Face Recognisation technology

8/7/2019 Face Recognisation technology

1/31

Made By : Navin Gupta


2/31

The information age is quicklyrevolutionizing the way transactions arecompleted. Everyday actions areincreasingly being handledelectronically, instead of with pencil andpaper or face to face.

This growth in electronic transactionshas resulted in a greater demand for fastand accurate user identification andauthentication.


3/31

A ccess codes for buildings, banksaccounts and computer systems often

use PIN's for identification and securityclearences. Using the proper PIN gains access, but

the user of the PIN is not verified. When

credit and ATM cards are lost or stolen,an unauthorized user can often come upwith the correct personal codes.

Face recognition technology may solvethis problem since a face is undeniablyconnected to its owner expect in thecase of identical twins.


4/31

Biometrics A biometric is a unique, measurable

characteristic of a human being that can beused to automatically recognize an individualor verify an individuals identity.

Biometrics can measure both physiologicaland behavioral characteristics.

Physiological biometrics -based onmeasurements and data derived from direct

measurement of a part of the human body. Behavioral biometrics -based on

measurements and data derived from anaction.


5/31

TYPES OF BIOMETR ICS

PHYSIOLO G ICAL

a.Finger-scanb. Facial Recognitionc. Iris-scand. Retina-scan

e. H and-scan

BE AVIO RAL

a. Voice-scanb. Signature-scanc. Keystroke-scan


6/31

WH Y W E CH OOSE FA CERE COGN ITION TE CH NOLOGY

It requires no physical interaction on behalf of the user.

It is accurate and allows for high enrolment

and verification rates. It does not require an expert to interpret the

comparison result. It can use your existing hardware

infrastructure, existing camaras and imagecapture Devices will work with no problems

It is the only biometric that allow you toperform passive identification in a one to.


7/31

FA CE RE COGN ITION

face recognition there are two types of comparisons .

VERIFIC ATION-in this the systemcompares the given individual with whothat individual says they are and gives ayes or no decision.

IDENTIFIC ATION- in this the systemcompares the given individual to all theOther individuals in the database andgives a ranked list of matches.


8/31

All identification or authentication technologiesoperate using the following four stages:

Capture: A physical or behavioural sample iscaptured by the system during Enrollment andalso in identification or verification process.

Extraction: unique data is extracted from thesample and a template is created. Comparison: the template is then compared

with a new sample. Match/non match: the system decides if the

features extracted from the new Samples area match or a non match.


9/31

CAPTUR ING OF IMAGE BY STANDARDVIDEO

CAMERAS The image is optical in characteristics and may be thought

of as a collection of a large number of bright and darkareas representing the picture details.

A t an instant there will be large number of picture detailsexisting simultaneously each representing the level of brightness of the scene to be reproduced.

Therefore it would require infinite number of channels totransmit optical information corresponding to pictureelements simultaneously.

There is practical difficulty in transmitting all informationsimultaneously so we use a method called scanning. the conversion of optical information to electrical form and

its transmission is carried out element by element one ata time in a sequential manner to cover the entire image.


10/31

W ORK ING OF VED IO CAMERA

A TV camera converts optical information intoelectrical information, the amplitude of whichvaries in accordance with variation of brightness.

A n optical image of the scene to betransmitted is focused by lense assembly onthe rectangular glass plate of the camera tube.

The inner side of this has a transparentcoating on which is laid a very thin layer of photoconductive material. The photolayer hasvery high resistance when no light is falling onit but decreases depending on the intensity of light falling on it.


11/31

A n electron beam is formed by an electrongun in the TV camera tube.

This beam is used to pick up the pictureinformation now avilable on the target plateof varying resistace at each point.

The electron beam is deflected by a pair of deflecting coils mounted on the glassenvelope and kept mutually perpendicular to

each other to achive scanning of the entiretarget area.


12/31

The deflecting coils are fed seperatelyfrom two sweep oscillators, eachoperating at different frequencies.

The magnetic deflection caused bycurrent in one coil gives horizontalmotion to the beam from left to right ata uniform rate and brings it back to theleft side to commence the trace of the

next line. The other coil is used to deflect the

beam from top to bottom.


13/31

A s the beam moves from element to element itencounters different resistance across thetarget plate depending on the resistance of thephotoconductive coating.

The result is flow of current which varies inmagnitude as elements are scanned.

The current passes through the loadresistance Rl connected to conductive coatingon one side of the DC supply source on theother.

Depending on the magnitude of current avarying voltage appears across the resistanceRl and this corresponds to the opticalinformation of the picture


14/31

PERFORMAN CE

False acceptance rate (F A R) - Theprobability that a system will incorrectlyidentify an individual or will fail to reject an

imposter. It is also called as type 2 error rateFA R= NF A /NIIA

Where

NFA = number of false acceptanceNIIA = number of imposter identification

attempts


15/31

False rejection rates (FRR) - Theprobability that a system will fail toidentify an enrollee. It is also called type1 error rate.

FRR= NFR/NEI A where NFR= number of false rejection rates NEIA = number of enrollee identification

attempt


16/31

Response time: The time period required bya biometric system to return a decision onidentification of a sample.

decision Threshold: The acceptance or rejection of a data is dependent on the matchscore falling above or below the threshold. Thethreshold is adjustable so that the system canbe made more or less strict; depending on therequirements of any given application.

Enrollment time: The time period a personmust spend to have his/her facial reference

templatesuccessfully created.


17/31

IMPLEMENTAT ION OF FA CERE COGN ITION TE CH NOLOGY

The implementation of face recognitiontechnology includes the following four stages:

Data acquisition Input processing Face image classification decision making


18/31

Da ta a cquisition

The input can be recorded video of thespeaker or a still image. A sample of 1sec duration consists of a 25 framevideo sequence.

More than one camera can be used toproduce a 3D representation of the face

and to protect against the usage of photographs to gain unauthorizedaccess.


19/31

Input processing

A pre-processing module locates the eyeposition and takes care of the surroundinglighting condition and colour variance.

First the presence of faces or face in ascene must be detected. Once the face isdetected, it must be localized andNormalization process may be required tobring the dimensions of the live facialsample in alignment with the one on thetemplate.


20/31

Fa ce im a ge cla ssific a tion

Some facial recognition approaches use the whole facewhile others concentrate on facial components and/ or regions (such as lips, eyes etc).

The appearance of the face can change considerably during

speech and due to facial expressions. In particular the mouthis subjected to fundamental changes but is also veryimportant source for discriminating faces.

So an approach to persons recognition is developed basedon patio- temporal modeling of features extracted fromtalking face. Models are trained specific to a persons speecharticulate and the way that the person speaks.


21/31

Person identification is performed by tracking mouth

movements of the talking face and by estimating thelikelyhood of each model of having generated the observedsequence of features.

The model with the highest likelyhood is chosen as therecognized person. Synergetic computer are used to classifyoptical and audio features, respectively.

A synergetic computer is a set of algorithm that simulatesynergetic phenomena. In training phase the BIOID createsa prototype called faceprint for each person.

A newly recorded pattern is preprocessed and comparedwith each faceprint stored in the database. A s comparisonsare made, the system assigns a value to the comparisonusing a scale of one to ten. If a score is above apredetermined threshold, a match is declared. .


22/31

Face recognition starts with a picture,attempting to find a person in the image.

This can be accomplished using severalmethods including movement, skin tones,or blurred human shapes. The facerecognition system locates the head and

finally the eyes of the individual. A matrix is then developed based on the

characteristics of the Individuals face. Themethod of defining the matrix variesaccording to the algorithm

This matrix is then compared to matricesthat are in a database and a similarityscore is generated for each comparison.


23/31

A rtificial intelligence is used to simulatehuman interpretation of faces. In order to increase the accuracy andadaptability, some kind of machine

learning has to be implemented.


24/31

MET H OD OF CAPTUR ING

VID EO IMA G ING TH ERM AL IMA G ING

Video imaging is more common asstandard video cameras can beused. The precise position and theangle of the head and thesurrounding lighting conditions may

affect the system performance. Thecomplete facial image is usuallycaptured and a number of points onthe face can then be mapped,position of the eyes, mouth and thenostrils as a example.

More advanced technologies make3-D map of the face whichmultiplies the possiblemeasurements that can be made .

Thermal imaging has better accuracy as it uses facialtemperature variations causedby vein structure as thedistinguishing traits. A s the heat

pattern is emitted from the faceitself without source of externalradiation these systems cancapture images despite thelighting condition, even in thedark.

The drawback is high cost. Theyare more expensive thanstandard video cameras.


25/31

H O W FA CE RE COGN ITIONSYSTEMS W ORK 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 hascertain distinguishable landmarks. These are the peaksand valleys that make up the different facial features.

Visionics defines these landmarks as nodal points. Thereare about 80 nodal points on a human face. Here are fewnodal points that are measured by the software.

1. distance between the eyes2. width of the nose3. depth of the eye socket4. cheekbones5. jaw line6. chin


26/31

TH E SOFT W ARE

D etection- when the system is attached to a videosurveilance system, the recognition softwaresearches the field of view of a video camera for faces. If there is a face in the view, it is detectedwithin a fraction of a second. A multi-scalealgorithm is used to search for faces in lowresolution. The system switches to a high-resolution search only after a head-like shape isdetected.

A lignment- O nce a face is detected, the systemdetermines the head's position, size and pose. A face needs to be turned at least 35 degreestoward the camera for the system to register it.


27/31

Normalization- The image of the head isscaled and rotated so that it can be registeredand mapped into an appropriate size andpose. Normalization is performed regardlessof the head's location and distance from thecamera. Light does not impact thenormalization process.

Representation- The system translates thefacial data into a unique code. This codingprocess allows for easier comparison of thenewly acquired facial data to stored facialdata.


28/31

Matching- The newly acquired facial data iscompared to the stored data and (ideally) linkedto at least one stored facial representation.

The heart of the Face It facial recognition systemis the L ocal Feature A nalysis ( L FA ) algorithm.

This is the mathematical technique the system

uses to encode faces. The system maps the face and creates a

faceprint, a unique numerical code for that face.O nce the system has stored a faceprint, it cancompare it to the thousands or millions of faceprints stored in a database.

Each faceprint is stored as an 84-byte file.


29/31

ADVANTAGES

There are many benefits to facerecognition systems such as its convinenceand Social acceptability.all you need is

your picturetaken for it to work. Face recognition is easy to use and in

many cases it can be performed without aPerson even knowing.

Face recognition is also one of the mostinexpensive biometric in the market and Itsprice should continue to go down.


30/31

D ISADVANTAGES

Face recognition systems cant tell the

difference between identical twins


31/31

APPL ICAT IONS

Security/Counterterrorism. A ccess control,comparing surveillance images to Knowterrorist.

Day Care: Verify identity of individuals pickingup the children.

Residential Security: A lert homeowners of approaching personnel

Voter verification: Where eligible politicians are

required to verify their identity during a votingprocess this is intended to stop voting wherethe vote may not go as expected.

Banking using ATM: The software is able toquickly verify a customers face.

Documents

Face Recognisation technology