18CSE357T – BIOMETRICS

Unit –2 : Session –6 : SLO -2

SRM Institute of Science and Technology 1

Face detection, feature extraction and
matching
 3
Motivation

Face is our primary focus of interaction with society, face

communicates identity, emotion, race and age. It is also
quite useful for judging gender, size and perhaps some
of the characteristics of the person.
 Applications [8], [12], [13],
4

[14]
Area Applications
Information Security Access Security (OS, Database)
Data Privacy (e.g. medical records)
User authentications (trading, on-line banking)
Access Management Secure Access Authentications (restricted facilities)
Permission based system
Access log or Audit Trails
Biometrics Personal identification (national IDs, passport, voter
registration, driver licenses)
Automated identity verification (boarder control)
Law Enforcement Video
Surveillance,
Suspect
identification
Suspect tracking (
investigation ) Simulated
aging
Forensic Reconstruction of faces from
remains, Tracing missing children.
Personal Security Home video surveillance system
Expression interpretation (driver monitoring system)
Entertainment Home video game
 Introducti 5

Face
on
Detection & Recognition
by
Humans

•Human brain is trained for

face detection and recognition.
Face detection and recognition
is an easy task for humans [1].

•Experimentally it has been

found that even one to three
day old babies are able to
distinguish between known
faces [2].

So how hard could it be for a computer ?

 6

Face Detection
• Identifies human faces in digital images.
• Identifies the pixels which represent the face in a given
image.
• Also referred to as the pre-processing phase.
 7

Face Recognition
• Identifying a face match.

Vijay Sharma
 8

Face Recognition
“Face Recognition is the task
of identifying an already
detected face as a KNOWN
or UNKNOWN face, and in
more advanced cases,
TELLING EXACTLY
WHO’S IT IS ! “ [8]

Face recognition problem statement:

Given still or video images of a scene, identify or verify one or more persons in the scene
using a stored database of faces.
 9

How it works?

Face Features Face

Detection Extraction Recognition

Figure: [8]

Features [9]
1. Distance between the eyes
2. Width of the nose
3. Depth of the eye socket
4. Cheekbones
5. Jaw line
6. Chin
 10

Face Detection using Haar Cascades [3],

[10]

• Devised by Paul Viola and Michael Jones in 2001.

• Robust and very quick.
• 15 times quicker than any technique at the time of release.
• Could be operated in real-time.
• (95% accuracy at around 17 fps.)
• Feature extraction and feature evaluation.
(Rectangular features are used)
• With a new image representation their calculation is very fast.
• Classifier training and feature selection using a method called
AdaBoost. (A long and exhaustive training process)
• A degenerate decision tree of classifiers is formed.
 11

Features [3],
[10]

Four basic types:

• They are easy to calculate.
• The white areas are subtracted from the black ones.

Edge Feature

Line Feature
 12

Features Extraction [3],

[10]
 13

Challenges in Haar Cascades [10]

• Variations in pose Head positions, frontal view, profile

view and head tilt, facial expressions.

• Illumination Changes Light direction and intensity changes,

cluttered background, low quality
images.

• Camera Parameters Resolution, color balance etc.

• Occlusion Glasses, facial hair and makeup.

 Advantages & Disadvantages
14

[11]

Advantages:

• High detection accuracy.

• Low false positive rate.

Disadvantages:

• Computationally complex and slow.

• Longer training time.
• Less accurate on black faces.
• Limitation in difficult lightening conditions.
• Less robust to occlusion/obstacle.
 Template creation

03/12/13 1
5
 • Enrollment templates are normally created from
a multiplicity of processed facial images.
• These templates can vary in size from less than
100 bytes, generated through certain vendors
and to over 3K for templates.
• The 3K template is by far the largest among
technologies considered physiological biometrics.
• Larger templates are normally associated with
behavioral biometrics,

03/12/13 1
6
 Template matching
• It compares match templates against enrollment
templates.
• A series of images is acquired and scored against
the enrollment, so that a user attempting 1:1
verification within a facial-scan system may have
10 to 20 match attempts take place within 1 to 2
seconds.
• facial-scan is not as effective as finger-scan or iris-
scan in identifying a single individual from a large
database, a number of potential matches are
generally returned after large-scale facial-scan
identification searches.
03/12/13 1
7
 How Facial Recognition System Works

• Facial recognition software is based on the ability to

first recognize faces, which is a technological feat in
itself. If you look at the mirror, you can see that your
face has certain distinguishable landmarks. These are
the peaks and valleys that make up the different
facial features.
• VISIONICS defines these landmarks as nodal points.
There are about 80 nodal points on a human face.

03/12/13 1
8
 Contd..
Here are few nodal points that are measured by the
software.
1. distance between the eyes
2. width of the nose
3. depth of the eye socket
4. cheekbones
5. jaw line
6. chin

03/12/13 1
9
 SOFTWARE
▪ Detection- when the system is attached to a video
surveilance system, the recognition software searches
the field of view of a video camera for faces. If there is
a face in the view, it is detected within a fraction of a
second. A multi-scale algorithm is used to search for
faces in low resolution. The system switches to a high-
resolution search only after a head-like shape is
detected.
▪ Alignment- Once a face is detected, the system
determines the head's position, size and pose. A face
needs to be turned at least 35 degrees toward the
camera for the system to register it.
03/12/13 2
0
 ▪ Normalization-The image of the head is scaled and
rotated so that it can be registered and mapped into
an appropriate size and pose. Normalization is
performed regardless of the head's location and
distance from the camera. Light does not impact the
normalization process.
▪ Representation-The system translates the facial data
into a unique code. This coding process allows for
easier comparison of the newly acquired facial data to
stored facial data.
▪ Matching- The newly acquired facial data is
compared to the stored data and (ideally) linked to at
least one stored facial representation.
03/12/13 2
1
 ▪ The system maps the face and creates a
faceprint, a unique numerical code for that face.
Once the system has stored a faceprint, it can
compare it to the thousands or millions of
faceprints stored in a database.
▪ Each faceprint is stored as an 84-byte file.

03/12/13 2
2
 Strengths

▪ It has the ability to leverage existing image

acquisition equipment.
▪ It can search against static images such as driver’s
license photographs.
▪ It is the only biometric able to operate without user
cooperation.

03/12/13 2
3
 Weaknesses
• reduce matching accuracy.
▪ Changes in acquisition environment
▪ Changes in physiological characteristics
reduce matching accuracy.
• It has the potential for privacy abuse due to
noncooperative enrollment and identification
capabilities.

03/12/13 2
4
 Applications
▪ Security/Counterterrorism. Access control, comparing
surveillance images to Know terrorist.
▪ Day Care: Verify identity of individuals picking up the
children.
▪ Residential Security: Alert homeowners of
approaching personnel
▪ Voter verification: Where eligible politicians are
required to verify their identity during a voting
process this is intended to stop voting where the vote
may not go as expected.
▪ Banking using ATM: The software is able to quickly
verify a customer’s face.
03/12/13 2
5