Welcome message from author

This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript

1

Wavelets in Pattern Wavelets in Pattern RecognitionRecognition

Lecture Notes in Pattern Recognition by W.Dzwinel

Uncertainty principle

2

Uncertainty principle

Tiling

3

Windowed FT vs. WT

Idea of “mother” wavelet

4

Scale and resolution

STFT vs. WT

5

Tiling

STFT vs. WT

6

Wavelets – continuous transform

Wavelets – continuous transform

7

Wavelets – discrete transform

Scaling function

8

Any function f(t) can be represented by the series of wavelets expansion:

∑∑∑= ∈∈

∈+=J

lj kjk

lJl RLtftkjdtlctf

ZZ

)()(,)(),()()()( 2ψϕ

flc Jlϕ=)(

fkjd jkψ=),(

c(l) – low frequency coefficients

d(j,k) – high frequency coefficients on different detail levels

Wavelet series

S

D1

D2

A1

D3

A2

A3

Consecutive iterations starting from a signal and decomposing it into approximations (A) and details (D).

Wavelet decomposition

9

Haar wavelet

Haar wavelet

10

Wavelet transformation - conditions

∫+∞

∞−

== 0)()( 2121 dttt ψψψψ

Wavelet ? (t) has to fulfill a few conditions:

∫+∞

∞−

= 0)( dttψ

∫∞

∞−

∞<dtt 2)(ψ

Wavelets represents a basis in the L2 Hilbert Space which CAN be orthogonal and /or orthonormal:

∫+∞

∞−

== 1)(2dttψψ

Haar wavelet

11

Haar decomposition

Wavelets and images

12

Wavelets and images

Wavelets – different bases

13

Multi-resolution

Wavelets construction

14

Wavelets – construction

Wavelets – construction

15

Wavelets – construction

Wavelets – construction

16

Wavelets in 2-D

Two dimensional wavelets

17

Wavelets – in multiple dimensions

Wavelets – in multiple dimensions

18

Wavelets – in multiple dimensions

Two-dimensional functionsidwt2, waverec2

Multi-level reconstructionwaverec

Single-level reconstruction of 1D signalidwt

Two-dimensional functionsdwt2, wavedec2

multi-level signal decompositionwavedec

One-dimensional single-level decomposition of a given signal

dwt

DescriptionFunction

wavemenu – starts graphical interface

Matlab and wavelets

19

Noise removal

Select first:

• Wavelet form

• Number of decomposition levels

1. Wavelet decomposition of signal S on level N.

2. Define the thresholds on all the levels from 1 to N and eliminate small wavelet coefficients of all the details.

3. Complete wavelet reconstruction by means of approximation and remaining coefficients of the details.

Thresholding and elimination

Two types (at least) of thresholding process:

• Hard elimination

• Soft elimination

≤>

=δδ

)(,0)(),(

)(tytyty

tytw

≤>−

=δ

δδ)(,0

)(),)(())(sgn()(

tytytyty

tymk

• Comparisons

twarda miekka

20

b) Hard

a) Soft

Symlets wavelets, and 4-level decomposition is used. The threshold values are the same.

Noise removal

Details and

threshold values

Decomposition coefficients before and

after thresholding

Noise removal (1D signal)

21

Noise removal (2D signal)

1. Signal decomposition2. Thresholding and elimination of coefficients3. Reconstruction.

Ad. 1, 3 – similar as in noise removal

Ad. 2 – different approaches exista) Fix the global threshold value and/or define a

quality of compression parameterb) Adaptive threshold setting on every

decomposition level

Signal compression

22

Histograms of some image before and after wavelet transform

Number of eliminated coefficients vs. the energy of the signal kept

Signal compression

Daubechies 3 wavelets decomposed on 3 levelsResult:

99.99% of signal energy preserved

Eliminated - 84.74%coefficients.

Signal compression (1D)

23

Signal compression-comparisons

Signal compression-comparisons

24

Signal compression-comparisons

Wavelet compression vs JPEG:

Original image - 786486 b waveletcompression - 7812 b

Signal compression-comparisons

25

Wavelet compression - 7812 b JPEG – 8071 b

Signal compression-comparisons

Wavelet compression vs JPEG:

coiflet waveletorder 5.

Detection of singularities(rapid change of frequency)

26

Two close discontinuities

Daubechies order 2.

Detection of singularities(singularity)

Daubechies order 1:

Detection of singularities(discontinuity of the second derivative)

27

N2NFWT

N2 log2NN log2NFFT

N3N2DFT

obrazsygnal 1D

Computational complexity

Which wavelets ???

l Continuous – very slow and redundant (overcompletness) but more reliable. The information cannot be lost easily.

l Biorthogonality – one set of wavelets for decomposition one for reconstruction (higher dimensions) are symmetric and have compact support but may amplify any error introduced on the coefficients

l Orthogonal – fast, concise but arbitrary scales – because orthogonal transformation are not translation invariant

l The number of vanishing moments determine what the wavelets do not see (first vanishing moment – linear function is not seen). More vanishing moments à search is focused on better selectivity in time but p à vanishing moments means that wavelet support must be at least 2p-1 larger support àmore computations.

28

Which wavelets ????l Image compression à 3-4 vanishing moments. l A few large singularities à more vanishing moments,

More singularities à smaller support à lesser number of vanishing moments

l Daubechies wavelets the most vanishing moments for the smallest possible support

l Regularity – regularity n à n+1 derivatives. Important for image encoding. Not important for audio. Most regular wavelets are splines.

l Frequency selectivity – not important for images, but important for audio.à freq.select == many vanishing moments. The best trade off is using Gabor functions or B-spline wavelets

Mammograms and radiology

29

Wavelets in Sci. Visulization

Wavelets transform in PDE solving

30

Ridglets

Ridglets tiling

31

Ridglets

Curvelets

32

Curvelets

Comparisons of different approaches

33

Comparisons of different approaches

Comparisons of different approaches

34

Web pages

1. Jan T. Bialasiewicz: Falki i aproksymacje, WNT Warszawa 2000.

2. B.B. Hubbard, The world according to waveletys, AK Peters Ltd, pp325

3. Andrew S. Glassner: Principles of digital image synthesis, MorganKaufmann Publishers 1995.

4. Wojciech Maziarz, Krystian Mikolajczyk: A course on wavelets for beginners- http://galaxy.uci.agh.edu.pl/~maziarz/Wavelets/, Kraków 1999.

5. The MathWorks Inc., Developers of Matlab & Simulink -http://www.mathworks.com.

6. Alain Fournier: Wavelets and their applications in computer graphics, SIGGRAPH'95 Course Notes, 1995.

7. Amara Graps: An Introduction to wavelets, IEEE Computational Sciencesand Engineering,Vol. 2, Number 2, Summer 1995, pp 50-61 -http://www.amara.com/IEEEwave/IEEEwavelet.html.

References

Related Documents