Transcript

Introduction to Wavelets

Olof Runborg

Numerical Analysis,School of Computer Science and Communication, KTH

RTG Summer School on Multiscale Modeling and AnalysisUniversity of Texas at Austin

2008-07-21 – 2008-08-08

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 1 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet multiresolution decomposition

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 2 / 8

Wavelet examples

Haar

Scaling function φ(x) Mother wavelet ψ(x)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 3 / 8

Wavelet examples

Daubechies 4

Scaling function φ(x) Mother wavelet ψ(x)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 3 / 8

Wavelet examples

Daubechies 6

Scaling function φ(x) Mother wavelet ψ(x)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 3 / 8

Wavelet examples

Daubechies 8

Scaling function φ(x) Mother wavelet ψ(x)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 3 / 8

Wavelet based image compression

Wavelets successful in image compression. Eg: JPEG 2000standard (Daubechies (9,7) biorthogonal wavelets), FBI fingerprintdatabase, . . .Consider image as a function and use 2D wavelets.

Compress by thresholding + coding + quantization.

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 4 / 8

Wavelet based image compression

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 5 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Time representation — total localization in time

f (n) =∑

k

f (k)δ(n − k)

Basis functions

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Time representation — total localization in time

f (n) =∑

k

f (k)δ(n − k)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Fourier representation — total localization in frequency

f (n) =∑

j

f̂ (j)exp(inj2π/N)

Basis functions

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Fourier representation — total localization in frequency

f (n) =∑

j

f̂ (j)exp(inj2π/N)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Wavelet representation — localization in time and frequency

f (n) =∑j,k

wj,kψj,k (n/2π)

Basis functions

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Wavelet representation — localization in time and frequency

f (n) =∑j,k

wj,kψj,k (n/2π)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of Signals

Given a discrete signal f (n), n = 0,1, . . ..

Wavelet representation — localization in time and frequency

f (n) =∑j,k

wj,kψj,k (n/2π)

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 6 / 8

Time Frequency Representation of SignalsExample

Time representation — total localization in time

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 7 / 8

Time Frequency Representation of SignalsExample

Fourier representation — total localization in frequency

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 7 / 8

Time Frequency Representation of SignalsExample

Wavelet representation — localization in time and frequency

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 7 / 8

Wavelets – some contributors

Strömberg – first continuous waveletMorlet, Grossman – "wavelet"Meyer, Mallat, Coifman – multiresolution analysisDaubechies – compactly supported waveletsBeylkin, Cohen, Dahmen, DeVore – PDE methods using waveletsSweldens – lifting, second generation wavelets

. . . and many, many more.

Olof Runborg (KTH) Introduction to Wavelets Austin, July 2008 8 / 8