16
Facial Facial Recognition Recognition

Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Embed Size (px)

Citation preview

Page 1: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Facial Facial RecognitionRecognition

Page 2: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Facial RecognitionFacial Recognition

1.1. takes a picture of a persontakes a picture of a person

2.2. runs that image through the runs that image through the databasedatabase

3.3. finds a match and identifies the finds a match and identifies the personperson

Humans have always had the innate ability Humans have always had the innate ability to recognize and distinguish between to recognize and distinguish between faces, yet computers only recently have faces, yet computers only recently have shown the same ability. shown the same ability.

Page 3: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Origin of Facial Recognition Origin of Facial Recognition SoftwareSoftware

► 1960s -1960s - scientists began work on using the scientists began work on using the computercomputer to recognize human faces to recognize human faces

Forty Years LaterForty Years Later::

In 2001 the Tampa Police Department In 2001 the Tampa Police Department installed police cameras equipped with installed police cameras equipped with facial recognition technology but it was facial recognition technology but it was scrapped in 2003 due to its ineffectivenessscrapped in 2003 due to its ineffectiveness

Page 4: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Example of facial recognition Example of facial recognition photosphotos

Page 5: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Facial recognition software Facial recognition software must:must:►Differentiate between the face and the Differentiate between the face and the

backgroundbackground►Measure the different features of the face to Measure the different features of the face to

determine its unique in digital formdetermine its unique in digital form►Determine a match in a databaseDetermine a match in a database

Page 6: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

What makes one face different What makes one face different from another face?from another face?

► distinguishable distinguishable landmarks called nodal landmarks called nodal points -points - the different peaks and valleys that the different peaks and valleys that make up facial featuresmake up facial features

► each human face has approximately 80 each human face has approximately 80 nodal pointsnodal points

Page 7: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

How does the software analyze a How does the software analyze a face? Step 1:face? Step 1:

►The software determines:The software determines:►Distance between the eyesDistance between the eyes►Width of the noseWidth of the nose►Depth of the eye socketsDepth of the eye sockets►The shape of the cheekbonesThe shape of the cheekbones►The length of the jaw lineThe length of the jaw line

Page 8: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

How does the software analyze a How does the software analyze a face? Step 2:face? Step 2:

►Result = a numerical code, called a Result = a numerical code, called a faceprintfaceprint

►The unique The unique faceprintfaceprint gets saved in the gets saved in the databasedatabase

►The software compares a 2D image to The software compares a 2D image to a 2D image in the database for a a 2D image in the database for a match. match.

Page 9: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

IssuesIssuesBasic Problem:Basic Problem: ► Images not taken in a controlled Images not taken in a controlled

environment resulted in a high failure to find environment resulted in a high failure to find a match in the database.a match in the database.

Match failures were caused by variations Match failures were caused by variations in:in:

► light reflections on the facelight reflections on the face► facial expressionsfacial expressions► facial angles to the camerafacial angles to the camera► aging – time intervals between photosaging – time intervals between photos

Page 10: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

SolutionsSolutions

►3D facial recognition uses distinctive 3D facial recognition uses distinctive features of the face -- where rigid features of the face -- where rigid tissue and bone is most apparent, tissue and bone is most apparent, such as the curves of the eye socket, such as the curves of the eye socket, nose and chin -- to identify the subject. nose and chin -- to identify the subject.

►These areas are all unique and don't These areas are all unique and don't change over time.change over time.

Page 11: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

EquipmentEquipment

► The Vision 3D + 2D ICAO camera is The Vision 3D + 2D ICAO camera is used to perform enrollment, used to perform enrollment, verification and identification of 3D and verification and identification of 3D and 2D face images2D face images..

Page 12: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Steps for ConsistencySteps for Consistency

►Alignment – the system determines Alignment – the system determines the head's position, size and pose the head's position, size and pose Measurement of facial curves on a Measurement of facial curves on a sub-millimeter (or microwave) scale sub-millimeter (or microwave) scale and creates a template.and creates a template.

►Measurement - the system then Measurement - the system then measures the curves of the face on a measures the curves of the face on a sub-millimeter (or microwave) scale sub-millimeter (or microwave) scale and creates a template.and creates a template.

Page 13: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always
Page 14: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Primary Uses of this Primary Uses of this TechnologyTechnology

► Law EnforcementLaw Enforcement► Other potential Other potential

applications applications include ATM and include ATM and check-cashing check-cashing security.security.

Page 15: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

ETHICAL ISSUESETHICAL ISSUES

► Invasion of Privacy – having a picture Invasion of Privacy – having a picture taken without knowingtaken without knowing

Page 16: Facial Recognition. 1. takes a picture of a person 2. runs that image through the database 3. finds a match and identifies the person Humans have always

Pros/ConsPros/Cons

AdvantagesAdvantages► harder to circumventharder to circumvent► easy to useeasy to use► no remembering passwords or forgetting id no remembering passwords or forgetting id

badgebadge

DisadvantagesDisadvantages► error rates are higherror rates are high