Carpet wear classification based on support vector machine pattern recognition approach

(1)

Carpet Wear Classification based on

Support Vector Machine Pattern

Recognition Approach

1,2

C. Copot,

2

S. Syafiie,

3

S. Vargas,

2

R. de Keyser,

4

L. van Langenhove,

1

C. Lazar

1

Department of Automatic Control and Applied Informatics, “Gheorghe Asachi”

Tehnical University of Iasi

2

Department of Electrical energy, Systems and Automation, Ghent University

3

Department of Telecommunication, Ghent University

(2)

● Introduction

● Features extraction

● PCA (Principal Component Analysis)

● Features classification

● Results

● Conclusion

(3)

Objective

Visual assessment by human expert (compared with standard) Mechanical wear

Label

5 4.5 4 . . . 1 5000 loops 10000 loops 15000 loops 20000 loops 25000 loops Carpet sample 3D Laser Scanner Co-occurrence matrix Computer LABEL

Human classifier

Automated

classifier

Computer classification 0 loops

Objective

Introduction

(4)

Data Acquisition

Fig 1. Digital image Fig 2. Laser image

(5)

1 3 4 2 7 1 1 2 5 6 3 6 7 1 2 1 1 2 7 5 2 3 1 0 0 0 0 0 0 0 0 1 0 2 0 0 0 1 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 0 0 0 1 0 0

)

1 &

0 (

o

d

=

Co-occurrence matrix

¾

Co-occurrence matrix composed

1 2 3 4 5 6 7 1 2 3 4 5 6 7 Gray-scale matrix Co-occurrence matrix

1 ))

,

(

),

,

(

))

,

(

),

,

(

g

i

j

g

i

+

k

j

+

l

=

C

g

i

j

g

i

+

k

j

+

l

+

C

_d _d 2 2

l

k

d

=

+

Where:

(6)

9 Energy

9 Sum Entropy

9 Entropy

9 Sum Variance

9 Inertia

9 Difference Variance

9 Homogeneity

9 Difference Entropy

9 Contrast

9 Informational Measures of Correlation

9 Correlation

9 Maximal Correlation Coefficient

9 Sum Average

Theoretical representation

Features extraction

(7)

Name of carpet Human expert classification 20 kl 517 BMW 20 kl 517 – 2 2.5 20 kl 517 – 4 2.5 20 kl 517 – 6 2.5 20 kl 517 – 8 2 20 kl 517 – 10 2 20 kl 517 – 12 2

Haralick feature - Energy

Table of human expert

(8)

¾

Principal Component Analysis is a useful technique used to reduce

the dimensionality of large data sets, such as those from micro array

analysis;

¾

When there are more than three variables, it is more difficult to

visualize graphically their relationships.

Principal Components PC1 = 94% 4% 2% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PC 9 PC 10 PC 11 PC 12 PC 13 PC 14

(9)

2D Plotting of first tow PC _{3D Plotting of first third PC} -150 -100 -50 0 50 100 -80 -60 -40 -20 0 20 40 60 carpet KL-517 PCA 1 PCA 2 class 2 class 2.5 class 5 -150 -100 -50 0 50 100 -100 -50 0 50 100 -50 -40 -30 -20 -10 0 10 20 30 40 PCA 1 carpet KL-517 PCA 2 PC A 3 class 2 class 2.5 class 5

PCA

PCA Results

(10)

¾

A separated hyperplane:

Binary SVM

0 =

+ b

x

w

T

⎥

⎦

⎤

⎢

⎣

⎡

−

+

=

+

1

0

1 b

x

w

T

1 if

0

1 if

0 −

=

<

+

=

>

+

y

b

x

w

y

b

x

w

T T

Where - x

is

training

data

- y is indicator vector

(11)

Nonlinear - SVM

p j i j i

x

k

(

,

)

=

(

+

1 )

2 2 / || ||

)

,

(

x

_i

x

_j

e

x y σ

k

=

− −

-

polynomial kernel

-

Gaussian kernel

The separable plane will be:

0 )

,

(

x

w

+ b

=

k

_i _j T

Input space Feature space

(12)

¾

One - Against - All decomposition

Multi – class SVM

-4 -2 0 2 4 6 8 -6 -4 -2 0 2 4 6 8 c lass 1 c lass 2 c lass 3 c lass 4 c lass 5 PC 1 PC 2

(13)

Type of carpet Polynomial kernel % Gaussian kernel % A8 – 501 84,4 84,4 A8 – 701 84,4 100 Big 4 100 100 20 KL 803 beige 84,4 84,4 LA 7 84,4 100 Pr 84 100 100 20 KL 517 100 100 Big 8 86 86 LA 9 72 100 20 KL 803 100 100 Over all 89,56 95,48

Testing Results

¾

For every carpet are used:

- 4 points for training

- 1 point for testing

Table of percentage classification

(14)

-150 -100 -50 0 50 100 -80 -60 -40 -20 0 20 40 60 -60 -40 -20 0 20 40 PCA 1 carpet KL-517 PCA 2 PC A 3 class 2 class 2.5 class 5 SVM