33
Introduction and Introduction and Overview Overview Dr Mohamed A. El-Gebeily Dr Mohamed A. El-Gebeily Department of Mathematical Department of Mathematical Sciences Sciences KFUPM KFUPM [email protected] [email protected]

Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM [email protected]

Embed Size (px)

Citation preview

Page 1: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Introduction and OverviewIntroduction and Overview

Dr Mohamed A. El-GebeilyDr Mohamed A. El-Gebeily

Department of Mathematical SciencesDepartment of Mathematical Sciences

KFUPMKFUPM

[email protected]@kfupm.edu.sa

Page 2: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Why Wavelets?Why Wavelets? Comparison With Fourier AnalysisComparison With Fourier Analysis What is Wavelet Analysis?What is Wavelet Analysis? The Continuous Wavelet The Continuous Wavelet

TransformTransform The Discrete Wavelet TransformThe Discrete Wavelet Transform Introduction to Wavelet FamiliesIntroduction to Wavelet Families ApplicationsApplications

Page 3: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Why Wavelets?Why Wavelets? Wavelets have scale and time zooming aspects

Scale zoomingScale zooming•Used to analyze the regularity of a signal.

•Biology for cell membrane recognition. •Metallurgy for the characterization of rough surfaces •Finance (which is more surprising), for detecting the properties of quick variation of values. •In Internet traffic description, for designing the services size.

Page 4: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Time zoomingTime zooming•Used to detect ruptures and short-time phenomena such as transient processes •Applications:

•Industrial supervision of gear-wheel faults•Non destructive control quality processes •Detection of short pathological events as epileptic crises or normal ones as evoked potentials in EEG (medicine) •SAR imagery •Automatic target recognition •Intermittence in physics

Page 5: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Wavelet Decomposition as a WholeWavelet Decomposition as a Whole•Many applications use the wavelet decomposition taken as a whole. •The common goals are de-noising or compression.

•Compression: FBI fingerprints•It is almost impossible to sum up several thousand papers written within the last 15 years.•It is difficult to get information on real-world industrial applications from companies.

Page 6: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Fourier AnalysisFourier AnalysisBreaks down a signal into constituent sinusoids of different frequencies.

Useful when the signal's frequency content is of great importance. Has a serious drawback: time information is lost.

Page 7: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Short-Time Fourier Analysis (STFT)Short-Time Fourier Analysis (STFT)Dennis Gabor (1946) adapted the Fourier transform to analyze only a small section of the signal at a time -- a technique called windowing the signal

Limited precision, because of the fixed size of the window for all frequencies.

Page 8: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Wavelet AnalysisWavelet Analysis

Wavelet: a windowing technique with variable-sized regions. Long time intervals where we want more precise low-frequency information, and shorter regions where we want high-frequency information.

Page 9: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

What Can Wavelet Analysis Do?What Can Wavelet Analysis Do?The ability to perform local analysis -- that is, to analyze a localized area of a larger signal. Example: a sinusoidal signal with a small discontinuity

Page 10: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

What Is Wavelet Analysis?What Is Wavelet Analysis?

A wavelet is a waveform of effectively limited duration that has an average value of zero. Compare wavelets with sine waves

Page 11: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

The Continuous Wavelet TransformThe Continuous Wavelet TransformThe continuous wavelet transform (CWT) is defined as:

scale.for stands and

(in time)position for stands where

,1

,

s

u

dts

ut

stfsuC

Page 12: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

ScalingScaling

ShiftingShifting

Page 13: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Five Easy Steps to a Continuous Five Easy Steps to a Continuous Wavelet TransformWavelet Transform1.Take a wavelet and compare it to a section at the start of the original signal. 2.Calculate a number, C, that represents how closely correlated the wavelet is with this section of the signal.

Page 14: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

3.Shift the wavelet to the right and repeat steps 1 and 2 until you've covered the whole signal.

4.Scale (stretch) the wavelet and repeat steps 1 through 3.

5.Repeat steps 1 through 4 for all scales.

C=.767

Page 15: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

A scalogram is a 3-D plot of the wavelet coefficients against time and scale.

Page 16: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

The Discrete Wavelet Transform The Discrete Wavelet Transform (DWT)(DWT)• The wavelet coefficients are computed at the dyadic points

It turns out that this is more efficient and enough to recover the original function from the wavelet coefficients.

jj ksu 2,2,

Page 17: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

One-Stage Filtering: Approximations One-Stage Filtering: Approximations and Detailsand Details

In wavelet analysis, we often speak of approximations and details. The approximations are the high-scale, low-frequency components of the signal. The details are the low-scale, high-frequency components.

Page 18: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

DownsamplingDownsampling

The WT produces two sequences called cA and cD.

Page 19: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

ExampleExample A pure sinusoid with high-frequency noise added to it.

s = sin(20.*linspace(0,pi,1000)) + 0.5.*rand(1,1000); [cA,cD] = dwt(s,'db2');

Page 20: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Multiple-Level DecompositionMultiple-Level Decomposition

wavelet decomposition tree.

Page 21: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Reconstruction FiltersReconstruction Filters

The low- and highpass decomposition filters (L and H), together with their associated reconstruction filters (L' and H'), form a system of what is called quadrature mirror filters

Page 22: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Reconstructing Approximations Reconstructing Approximations and Detailsand Details

It is possible to reconstruct the approximations and details themselves from their coefficient vectors.

Page 23: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Extending this technique to the components of a multilevel analysis, we find that similar relationships hold for all the reconstructed signal constituents. That is, there are several ways to reassemble the original signal:

Page 24: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

An Introduction to the Wavelet An Introduction to the Wavelet FamiliesFamilies

Several families of wavelets have proven to be especially useful:HaarHaar wavelet is the first and simplest. Haar wavelet is discontinuous, and resembles a step function. Same as Daubechies db1.

Page 25: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

DaubechiesDaubechiesIngrid Daubechies invented what are called compactly supported orthonormal wavelets -- thus making discrete wavelet analysis practicable.

Page 26: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

BiorthogonalBiorthogonal

Page 27: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

CoifletsCoiflets

Page 28: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

SymletsSymlets

Page 29: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

MorletMorlet

Page 30: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

Mexican HatMexican Hat

Page 31: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

MeyerMeyer

Page 32: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

The purpose of this example is to show how analysis by wavelets can detect the exact instant when a signal changes. The discontinuous signal consists of a slow sine wave abruptly followed by a medium sine wave.

Page 33: Introduction and Overview Dr Mohamed A. El-Gebeily Department of Mathematical Sciences KFUPM mgebeily@kfupm.edu.sa

The image below is compressed to a ratio of 25:1 of its original size by using the two dimensional wavelet transform