1. A Completed Modeling of Local Binary
Pattern Operator for Texture Classification
Zhenhua Guo,
Lei Zhang,
David Zhang
2012 IEEE Transactions on Image Processing
NTNU-CSIE Chen-Lin Yu , Total page:23

2. Outline





Introduction
Brief Review of LBP
Completed LBP (CLBP)
Experiments
Conclusion
2

3. INTRODUCTION
3

4. Texture Classification
 Texture classification is an active research topic in
computer vision and pattern recognition.
 In texture classification, the goal is to assign an
unknown sample image to one of a set of known texture
classes.
 Texture classification process involves two phases:
(1) learning phase and (2) the recognition phase
4

5. Divided texture analysis methods into
four categories:
Statistical
Geometrical
Model-based
Signal processing
http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/OJALA1/texclas.htm

6. Goal:
 Proposing a new local feature extractor to generalize
and complete LBP(CLBP).
 Some information is missed in LBP code. We
attempts to address that how to effectively represent
the missing information in the LBP so that better
texture classification.
 In CLBP, a local region is represented by its center
pixel and a local difference sign-magnitude transform
(LDSMT).
6

7. BRIEF REVIEW OF LBP
7

8. LBP
8

9. Rotation variance
 We define formula U which value of an LBP pattern
is defined as the number of spatial transitions
(bitwise 0/1 changes) in that pattern.
0
1
1
0
1
0
0
0
01110000
U(LBPP,R) = 2
1
1
1
1
1
0
0
1
11110101
U(LBPP,R) = 4
9

10.  The uniform LBP patterns refer to the patterns
which have limited transition or discontinuities
(U<=2) in the circular binary presentation [13].
 (superscript “riu2” means rotation invariant “uniform”
patterns with U<=2)
[13] Multiresolution gray-scale and rotation invariant texture classification with Local Binary Pattern
10

11. COMPLETED LBP (CLBP)
11

12. CLBP Framework
S
M
Local
Difference
CLBP_S
CLBP_M
LDSMT
Original
Image
Center Gray
Level
Classifier
CLBP_C
CLBP
Histogram
CLBP Map
(nearest neighborhood)
12

13. I. Local Difference Sign-Magnitude
Transform (LDSMT)
a 3*3 sample block
sign components
local differences
magnitude components
13

14.  Fig(b) local differences vector
 Fig(c) sign vector
 Fig(d) magnitude vector
 Ex: different vector is [3,9,-13,-16,-15,74,39,31]
after LDMST sign vector is [1,1,-1,-1,-1,1,1,1]
and magnitude vector is [3,9,13,16,15,74,39,31]
14

15. II. CLBP Map
 By the LDSMT ,three operators, namely CLBP_C,
CLBP_S and CLBP_M, are proposed to code the C,
S and M features, respectively
 Combine CLBP_S and CLBP_M
• 1. Concatenation (CLBP_S_M)
• 2. Joint (CLBP_S/M)
 Combine three operators
• 1. Joint (CLBP_S/M/C)
• 2. Hybrid (CLBP_M_S/C or CLBP_S_M/C)
15

16. Dissimilarity Metric and Multi-scale CLBP
 There are various metrics to evaluate between two
histograms, such as histogram intersection, loglikelihood ratio, and chi-square statistic [13].
 The nearest neighborhood classifier with the chisquare distance is used to measure the dissimilarity
between two histograms
16

17. EXPERIMENTS
17

18. Outex Database
 Includes 24 classes of textures , each texture available at the
site is captured using three different simulated illuminants provided in the light source:
– H : 2300K (陽光 左下)
– Inca : 2856K (日光燈 右下)
– TL84 : 4000K(螢光燈 右上)
and nine rotation angles
(0o,5º,10º,15º,30º,45º,60º,75º and 90º)
18

19. 19

20. CONCLUSION
20

21.  We analyzed LBP from a viewpoint of local
difference sign-magnitude transform (LDSMT), and
consequently a new scheme, namely completed
LBP (CLBP)
 By fusing CLBP_C、CLBP_S、CLBP_M codes, it
will much better texture classification accuracy than
the state-of-the-arts LBP.
21

22. Thank you.
END
22