Apple launched the iPhone X – the first iPhone
users could unlock with FaceID – Apple’s marketing term for facial
recognition.

22OBM103- BIOMETRICS AND ITS APPLICATION


49
UNIT-3

FACE RECOGNITION
NEC and Facial Recognition
Border controls, airlines, airports, transport hubs, stadiums,
mega-events, concerts, and conferences.

Biometrics is playing a growing role not only in the real-time policing


and securing of increasingly crowded and varied venues worldwide
but also in ensuring a smooth, enjoyable experience for the citizens
who visit them.
As a long-time committed pioneer of biometric research and
solutions, NEC has developed multi-modal technologies including
face, irisAs a long-time committed pioneer of biometric research and
solutions, NEC has developed multi-modal technologies including
face, iris and voice recognition,
22OBM103- finger
BIOMETRICS ANDand
ITSpalmprint identificationAs
APPLICATION
a long-time committed pioneer of biometric research and solutions, 50
UNIT-3

FACE RECOGNITION
Face recognition can often prove one of the best biometrics
because images can be taken without touching or interacting with
the individual being identified, and those images are recorded and
instantly checked against existing databases.

NEC’s face recognition offers high-performance, scalable solutions


for the most demanding real-time or post-event requirements. With
face recognition, search, identification and verification functions all
on a single platform, it can be easily integrated into existing security
systems to extract faces in real-time, match them against an
existing database or watchlist and produces real-time alerts to help
reduce public safety risks.
22OBM103- BIOMETRICS AND ITS APPLICATION
51
UNIT-3

FACE RECOGNITION
With the ability to process and analyse multiple camera feeds and
thousands of faces per minute, NEC’s powerful face recognition is
able to address the largest and most difficult security challenges
with unparalleled efficiency, sensitivity, and perception.

22OBM103- BIOMETRICS AND ITS APPLICATION


52
UNIT-3

FACE RECOGNITION
With the ability to process and analyse multiple camera feeds and
thousands of faces per minute, NEC’s powerful face recognition is
able to address the largest and most difficult security challenges
with unparalleled efficiency, sensitivity, and perception.

22OBM103- BIOMETRICS AND ITS APPLICATION


53
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
A throughout survey has revealed that various methods and
combination of these methods can be applied in development of a
new face recognition system.

Among the many possible approaches, we have decided to use a


combination of knowledge-based methods for face detection part
and neural network approach for face recognition part.

The main reason in this selection is their smooth applicability and


reliability issues. Our face recognition system approach is given in
Figure 2.
22OBM103- BIOMETRICS AND ITS APPLICATION
54
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM

22OBM103- BIOMETRICS AND ITS APPLICATION


55
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
This stage consists of image capturing, database of faces, feature
extraction process and another database for storing templates of
features extracted using a suitable training method.

Image enhancement is also a part of face recognition system so that


poor image quality should not affect the performance of the system
adversely.
In the recognition stage, either verification or identification or both
are done. In this stage, an input facial image is subjected to
recognition by using suitable user interface. This is passed through
an image enhancement method.
22OBM103- BIOMETRICS AND ITS APPLICATION
56
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
The features are extracted and converted into equivalent
representations as templates.

These templates are compared with the template databases where


the templates of already trained images are stored. The matching of
templates results in the image being recognized.

Figure 3.2 shows recognition stage of face biometrics. In both


enrollment and recognition stages, preprocessing may be required
to make an image being subjected to computer-based algorithm.
22OBM103- BIOMETRICS AND ITS APPLICATION
57
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
The features are extracted and converted into equivalent
representations as templates.

These templates are compared with the template databases where


the templates of already trained images are stored. The matching of
templates results in the image being recognized.

Figure 3.2 shows recognition stage of face biometrics. In both


enrollment and recognition stages, preprocessing may be required
to make an image being subjected to computer-based algorithm.
22OBM103- BIOMETRICS AND ITS APPLICATION
58
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM

22OBM103- BIOMETRICS AND ITS APPLICATION


59
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
As discussed in the previous section, there are several ways in
which face recognition systems can be designed. They are based
on templates or the training methods used. Some of the face
recognition systems are template and geometrical feature-based
methods.

Few examples of template-based methods include Eigen faces,


Fisher faces, methods based on PCA, HMM, discrete cosine
transform (DCT), SVM, Fourier transform, etc. These methods
require certain features of faces which are extracted by training
them, using any one of these methods. Sometimes, hybrid approach
is also used22OBM103-
to extract features.
BIOMETRICS AND ITS APPLICATION
60
UNIT-3

FACE RECOGNITION
DESIGN OF A FACE RECOGNITION SYSTEM
Other methods under template-based methods use NNs in different
ways.

The different ways decide different methods, such as single layer,


multilayer NN methods, self-organizing map method or Hopfield
method. Another category of face recognition systems is using some
geometric features, such as wavelet

22OBM103- BIOMETRICS AND ITS APPLICATION


61
UNIT-3

FACE RECOGNITION
neural network for face Recognition

The NN is an interconnection of neurons which are cells the


network. the computer simulation of human brain. It is just like a
human brain, consisting of millions of neurons connected together in
different layers.
When a human brain receives ideas, the thinking process starts.
The idea is then selected and decision is made and is converted
into action. The NN is arranged in different layers: input, output and
hidden. There may be multiple hidden layers.
Each layer has a certain number of nodes or neurons with some
value of weight. For example, whenever we are in deep thought or
in trouble, we place ourBIOMETRICS
22OBM103- fingers on our
ANDhead
ITS and try to press the fore
APPLICATION
head and then act. In a similar manner, the NN is trained. 62
UNIT-3

FACE RECOGNITION
neural network for face Recognition

Each neuron has certain weight and whenever e input is received by


the network and passed through the layers, weights of the neurons
are adjusted to the obtain a specified output.

The comparison of human brain with NN is also known as "neural


network as human brain analogy".

The NN is a soft computing tool among fuzzy logic, artificial


intelligence, genetic algorithm, etc. The network is trained and used
to extract the features of the images.
22OBM103- BIOMETRICS AND ITS APPLICATION
63
UNIT-3

FACE RECOGNITION
neural network for face Recognition

Because of human brain analogy or cognitive capability of brain, NN


has got large applications in biometrics implementations. The
resemblance of NN with the brain of a human being can be
explained as follows:

1. We gain knowledge through training. Similarly, some amount of


knowledge is acquired by the computer or a system by learning and
training of NN.
2. Neurons correspond to nodes in NN.
3. Each neuron has an activation or threshold function, and
accordingly 22OBM103-
decision is made. This is
BIOMETRICS similar
AND to a brain that decides
ITS APPLICATION
either to do a particular task or not. 64
UNIT-3

FACE RECOGNITION
neural network for face Recognition

The NN is characterized by four important components:


preprocessing, training, decision and assessment, like brain. One
does not accept all the inputs and some filtering is always applied.
Using some logic, one thinks about it and then a decision is taken.

All inputs subjected to a NN are preprocessed and when accepted


by the network, training process starts and a decision is taken.
Sometimes, decision is required to be changed and by adjusting the
weights of nodes, it can be done. The decision made can also be
evaluated by NN.
22OBM103- BIOMETRICS AND ITS APPLICATION
65
UNIT-3

FACE RECOGNITION
neural network for face Recognition

The simplest NN can be expressed as a single neuron consisting of


an input signal x, a weight matrix W. output & and an activation
function f(s). The choice of activation function is usually a sigmoid
function or hyperbolic tangent function.

The network may learn non-linear patterns because of the


non-linear activation function used. Figure 3.3 shows a model of NN
with input x and final output y. The output of neuron is a = WX = X₁
W₁ + X₂ W₂+...+XW. Based on the activation function used, the final
decision is produced as y.
22OBM103- BIOMETRICS AND ITS APPLICATION
66
UNIT-3

FACE RECOGNITION
neural network for face Recognition

22OBM103- BIOMETRICS AND ITS APPLICATION


67
UNIT-3

FACE RECOGNITION
neural network for face Recognition

The NNs have several layers of neurons called hidden layers. The
number of neurons may be different for different hidden layers and
these layers are chosen depending on the specific data to learn.

A multilayer NN is usually employed in face recognition. The


network is used to train the images and some features of faces are
extracted. The NN may have a number of hidden layers consisting
of certain number of neurons.

22OBM103- BIOMETRICS AND ITS APPLICATION


68
UNIT-3

FACE RECOGNITION
neural network for face Recognition

Network architectures may be used in real-time applications for


access control using face image. The weights of neurons are
adjusted by a suitable training procedure called learning of NN.

Artificial NNs are influenced by a human brain that consists of


neurons.

The NNs are extensively used in computer vision, signal processing,


pattern recognition, feature extraction and applications based on
classification. The common factor in brain and NN is the ability to
learn and recognize
22OBM103-patterns.
BIOMETRICS AND ITS APPLICATION
69
UNIT-3

FACE RECOGNITION
Training of Neural Network
Like a brain is trained and weights of neurons are adjusted to give a
decision, the NNs are trained so that learning mechanism takes
place in the network. There are many different techniques for
training the NNs, but they are mainly categorized into two
types-supervised and unsupervised learning. The NNs can be
operated in an adaptive manner just by adjusting the weights of
neurons or nodes of the network.

22OBM103- BIOMETRICS AND ITS APPLICATION


70
UNIT-3

FACE RECOGNITION
Weights of neurons are modified by a learning process.
Usually, in order to get the desired response of network or desired
output, the w weights need to be modified.
The process of changing weights for getting the desired output is
called learning of network.
Again, the learning can be compared to a human brain. Many things
are not taught to us- children learn from the environment without
being taught by anyone.
This is called unsupervised learning. When we are taught the
learning process by someone else, then it is called supervised
learning.

22OBM103- BIOMETRICS AND ITS APPLICATION


71
UNIT-3

FACE RECOGNITION
Weights of neurons are modified by a learning process.
Usually, in order to get the desired response of network or desired
output, the w weights need to be modified.
The process of changing weights for getting the desired output is
called learning of network.
Again, the learning can be compared to a human brain. Many things
are not taught to us- children learn from the environment without
being taught by anyone.
This is called unsupervised learning. When we are taught the
learning process by someone else, then it is called supervised
learning.

22OBM103- BIOMETRICS AND ITS APPLICATION


72
UNIT-3

FACE RECOGNITION
An overview of NN and how it helps in biometrics or computer vision
has been answered using a practical example has been discussed
in this chapter.
1.Supervised learning: A single neuron can be trained by an
ordinary least mean square algorithm. This algorithm
minimizes the mean square error and calculates the inverse of
an auto-correlation matrix.
2. If the network is multilayered consisting of many hidden
layers, then the most popularly used back propagation
algorithm can be used.

22OBM103- BIOMETRICS AND ITS APPLICATION


73
UNIT-3

FACE RECOGNITION
The back propagation algorithm has a slow convergence
speed and is therefore affected by many variations and
improvements, such as variable learning speed.
Weights of neurons are adjusted going back to previous
layers so as to achieve the desired output. The supervised
learning or training is used to obtain a desired output.

22OBM103- BIOMETRICS AND ITS APPLICATION


74
UNIT-3

FACE RECOGNITION
2. Unsupervised learning: Unsupervised learning is often used
when there is no target signal.

Clustering based applications require this type of competitive


learning. Self-organized map is a very popular.

method of unsupervised learning. The neurons are spread over


the signal space and a winning neuron act as a classifier of the
input.

The PCA is also used as a learning method to find principal


components or Eigen vectors in compression and recognition
applications.
22OBM103- BIOMETRICS AND ITS APPLICATION
75
UNIT-3

FACE RECOGNITION
Detecting faces in video sequences involves applying face detection
algorithms to each frame of the video. One popular approach is to
use a method like the Viola-Jones algorithm or deep learning-based
methods like Single Shot Multibox Detector (SSD), Faster R-CNN,
or You Only Look Once (YOLO).
Here's a general overview of how you could approach face
detection in video sequences:
Frame Extraction: Extract frames from the video sequence at a
certain frame rate or keyframes.
Face Detection: Apply a face detection algorithm to each frame.
This could be a classical method like Haar cascades (used in
Viola-Jones) or a deep learning-based approach like SSD, Faster
R-CNN, or YOLO.
22OBM103- BIOMETRICS AND ITS APPLICATION
76
UNIT-3

FACE RECOGNITION
Tracking (Optional): Optionally, you might want to incorporate
tracking algorithms to track faces across frames to improve
robustness and efficiency. Popular tracking algorithms include
Kalman filters, Particle filters, or Deep SORT.

Post-processing (Optional): Perform post-processing steps like


non-maximum suppression to remove duplicate detections or
filtering based on size or position to refine the detections.

22OBM103- BIOMETRICS AND ITS APPLICATION


77
UNIT-3

FACE RECOGNITION
Visualization/Output: Visualize the detected faces in the video
frames or output the results in a desired format (e.g., bounding box
coordinates, timestamps, etc.).

It's essential to choose the right balance between accuracy and


computational efficiency depending on your specific requirements
and constraints.

Deep learning-based methods usually offer better accuracy but


might be computationally more expensive compared to classical
methods.

22OBM103- BIOMETRICS AND ITS APPLICATION


78
UNIT-3

FACE RECOGNITION
Visualization/Output: Visualize the detected faces in the video
frames or output the results in a desired format (e.g., bounding box
coordinates, timestamps, etc.).

It's essential to choose the right balance between accuracy and


computational efficiency depending on your specific requirements
and constraints.

Deep learning-based methods usually offer better accuracy but


might be computationally more expensive compared to classical
methods.

22OBM103- BIOMETRICS AND ITS APPLICATION


79
UNIT-3

FACE RECOGNITION
1. Frame Extraction:
Before detecting faces in a video sequence, you need to break it
down into individual frames.

This process involves extracting frames from the video at regular


intervals or based on keyframes.

Most video processing libraries provide tools for frame extraction,


allowing you to specify the frame rate or exact timestamps for
extraction.

22OBM103- BIOMETRICS AND ITS APPLICATION


80
UNIT-3

FACE RECOGNITION
2. Face Detection:
Once you have extracted frames from the video, you apply a face
detection algorithm to each frame. Here are some common
approaches:

Haar Cascades (Viola-Jones Algorithm): This is a classic method


for face detection. It works by using a set of pre-trained classifiers to
identify features (Haar-like features) in an image and then using a
cascade of these classifiers to detect faces.

22OBM103- BIOMETRICS AND ITS APPLICATION


81
UNIT-3

FACE RECOGNITION
Deep Learning-Based Methods:
Single Shot Multibox Detector (SSD): SSD is a deep
learning-based object detection algorithm. It predicts bounding
boxes and class labels for multiple objects in a single pass
through the network.
Faster R-CNN: Faster R-CNN is another deep learning-based
object detection algorithm. It consists of two main modules: a
region proposal network (RPN) for generating region proposals
and a network for detecting objects within those proposals.
You Only Look Once (YOLO): YOLO is a real-time object
detection system. It divides the input image into a grid and
predicts bounding boxes and probabilities for each grid cell.
22OBM103- BIOMETRICS AND ITS APPLICATION
82
UNIT-3

FACE RECOGNITION
Deep Learning-Based Methods:
Single Shot Multibox Detector (SSD): SSD is a deep
learning-based object detection algorithm. It predicts bounding
boxes and class labels for multiple objects in a single pass
through the network.
Faster R-CNN: Faster R-CNN is another deep learning-based
object detection algorithm. It consists of two main modules: a
region proposal network (RPN) for generating region proposals
and a network for detecting objects within those proposals.
You Only Look Once (YOLO): YOLO is a real-time object
detection system. It divides the input image into a grid and
predicts bounding boxes and probabilities for each grid cell.
22OBM103- BIOMETRICS AND ITS APPLICATION
83
UNIT-3

FACE RECOGNITION
3. Tracking (Optional):
While face detection in individual frames can be effective, it may not
maintain consistency across frames. Therefore, incorporating
tracking algorithms can be beneficial. Tracking algorithms help
associate detections in successive frames, enabling you to track
faces over time.

22OBM103- BIOMETRICS AND ITS APPLICATION


84
UNIT-3

FACE RECOGNITION
Common tracking algorithms include:
Kalman Filters: Kalman filters are recursive estimators that use a
series of measurements observed over time to estimate the state of
a process.
Particle Filters: Particle filters, also known as sequential Monte
Carlo methods, represent the posterior density by a set of random
samples, known as particles.
Deep SORT (Simple Online and Realtime Tracking with a Deep
Association Metric): Deep SORT combines deep learning-based
object detection with a Kalman filter-based tracking algorithm.

22OBM103- BIOMETRICS AND ITS APPLICATION


85
UNIT-3

FACE RECOGNITION
4. Post-processing (Optional):
After face detection and possibly tracking, you may want to perform
post-processing steps to refine the results. This could include:
Non-maximum Suppression: This technique is used to eliminate
redundant detections by keeping only the most confident ones
among overlapping bounding boxes.
Size or Position Filtering: Filtering out detections based on their
size or position within the frame can help remove false positives or
irrelevant detections.

22OBM103- BIOMETRICS AND ITS APPLICATION


86
UNIT-3

FACE RECOGNITION
5. Visualization/Output:
Finally, you can visualize the detected faces in the video frames by
drawing bounding boxes around them or output the results in a
desired format, such as bounding box coordinates, timestamps, or
labels.

By following these steps and selecting appropriate algorithms and


parameters, you can effectively detect faces in video sequences.
The choice of approach depends on factors such as accuracy
requirements, computational resources, and real-time processing
constraints.

22OBM103- BIOMETRICS AND ITS APPLICATION


87
UNIT-3

FACE RECOGNITION
Face recognition methods
Aim to identify or verify individuals by analyzing and comparing their
facial features. Here are some common approaches to face
recognition:
1. Eigenfaces:
Eigenfaces is a classic method for face recognition that uses
Principal Component Analysis (PCA) to represent faces as vectors
in a high-dimensional space. It reduces the dimensionality of face
images and extracts a set of "eigenfaces" that represent the
principal components of facial variation. Face recognition is then
performed by comparing the projections of a face onto the
eigenfaces.
22OBM103- BIOMETRICS AND ITS APPLICATION
88
UNIT-3

FACE RECOGNITION
Face recognition methods
2. Fisherfaces (Linear Discriminant Analysis):
Similar to Eigenfaces, Fisherfaces uses dimensionality reduction
techniques like Linear Discriminant Analysis (LDA) to find a set of
discriminant features that maximize the ratio of between-class
variance to within-class variance.

This results in a more discriminative representation of faces, which


can improve recognition accuracy compared to Eigenfaces.

22OBM103- BIOMETRICS AND ITS APPLICATION


89
UNIT-3

FACE RECOGNITION
Face recognition methods
2. Fisherfaces (Linear Discriminant Analysis):
Similar to Eigenfaces, Fisherfaces uses dimensionality reduction
techniques like Linear Discriminant Analysis (LDA) to find a set of
discriminant features that maximize the ratio of between-class
variance to within-class variance.

This results in a more discriminative representation of faces, which


can improve recognition accuracy compared to Eigenfaces.

22OBM103- BIOMETRICS AND ITS APPLICATION


90
UNIT-3

FACE RECOGNITION
Face recognition methods
3. Local Binary Patterns (LBP):
Local Binary Patterns is a texture descriptor that characterizes the
local structure of images. In face recognition, LBP extracts texture
features from facial images by comparing the intensity of pixels with
neighboring pixels. These features are then used for classification or
matching faces.

22OBM103- BIOMETRICS AND ITS APPLICATION


91
UNIT-3

FACE RECOGNITION
Face recognition methods
3. Local Binary Patterns (LBP):
Local Binary Patterns is a texture descriptor that characterizes the
local structure of images. In face recognition, LBP extracts texture
features from facial images by comparing the intensity of pixels with
neighboring pixels. These features are then used for classification or
matching faces.

22OBM103- BIOMETRICS AND ITS APPLICATION


92
UNIT-3

FACE RECOGNITION
Face recognition methods
4. Histogram of Oriented Gradients (HOG):
HOG is another feature descriptor commonly used in object
detection and recognition tasks, including face recognition. It
computes the distribution of gradient orientations in localized
portions of an image. In face recognition, HOG descriptors capture
the spatial layout of facial features such as edges and corners.

22OBM103- BIOMETRICS AND ITS APPLICATION


93
UNIT-3

FACE RECOGNITION
Face recognition methods
5. Convolutional Neural Networks (CNNs):
Deep learning-based approaches, particularly Convolutional Neural
Networks (CNNs), have gained significant popularity in recent years
for face recognition tasks.

CNNs can automatically learn hierarchical representations of facial


features directly from raw pixel data. Architectures like VGG,
ResNet, and Siamese networks have been adapted for face
recognition tasks with impressive results.

22OBM103- BIOMETRICS AND ITS APPLICATION


94
UNIT-3

FACE RECOGNITION
Face recognition methods
6. Siamese Networks:
Siamese networks are neural networks with shared weights that
take pairs of images as input and learn to output a similarity score
indicating whether the images belong to the same person or not.

They are commonly used for one-shot or few-shot face recognition


tasks, where limited training data is available for each individual.

22OBM103- BIOMETRICS AND ITS APPLICATION


95
UNIT-3

FACE RECOGNITION
Face recognition methods
7. Triplet Loss:
Triplet loss is a loss function used in training siamese networks and
other deep learning-based face recognition models. It minimizes the
distance between the embeddings of matching pairs of images while
maximizing the distance between embeddings of non-matching
pairs, encouraging the network to learn discriminative features.
8. Ensemble Methods:
Ensemble methods combine multiple base classifiers to improve
overall recognition performance. Techniques such as bagging,
boosting, and stacking can be applied to face recognition models to
enhance robustness and generalization.
22OBM103- BIOMETRICS AND ITS APPLICATION
96
UNIT-3

FACE RECOGNITION
Face recognition methods
Principal Component Analysis (PCA) is a dimensionality reduction
technique commonly used in various fields, including image
processing, pattern recognition, and data analysis. PCA aims to
transform high-dimensional data into a lower-dimensional space
while preserving most of the variance in the data. It achieves this by
identifying the directions, or principal components, along which the
data varies the most..

22OBM103- BIOMETRICS AND ITS APPLICATION


97
UNIT-3

FACE RECOGNITION
Face recognition methods
1. Data Standardization:
PCA requires that the data is standardized (i.e., centered around
zero with unit variance) since it's sensitive to the scale of the
features.
2. Covariance Matrix Computation:
Given a dataset with �m samples and �n features, PCA computes
the covariance matrix ΣΣ of the data. The covariance between two
features measures how they vary together.
Σ=1����Σ=m1XTX
where �X is the �×�m×n data matrix with each row representing
a sample and each column representing a feature.
22OBM103- BIOMETRICS AND ITS APPLICATION
98
UNIT-3

FACE RECOGNITION
Face recognition methods
3. Eigendecomposition:
PCA then performs an eigendecomposition of the covariance matrix
ΣΣ to find its eigenvectors and eigenvalues.
The eigenvectors represent the directions (principal components)
along which the data varies the most, and the eigenvalues represent
the amount of variance explained by each principal component.

22OBM103- BIOMETRICS AND ITS APPLICATION


99
UNIT-3

FACE RECOGNITION
Face recognition methods
4. Principal Component Selection:
The principal components are sorted in descending order of their
corresponding eigenvalues. The principal component with the
highest eigenvalue explains the most variance in the data, followed
by the second principal component, and so on.
Typically, only a subset of the principal components that capture
most of the variance (e.g., 95%) is retained.

22OBM103- BIOMETRICS AND ITS APPLICATION


100
UNIT-3

FACE RECOGNITION
Face recognition methods
5. Projection:
Finally, PCA projects the original data onto the selected principal
components to obtain the lower-dimensional representation.

The projected data can be obtained by multiplying the original data


matrix �X by the matrix of selected principal components.

22OBM103- BIOMETRICS AND ITS APPLICATION


101
UNIT-3

FACE RECOGNITION
Face recognition methods
PCA is widely used for various purposes, including:

Dimensionality reduction: By selecting only the most important


principal components, PCA reduces the dimensionality of the data,
making it easier to visualize and analyze.

Feature extraction: The principal components can serve as new


features that capture the most significant variations in the data.

Noise reduction: Since PCA focuses on capturing the most


significant variations, it can help reduce the impact of noisy or
irrelevant features.
22OBM103- BIOMETRICS AND ITS APPLICATION
102

You might also like