WaveletsWavelets
An Introductory ExpositionAn Introductory Exposition
Features, Analysis Structures, and Features, Analysis Structures, and
Selected ApplicationsSelected Applications
Entry of Wavelets into theEntry of Wavelets into the
Domain of SPDomain of SP
�� Most significant event in SP after FT.Most significant event in SP after FT.
�� Reference to Wavelet Transform: Reference to Wavelet Transform:
“Fourier Transform of 20“Fourier Transform of 20thth Century”Century”
�� Wavelets fills the missing link in signal Wavelets fills the missing link in signal
processing: link between time and processing: link between time and
frequency studiesfrequency studies
�� ““Little wave with big future”Little wave with big future”
Groups Active in Wavelet StudiesGroups Active in Wavelet Studies
Wavelets
Mathematics: Function
Analysis and Approximation
Engineering, Applied Scientists
In Signal Processing
Applications
Industrial Groups
Key Idea Key Idea
Alternative Signal Representations Alternative Signal Representations
and Transformationand Transformation
�� Signal representation in a suitable Signal representation in a suitable
domain for information extraction,….domain for information extraction,….
Examples:Examples:
�� Fourier Transform for spectral analysisFourier Transform for spectral analysis
�� Hilbert Transform in envelop detectionHilbert Transform in envelop detection
�� KLT(PCA) for optimal function approx.KLT(PCA) for optimal function approx.
�� Laguerre basis functionLaguerre basis function
�� Numerous other transforms( DCT,Radon,…)Numerous other transforms( DCT,Radon,…)
Two Domains for Signal Two Domains for Signal
Representation Representation
Different projection spaces different signal representaDifferent projection spaces different signal representa--
tionstions
Signal Domain
Projection space,
Basis functions
Transformed Domain
Transformation
Representation in Time
f(t)Representation in Freq
F(ω)
Time Domain Studies Frequency Domain Studies
Fourier Transform
Shortcomings of Fourier TransformShortcomings of Fourier Transform
0 5 0 1 0 0 1 5 0 2 0 0 2 5 0- 0 . 0 2
0
0 . 0 2
0 . 0 4s i g n a l f ( t ) w i t h l o c a l i z e d c h a n g e s
f(t)
0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0- 1
0
1
co
s(t)
S i n e a n d c o s i n e a s b a s i s f u n c t i o n s o f F o u r i e r T r a n s f o r m
0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0- 1
0
1
sin
(t)
t i m e t
-- Basis function vary within Basis function vary within ±±±±±±±± ∞, no localization∞, no localization
-- Only one Basis function: Sine and Cosine or its complex formOnly one Basis function: Sine and Cosine or its complex form
-- Fourier Coefficients: Projection of function f(t) onto sine and Fourier Coefficients: Projection of function f(t) onto sine and
cosine bases functionscosine bases functions
-- Information about the entire range of a function is contained Information about the entire range of a function is contained
in the coefficients, no localizationin the coefficients, no localization
--
Fourier TransformFourier Transform
Loss of Local InformationLoss of Local InformationAn IllustrationAn Illustration
�� Information about singularities and sharp changes spread across many Information about singularities and sharp changes spread across many frequencies and many basic functionsfrequencies and many basic functions
�� Cost of computations for transients is highCost of computations for transients is high
x
freq ω
f(x)
F(ω)
Starting Point of WaveletsStarting Point of Wavelets
RealReal--world Signalsworld Signals
�� Dominance of transient and nonDominance of transient and non--
stationary signalsstationary signals
�� Information often reside in transients, Information often reside in transients,
changes changes
�� The need to develop tools for TThe need to develop tools for T--F F
analysis analysis
�� Shortcomings of STFT: Inability to Shortcomings of STFT: Inability to
model most of the nonstationary realmodel most of the nonstationary real--
world signalsworld signals
What Are WaveletsWhat Are Wavelets
�� Wavelets are Wavelets are wavelikewavelike oscilatory oscilatory
signals of finite bandwidth both signals of finite bandwidth both
in Time and in Frequencyin Time and in Frequency
�� Wavelets are Wavelets are basis functions of basis functions of
spacesspaces with certain propertieswith certain properties
Examples of WaveletsExamples of Wavelets
Db 4, Db 10 are from Daubechies family of wavelets.Db 4, Db 10 are from Daubechies family of wavelets.
Db wavelets have no analytical expression, they are Db wavelets have no analytical expression, they are
constructed numericallyconstructed numerically
Db4 Db10
Examples of Wavelet FunctionsExamples of Wavelet Functions
0 20 40 60 80 100 120 140 160 180-4
-2
0
2
WP
3,5
0 20 40 60 80 100 120 140 160 180-1
0
1
2Coiflet
WP
1,5
0 20 40 60 80 100 120 140-4
-2
0
2
WP
2,5
Wavelet Functions Generated from Wavelet Functions Generated from
Bior 3.9Bior 3.9
0 50 100 150 200 250 300 350-10
-5
0
5
10bior3.9, wavelet packet W30
0 100 200 300 400 500 600 700-4
-2
0
2
4Bior3.9, wavelet packet W 10
Representation in Time
f(t)Representation in Freq
F(ω)
Time Domain Studies Frequency Domain Studies
Shortcomings of Fourier Transform
Loss of Time Information in FTLoss of Time Information in FT
0 0.5 1 1.5 2 2.5 3 3.5
x 104
-1
-0.5
0
0.5
1
shift
ed
0 0.5 1 1.5 2 2.5 3 3.5
x 104
-1
-0.5
0
0.5
1 SALAAM with switching the 1st 5000 samples with the tail segment
Ori
gin
al
Fourier TransformFourier TransformLoss of time informationLoss of time information
0 1000 2000 3000 4000 50000
1000
2000
3000
4000
0 1000 2000 3000 4000 50000
1000
2000
3000
4000abs(fft) of SALAAM with shifting the 1st 5000 samples to the tail
Representation in Time
f(t)Representation in Freq
F(ω)
Time Domain Studies Frequency Domain Studies
Wavelets
An Important Property of Wavelets:
Wavelets Filling the Gap
Transient Nature of SignalsTransient Nature of Signals
��Most of the signals we deal areMost of the signals we deal are
�� Nonstationary: Nonstationary: ��NonNon--stationary refers to time variancy and stationary refers to time variancy and
spectral variation of the signals with timespectral variation of the signals with time
�� Unpredictable Unpredictable ��changes in statistics of the signals including changes in statistics of the signals including
changes in pdf function or statistical parameterschanges in pdf function or statistical parameters
��The degree of unpredictability varies for The degree of unpredictability varies for different signals and applicationsdifferent signals and applications
An Example of a Transient SignalAn Example of a Transient Signal
Speech SignalSpeech Signal
0 0.5 1 1.5 2 2.5 3 3.5
x 104
-1
0
1SALAAM
0 500 1000 1500 2000 2500 3000-0.5
0
0.5
sAla
am
0 500 1000 1500-1
0
1
saL
aam
0 500 1000 1500 2000 2500 3000-1
0
1
saL
aam
Three segments of SALAAM
EEG Signals, Awake and AsleepEEG Signals, Awake and Asleep
EEG (EP)EEG (EP)
Examples of Transient SignalsExamples of Transient SignalsEngine VibrationsEngine Vibrations
500 1000 1500 2000 2500 3000
-2
0
2
Hea
lthy
500 1000 1500 2000 2500 3000
-2
0
2
Time (Samplings)
Fau
lty
An Example of RealAn Example of Real--world Signalsworld Signals
Engine VibrationsEngine Vibrations
Wavelet Functions, WPWavelet Functions, WP
0 20 40 60 80 100-40
-20
0
20
40Bior3. 1, wavelet packet W2
0 20 40 60 80 100-4
-2
0
2
4db4, wavelet packet W15
Main Stages in Signal AnalysisMain Stages in Signal Analysis
Signal domain Signal domain basis functionsbasis functions Transformed domain, CoeffsTransformed domain, Coeffs
Transformation
Analysis
Reconstruction
Signals
Recon Signal
basis functions
Modified
Coefficients
Information
extraction
Examples of Processing in Examples of Processing in
Coefficient DomainCoefficient Domain
��Coding for communication, transmissionCoding for communication, transmission
��Compression and Data StorageCompression and Data Storage
��Detection and Pattern RecognitionDetection and Pattern Recognition
��Modification for Enhancement purposesModification for Enhancement purposes
��Noise ReductionNoise Reduction
��WatermarkingWatermarking
��Signal SeparationSignal Separation
��ModelingModeling
Two Main Tasks of Wavelet Two Main Tasks of Wavelet
AnalysisAnalysis
��Decomposition:Decomposition:
�� Information Reside in Signal Constituents / Information Reside in Signal Constituents /
ComponentsComponents
��TimeTime--Frequency RepresentationFrequency Representation
�� Transformation of a signal into timeTransformation of a signal into time--
frequency representation frequency representation
�� Different basis and transformations result in Different basis and transformations result in
different constituents and Tdifferent constituents and T--f informationf information
Key Concern in SPKey Concern in SP
Resolution/Localization in Signal Resolution/Localization in Signal
ProcessingProcessing��Resolution: Information Extraction at a Resolution: Information Extraction at a
Narrow Band of Signal Span.Narrow Band of Signal Span.
��Localization both in Time and in FrequencyLocalization both in Time and in Frequency
High freq
Low freq
Time/space
freq
Resolution/LocalizationResolution/Localization
in Timein Time
33
3
Resolution/Localization in Resolution/Localization in
FrequencyFrequency
0 2000 4000 6000 8000 10000 12000 140000
50
100
150Power Spectrum, Vibrat ion data at combustion zone, original and denois ed
Ori
gin
al
0 2000 4000 6000 8000 10000 12000 140000
50
100
150
de
no
ise
d,
soft
Freq. Hz
22
Need for Joint TimeNeed for Joint Time--Frequency Frequency
Analysis of High ResolutionAnalysis of High Resolution
��Why Joint TWhy Joint T--F Analysis:F Analysis:
�� Transient signal information reside in Transient signal information reside in
different bands of time and frequency different bands of time and frequency
domaindomain
�� The need for the study of signal at the The need for the study of signal at the
limits of resolutionlimits of resolution determined by determined by
Heisenberg Uncertainty LimitsHeisenberg Uncertainty Limits
Heisenberg Uncertainty PrincipleHeisenberg Uncertainty Principle
�� Heisenberg uncertainty principle defined in Heisenberg uncertainty principle defined in quantum physics and is applicable to signal quantum physics and is applicable to signal processing problems as well. processing problems as well.
�� Position and momentum of a particle can not Position and momentum of a particle can not be determined simultaneouslybe determined simultaneously
�� Sets a limit given as follows:Sets a limit given as follows:
�� ∆t. ∆f ≥ 1/4π ∆t. ∆f ≥ 1/4π
�� ∆x∆x22=⌠{(x=⌠{(x--µµm m ))22 |f(x)||f(x)|2 2 dxdx /⌠|f(x)|/⌠|f(x)|2 2 dxdx
µµm m is center of mass of the function f(x): is center of mass of the function f(x):
µµm=m=⌠{x| f(x)|⌠{x| f(x)|2 2 dx dx /⌠|f(x)|/⌠|f(x)|2 2 dxdx
Second ConcernSecond ConcernAlternative TimeAlternative Time––Frequency TilingFrequency Tiling
�� Information about Different signal behavior reside at Information about Different signal behavior reside at
different Tdifferent T--F BandsF Bands
�� Need to have alternative TNeed to have alternative T--F tiling and cell structuresF tiling and cell structures
Time/scale
Freq/scale
Third ConcernThird Concern
Adaptive Adaptive MultiMulti--Resolution StudyResolution Study
�� Different information reside at different resolutions. Different information reside at different resolutions.
�� Redundancy in Signal RepresentationRedundancy in Signal Representation
Time/scale
Freq/scale
Redundancy
in tiling
High Frequency
Low Frequency
Wavelets Wavelets
Tools for TTools for T--F AnalysisF AnalysisWavelets allow:Wavelets allow:
�� High resolutionHigh resolution and focused study of signals in time and focused study of signals in time
and in frequencyand in frequency
�� Low resolution, Coarse Low resolution, Coarse and more general picture and more general picture
and trend analysis and trend analysis
Comment:Comment:
Wavelet analysis resembles human Wavelet analysis resembles human mental activitymental activity i.e i.e
capability for a detail focusing as well as general and capability for a detail focusing as well as general and
a broadlya broadly--based data analysis based data analysis
Fourier vs Wavelet TransformFourier vs Wavelet Transform
FourierFourier TransformTransform WaveletWavelet TransformTransform
Numerous Analysis structures:
� CWT,
�DWT(2 Band and M Band),
�WP
�DDWT, SWT,
�Adaptive Signal Transform, Numerous Best
basis selection algorithms
�Frame structure
Analysis Structures:
- FS( periodic functions only)
- FT and DFT
Low computational costsComputational Cost High
Many Basis Functions Single Basis Function
Joint Time and Frequency InformationFrequency Information only,
Time/space information is lost
Nonstationary Transient signal AnalysisStationary signal Analysis
Examples of WaveletsExamples of Wavelets
0 5 10 15 20 25-0.2
0
0.2
0.4
0.6
0.8
1
1.2c oif4 : phi
0 5 10 15 20 25-1
-0.5
0
0.5
1
1.5c oi f4 : psi
Scaling function Wavelet function
Coiflet 4
Wavelet Functions Generated from Wavelet Functions Generated from
HAAR functionHAAR function
0 5 10 15 20 25 30 35-2
-1
0
1
2Haar, wavelet packet W13
0 5 10 15 20 25 30 35-2
-1
0
1
2dmey , wavelet packet W10
Gabor 2D WaveletGabor 2D Wavelet
Gabor Wavelet at different scale Gabor Wavelet at different scale and Orientationand Orientation
Laguerre Gaussian WaveletLaguerre Gaussian Wavelet
Laguerre Gaussian WaveletLaguerre Gaussian Wavelet
Wavelet TransformWavelet Transform
��Wavelet Transform is defined as in other Wavelet Transform is defined as in other
transforms:transforms:
�� W W = <= <f(t),f(t),ψ(t)ψ(t)> = > = ∫∫f(t) ψ*(t)dtf(t) ψ*(t)dt
�� wwjkjk = <= <f(t),f(t),ψψjj,,kk(t)(t)> = > = ∫∫f(t) ψf(t) ψj,kj,k*(t)dt*(t)dt
ψj,k(t) is a shifted and scaled wavelet function, j is ψj,k(t) is a shifted and scaled wavelet function, j is
scaling and k translation parameters, J,K can be any scaling and k translation parameters, J,K can be any
scalars. In DWT they assume integer valuesscalars. In DWT they assume integer values..
jkkjkj wtf*
,, )( ψ=
Physical Interpretation of Physical Interpretation of
Wavelet TransformWavelet Transform
��Correlation (extent of matching) of Correlation (extent of matching) of ff(t) with (t) with
the wavelet the wavelet ψψjkjk(t(t) at scale j and location ) at scale j and location k. k.
��It carries signal information at scale j, It carries signal information at scale j, and location k. It gives localized and location k. It gives localized information of a signal.information of a signal.
��Representation of a signal in a domain Representation of a signal in a domain described by wavelet basis functionsdescribed by wavelet basis functions
Wavelet TransformWavelet Transform
Given a function f(t) єL^2Given a function f(t) єL^2
f(t) wab
Ψab(t)= Ψ(atΨab(t)= Ψ(at--b)b)
Function Domain
single dimensional
Transformed Domain:
two dimensional,
paramters a,b
Translation b
Scale a /freq
f(t)
Time/space
coeffs
Illustration of TranslationIllustration of Translation
by two different Waveletsby two different Wavelets f(t)
t
φ0k(t) shifted wavelets
t
φ0k(t) Haar
shifted Haar wavelet
t
Illustrative Examples of WT Illustrative Examples of WT Signal f(t)
t
φ0k(t) shifted wavelets
t t
φ0k(t) Haar φjk(t)
shifted Haar wavelet scaled Haar wavelet t t t
Wavelet TransformWavelet TransformWindow Function InterpretationWindow Function Interpretation
�� Different information are extracted form the same Different information are extracted form the same signal using different wavelets. The need to access signal using different wavelets. The need to access multitude of transforms and wavelet basesmultitude of transforms and wavelet bases for for extraction of different information of a given extraction of different information of a given signalsignal
�� Wavelets as window functions may be considered Wavelets as window functions may be considered as lenses having different resolutions. Different as lenses having different resolutions. Different information are extracted by different lenses at information are extracted by different lenses at different scales and different locationsdifferent scales and different locations of a given of a given signalsignal
�� STFT has only one window function at a given STFT has only one window function at a given scalescale
Wavelet TransformWavelet Transform, , Translation of Window FunctionTranslation of Window Function
Illustration of Wavelet Illustration of Wavelet
Scaling/DilationScaling/Dilation
Common WaveletsCommon Wavelets� Old Wavelets:
� Haar function� Gabor function and wavelet(1D,2D) � Morlet wavelets� Shannon function� 1st and 2nd Derivatives of Gaussian function(mexican hat)� Truncated and lapped Sine or Cosine functions
� Recent Wavelets � Db Wavelets (orthogonal, biorthogonal), � Coiflets, � Symlets� Biorthgonal wavelets� Mallat Wavelet (1D,2D)� Bathlets� Curvelets� Ridgelets� Meyer wavelets� Banana Wavelets� Your Wavelets
� Second Generation Wavelets( wavelets constructed by lifting scheme)
Commonly Used Numerical Commonly Used Numerical
WaveletsWavelets
��Daubechies WaveletsDaubechies Wavelets
��Biorthogonal WaveletsBiorthogonal Wavelets
��CoifletsCoiflets
��SymmletsSymmlets
��BathletsBathlets
��DmeyerDmeyer
How Wavelets are GeneratedHow Wavelets are Generated
��A few number of wavelets are constructed A few number of wavelets are constructed
using known analog functions. Some are using known analog functions. Some are
expressed in analog function form, e.g. expressed in analog function form, e.g.
Morlet waveletsMorlet wavelets
��Many others are designed and constructed Many others are designed and constructed
numericallynumerically. Db wavelets are examples of . Db wavelets are examples of
these waveletsthese wavelets
�� New wavelets are introduced by individuals New wavelets are introduced by individuals
each yeareach year
Categorization of WaveletsCategorization of Wavelets
�� Analytical or Numerical Analytical or Numerical
�� Real or ComplexReal or Complex
�� Symmetric, antisymmetric, asymmetric (e.g. db)Symmetric, antisymmetric, asymmetric (e.g. db)
�� Compactly supported or not compactCompactly supported or not compact
�� Causal, NonCausal, Non--causal waveletscausal wavelets
�� Do they have efficient computation algorithms such as FWTDo they have efficient computation algorithms such as FWT
�� Wavelets with special features e.g. complex wavelet with real part Wavelets with special features e.g. complex wavelet with real part (high freq) and complex part carry low frequency content(high freq) and complex part carry low frequency content
�� Maximally flat waveletes (Flatness of spectrum, Rate of decay at ω=pi Maximally flat waveletes (Flatness of spectrum, Rate of decay at ω=pi and origin)and origin)
�� Have Orthogonal, biorthogonal analysis system Have Orthogonal, biorthogonal analysis system
�� Wavelets of high Resolution in timeWavelets of high Resolution in time--frequency domain: frequency domain: ∆t ∆ω. ∆t ∆ω. Gabor: Gabor: H=0.5, Haar: H=0.58 in normalized frequency (0,1). H=0.5, Haar: H=0.58 in normalized frequency (0,1).
�� Balanced Wavelets (Bathlets) are designed under a criteria based on Balanced Wavelets (Bathlets) are designed under a criteria based on balancing balancing ∆t and ∆ω.∆t and ∆ω.
Two Common Analysis StructuresTwo Common Analysis Structures
11-- Standard Two Band Discrete Wavelet TransformStandard Two Band Discrete Wavelet Transform
High Pass
Low Pass
Low pass
High pass
Decimated Wavelet Transform-Analysis Stage
High Freq Details
Low freq Approx
Signal
2
2
MultiresolutionMultiresolution
Pyramidal Signal Analysis StructurePyramidal Signal Analysis Structure
��Multiresolution as a subband coding of a Multiresolution as a subband coding of a
signal, signal,
��Signal components at different subbands Signal components at different subbands
are extractedare extracted
��Different analyzing wavelets results in Different analyzing wavelets results in
different components at different bands different components at different bands
An Illustration of MultiresolutionAn Illustration of Multiresolution
Signal DecompositionSignal Decomposition
Binary Tree of DWTBinary Tree of DWT
Common Analysis StructuresCommon Analysis Structures
Wavelet PacketsWavelet Packets
High Pass
Low Pass
2
2
2
Low pass
High pass
Wavelet Packets -Analysis Stages
2
High freq details
High pass
Low pass 2
2
V space of the signal
Wavelet Packets TreeWavelet Packets Tree
Standard Two Channel DWT and Standard Two Channel DWT and
WP TWP T--F TilingF Tiling
TimeTime--Freq. TilingFreq. Tiling
Standard DWT Standard WP
Perfect Reconstruction FilterbankPerfect Reconstruction Filterbank
Filterbank implementation of DWTFilterbank implementation of DWT
High Pass
Low Pass Low pass
High pass
Perfect Reconstrution Filter bank
Wavelet
Subspace W
Alternative Signal Analysis Alternative Signal Analysis
ArchitecturesArchitectures
Wavelets are rich in analysis structures and Wavelets are rich in analysis structures and
algorithmsalgorithms
Main Analysis StructuresMain Analysis Structures
�� ORTHOGONALORTHOGONAL
�� BIORTHOGONALBIORTHOGONAL
�� REDUNDANT WAVELET TRANSFROMREDUNDANT WAVELET TRANSFROM
�� FRAMEFRAME--BASED REDUNDANT BASED REDUNDANT
TRANSFORMTRANSFORM
Beyond Orthogonality: Compression Beyond Orthogonality: Compression
RequirementsRequirements
Slow rate of decay Fast rate of decay
Coeffs 2Coeffs 1
Orthogonal Expansion Non-orthogonal Expansion
Primary Features of Wavelets used Primary Features of Wavelets used
in Signal Processingin Signal Processing
1.1. TimeTime-- frequency Localizationfrequency Localization (useful for transient (useful for transient data analysis)data analysis)
2.2. Sparsity: Sparsity: Coefficients are often grouped into high Coefficients are often grouped into high and low amplitudeand low amplitude
3.3. Function Approximation: Function Approximation: near Optimal Function near Optimal Function ApproximationApproximation
4.4. Noise Reduction, Noise Reduction, No wavelet can model white noise No wavelet can model white noise leading to small amplitude coefficientsleading to small amplitude coefficients
5.5. Decorrelation: Signal components( coefficients) in Decorrelation: Signal components( coefficients) in wavelet domain have a lower degree of correlation wavelet domain have a lower degree of correlation
Categorization ofCategorization of
Wavelet ApplicationsWavelet Applications
Applications: Methodology/Algorithm IntensiveApplications: Methodology/Algorithm Intensive
�� Change Detection ( of all kinds)Change Detection ( of all kinds)
�� CompressionCompression
�� Feature Extraction and Pattern Recognition with Numerous ApplicationsFeature Extraction and Pattern Recognition with Numerous Applications
�� Image Enhancement Image Enhancement
�� Noise reduction with Numerous ApplicationsNoise reduction with Numerous Applications
�� Diagnostics (such as NDT, NDE)Diagnostics (such as NDT, NDE)
�� Signal Separation, ICA using WaveletsSignal Separation, ICA using Wavelets
�� Solution of Differential and Partial Differential EquationsSolution of Differential and Partial Differential Equations
�� Analysis of FractalsAnalysis of Fractals
�� System Identification in Control systemsSystem Identification in Control systems
�� Data and Image FusionData and Image Fusion
�� WatermarkingWatermarking
Applications, Industry Applications, Industry
CategorizationCategorization
Application Areas: Industry IntensiveApplication Areas: Industry Intensive
�� Power Systems, Transient data Analysis, Component Diagnosis, Signal Power Systems, Transient data Analysis, Component Diagnosis, Signal detectiondetection
�� NDT, Ultra sonic Flaw Detection, DiagnosisNDT, Ultra sonic Flaw Detection, Diagnosis
�� Communications, Coding, TransmissionCommunications, Coding, Transmission
�� Medical Applications: Compression, Noise Reduction, signal separation, Medical Applications: Compression, Noise Reduction, signal separation, signal classification, image enhancementsignal classification, image enhancement
�� Machine Vision, Robotics, geometry and spatial edge detectionMachine Vision, Robotics, geometry and spatial edge detection
�� Industrial Machinery and Machine Diagnosis,Industrial Machinery and Machine Diagnosis,
�� Remote Sensing, Satellite Image Analysis, Image segmentation, fusionRemote Sensing, Satellite Image Analysis, Image segmentation, fusion
�� Biometrics: Fingerprint, Iris and Facial IdentificationBiometrics: Fingerprint, Iris and Facial Identification
�� Nuclear Reactors, Boiler Pipe fault analysis, Nuclear Reactors, Boiler Pipe fault analysis,
�� Oil Industry, Oil ExplorationOil Industry, Oil Exploration
�� Transportation Industry (railroad engine diagnosis, Aero jet engine and Transportation Industry (railroad engine diagnosis, Aero jet engine and Helicopter drive and gear diagnosisHelicopter drive and gear diagnosis
�� Seismic Data analysisSeismic Data analysis
�� GeneticsGenetics
Applications, Fields of WorkApplications, Fields of Work
Applications in different fieldsApplications in different fields�� Physics, Geophysics, Physics, Geophysics,
�� Mathematics, function approximation, signal Mathematics, function approximation, signal processing, Harmonic analysis, Differential processing, Harmonic analysis, Differential Equation, Frames, Function Space AnalysisEquation, Frames, Function Space Analysis
�� Electrical Engineering, numerous aplicationsElectrical Engineering, numerous aplications
�� Biomedical Engineering, EEG, ECG, MRIBiomedical Engineering, EEG, ECG, MRI
�� Finance and Financial Data AnalysisFinance and Financial Data Analysis
�� Mechanical, Civil, Ocean Science and EngineeringMechanical, Civil, Ocean Science and Engineering
�� Remote Sensing, ForestryRemote Sensing, Forestry
�� Nontechnical: Historial document retrievalNontechnical: Historial document retrieval
Details of ApplicationsDetails of Applications
�� To Be Presented in a Separate To Be Presented in a Separate
�� Slide Presentation Slide Presentation