Face Recognition Using Eigen Faces and Artificial Neural Network

parathyroidsanchovyΤεχνίτη Νοημοσύνη και Ρομποτική

17 Νοε 2013 (πριν από 3 χρόνια και 8 μήνες)

85 εμφανίσεις

Face Recognition Using Eigen Faces and
Artificial Neural Network

資訊所

P76994270
陳智凱

Outline


1

Introduction

1

Proposed Technique

2

Experiment and Analysis

3

Conclusion

4


2

Introduction


Face

recognition

has

become

an

important

issue

in

many

application


Developing

a

computational

model

of

face

recognition

is

quite

difficult,

because

faces

are

complex,

multi
-
dimensional

visual

stimuli



The

starting

step

involves

extraction

of

the

relevant

features

from

facial

image


A

challenge

is

how

to

quantize

facial

features

so

that

a

computer

is

able

to

recognize

a

face

3

Introduction


There

are

two

basic

methods

for

face

recognition


Based

on

extracting

feature

vectors

from

the

basic

parts

of

a

face


Based

on

the

information

theory

concepts

viz
.

principal

component

analysis

method


An

unsupervised

pattern

recognition

scheme

is

proposed



Based

on

eigenface,

PCA,

and

ANN


PCA

for

face

recognition

is

based

on

the

information

theory

approach

in

which

the

relevant

information

in

a

face

image

extracted

as

efficiently

as

possible


ANN

is

used

for

classification



4


5

Proposed Technique


6

Proposed Technique

1)
Preprocessing

and

face

library

formation


Preprocessing


Image

size

normalization



Histogram

equalization


Conversion

into

gray

scale


Face

library

is

divided

into

two

sets



training

dataset

and

testing

dataset

7

Proposed Technique

8

Proposed Technique

9

Fig. 1


Training faces

Proposed Technique

10

Proposed Technique

11

Proposed Technique

12

Proposed Technique

13

Proposed Technique

14

Proposed Technique

15

Proposed Technique

4)
Training

of

Neural

Networks


Face

descriptors

are

used

as

inputs

to

train

networks



Face

descriptors

that

belong

to

same

person

are

used

as

positive

examples

for

the

person’s

network

(such

that

network

gives

1

as

output),

and

negative

examples

for

the

others

network

(such

that

network

gives

0

as

output)

16

Proposed Technique

17

Fig. 5


Training of Neural Network

Proposed Technique

5)
Simulation

of

ANN

for

recognition

1.
New

test

image

is

taken

for

recognition

and

its

face

descriptor

is

calculated

from

the

eigenfaces

2.
The

new

descriptor

is

given

as

an

input

to

every

network,

then

these

networks

are

simulated

3.
Compare

the

simulated

results

and

if

the

maximum

output

exceeds

the

predefined

threshold

level,

then

it

is

confirmed

that

this

new

face

belongs

to

the

recognized

person

with

the

maximum

output


18

Proposed Technique

19

Fig. 6


Testing of Neural Network

Proposed Technique

20


21

Experiment and Analysis


The proposed method is tested on ORL face database



There are more than one image of an individual’s face
with different confitions in database



For each of 40 distinct subjects, there are 10 different
images

22

Experiment and Analysis


For some subject, the images were taken at different
details and all the images were taken against a dark
homogeneous background



The original pictures of 112
х

92 pixels have been
resized to 56
х

46



Eigenfaces are calculated by using PCA algorithm

23

Experiment and Analysis


The number of networks is equal to the number of
people in the database


The initial parameters of the Neural Network used are
given below

24

Experiment and Analysis


The number of networks is equal to the number of
people in the database, therefore forty networks




Among the ten images, first 6 of them are used for
training the neural network, and these networks
would be used later on for recognition



For testing the whole database, the faces used in
training, testing, and recognition are changed

25

Experiment and Analysis


The complete face recognition process

26

Experiment and Analysis


The proposed technique is analyzed by varying the
number of eigenfaces used feature extraction

27

Experiment and Analysis


28

Experiment and Analysis


The result derived from proposed method is
compared with K
-
means and Fuzzy Ant with fuzzy C
-
means

29


30

Conclusion


The paper presents a face recognition approach using
PCA and Neural Network techniques


The eigenface method is sensitive to head
orientations, and most of the mismatches occur for
the images with large head orientation


The proposed recognition method has advantage over
K
-
means method and Fuzzy Ant with fuzzy C
-
means
based algorithm

31


32

Q

&

A

heartthrob.kai@gmail.com