Wavelets (Chapter 7)

CS474/674 Prof. Bebis

STFT (revisited)
Time/Frequency localization depends on window size.
Once you choose a particular window size, it will be the same
for all frequencies.
Many signals require a more flexible approach - vary the
window size to determine more accurately either time or
frequency.

The Wavelet Transform

Overcomes the preset resolution problem of the STFT by
using a variable length window:
Narrower windows are more appropriate at high frequencies
(better time localization)
Wider windows are more appropriate at low frequencies
(better frequency localization)

The Wavelet Transform (contd)

Wide windows do not provide good localization

at high frequencies.

The Wavelet Transform (contd)

A narrower window is more appropriate at high frequencies.

The Wavelet Transform (contd)

Narrow windows do not provide good localization

at low frequencies.

The Wavelet Transform (contd)

A wider window is more appropriate at low frequencies.

What is a wavelet?
It is a function that waves above and below the x-axis;
it has (1) varying frequency, (2) limited duration, and (3)
an average value of zero.
This is in contrast to sinusoids, used by FT, which have
infinite energy.
Sinusoid

Wavelet

Wavelets
Like sine and cosine functions in FT, wavelets can define
basis functions k(t):

f (t ) ak k (t )
k

Span of k(t): vector space S containing all functions f(t)

that can be represented by k(t).

Wavelets (contd)
There are many different wavelets:
Haar

Morlet

Daubechies

Basis Construction - Mother Wavelet

jk (t )

(dyadic/octave grid)

Basis Construction - Mother Wavelet

(contd)
jk (t ) 2 j / 2 2 j t k

scale =1/2j
(1/frequency)

Continuous Wavelet Transform (CWT)

translation parameter,
measure of time

Forward
CWT:

scale parameter
(measure of frequency)

1
C ( , s )
s

normalization
constant

t
f t
dt
s

Mother wavelet
(window)
scale =1/2j=1/frequency

CWT: Main Steps

1. 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. The higher C is, the more the similarity.

CWT: Main Steps (contd)

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

CWT: Main Steps (contd)

4. Scale the wavelet and repeat steps 1 through 3.

5. Repeat steps 1 through 4 for all scales.

Coefficients of CTW Transform

Wavelet analysis produces a time-scale view of the input
signal or image.

C ( , s)

1
s

t
dt
s

f t

Continuous Wavelet Transform (contd)

Inverse CWT:

1
f (t )
s

t
s C ( , s) ( s )d ds
note the double integral!

FT vs WT
weighted by F(u)

weighted by C(,s)

Properties of Wavelets
Simultaneous localization in time and scale
- The location of the wavelet allows to explicitly represent

the location of events in time.

- The shape of the wavelet allows to represent different
detail or resolution.

Properties of Wavelets (contd)

Sparsity: for functions typically found in practice, many of the
coefficients in a wavelet representation are either zero or very
small.
1
f (t )
s

t
s C ( , s) ( s )d ds

Adaptability: Can represent functions with discontinuities or

corners more efficiently.
Linear-time complexity: many wavelet transformations can be
accomplished in O(N) time.

Discrete Wavelet Transform (DWT)

a jk f (t ) *jk (t )

(forward DWT)

f (t ) a jk jk (t ) (inverse DWT)
k
j
j/2
j

(
t
)

2
t k
jk
where

DFT vs DWT
FT expansion:

one parameter basis

or

f (t ) a l l (t )
l

two parameter basis

WT expansion

f (t ) a jk jk (t )
k

Multiresolution Representation using

jk

(t )

f (t )

fine
details

wider, large translations

f (t )
k

a
j

jk (t )

jk

j
coarse
details

Multiresolution Representation using

jk

(t )

f (t )

fine
details

f (t )
k

a
j

jk (t )

jk

j
coarse
details

Multiresolution Representation using

jk

(t )

f (t )

narrower, small translations

f (t )
k

a
j

jk (t )

jk

fine
details

j
coarse
details

Multiresolution Representation using

jk

(t )

high resolution

f (t )

(more details)

f1 (t )
f2 (t )

low resolution

fs (t )

(less details)

f (t )
k

a
j

jk (t )

jk

Prediction Residual Pyramid - Revisited

In the absence of quantization errors, the approximation
pyramid can be reconstructed from the prediction residual
pyramid.
Prediction residual pyramid can be represented more
efficiently.

(with sub-sampling)

Efficient Representation Using Details

details D3

details D2

details D1

L0
(without sub-sampling)

Efficient Representation Using Details (contd)

L0 D1 D2 D3
in general: L0 D1 D2 D3DJ

representation:

A wavelet representation of a function consists of

(1)a coarse overall approximation
(2)detail coefficients that influence the function at various scales.

Reconstruction (synthesis)
H3=H2+D3

H2=H1+D2
details D3

details D2

H1=L0+D1

details D1
L0
(without sub-sampling)

Example - Haar Wavelets

Suppose we are given a 1D "image" with a resolution
of 4 pixels:

[9 7 3 5]
The Haar wavelet transform is the following:

L 0 D 1 D2 D3

(with sub-sampling)

Example - Haar Wavelets (contd)

Start by averaging the pixels together (pairwise) to get
a new lower resolution image:
[9 7 3 5]

To recover the original four pixels from the two

averaged pixels, store some detail coefficients.
1

Example - Haar Wavelets (contd)

Repeating this process on the averages (i.e., low
resolution image) gives the full decomposition:

Harr decomposition:

Example - Haar Wavelets (contd)

We can reconstruct the original image by adding or

subtracting the detail coefficients from the lowerresolution versions.
2

1 -1

Example - Haar Wavelets (contd)

Detail coefficients
become smaller and
smaller as j increases.

Dj
Dj-1

L0

D1

How to
compute Di ?

Multiresolution Conditions
If a set of functions can be represented by a weighted
sum of (2jt - k), then a larger set,
set including the
original, can be represented by a weighted sum of
(2j+1t - k):
high
resolution

j
low
resolution

Multiresolution Conditions (contd)

If a set of functions can be represented by a weighted
sum of (2jt - k), then a larger set, including the
original, can be represented by a weighted sum of
(2j+1t - k):
Vj: span of (2jt - k):

f j (t ) ak jk (t )

Vj+1: span of (2j+1t - k):

f j 1 (t ) bk ( j 1) k (t )

V j V j 1

Nested Spaces Vj
Vj : space spanned by (2jt - k)

Basis functions:
(t - k)

V0

(2t - k)

f(t) Vj

f (t )
k

jk (t )

jk

V1

(2jt - k)

Vj

Multiresolution conditions nested spanned spaces: V j V j 1

i.e., if f(t) V j then f(t) V j+1

How to compute Di ? (contd)

f(t) Vj

f (t )
k

jk (t )

jk

IDEA: define a set of basis

functions that span the
difference between Vj+1 and
Vj

Orthogonal Complement Wj
Let Wj be the orthogonal complement of Vj in Vj+1
Vj+1 = Vj + Wj

How to compute Di ? (contd)

If f(t) Vj+1, then f(t) can be represented using basis
functions (t) from Vj+1:

f (t ) ck (2 j 1 t k )

Vj+1

Alternatively, f(t) can be represented using two sets of basis functions,

(t) from Vj and (t) from Wj:
Vj+1 = Vj + Wj

f (t ) ck (2 j t k ) d jk (2 j t k )
k

How to compute Di ? (contd)

Think of Wj as a means to represent the parts of a function
in Vj+1 that cannot be represented in Vj

f (t ) ck (2 j 1 t k )
k

f (t ) ck (2 j t k ) d jk (2 j t k )
k

Vj

Wj

differences
between
Vj and Vj+1

How to compute Di ? (contd)

Vj+1 = Vj + Wj using recursion on Vj:

Vj+1 = Vj-1+Wj-1+Wj = = V0 + W0 + W1 + W2 + + Wj
if f(t) Vj+1 , then:
f (t ) ck (t k ) d jk (2 j t k )
k

V0
basis functions

W0, W1, W2,

basis functions

Summary: wavelet expansion (Section 7.2)

Wavelet decompositions involve a pair of waveforms
(mother wavelets):
encode low
resolution info

(t)

(t)

encode details or
high resolution info

f (t ) ck (t k ) d jk (2 j t k )
k

scaling function

wavelet function

1D Haar Wavelets
Haar scaling and wavelet functions:

(t)

computes average

(t)

computes details

1D Haar Wavelets (contd)

V0 represents the space of one pixel images
Think of a one-pixel image as a function that is constant
over [0,1)

Example:

width: 1
0

1D Haar Wavelets (contd)

V1 represents the space of all two-pixel images
Think of a two-pixel image as a function having 21
equal-sized constant pieces over the interval [0, 1).
Examples:

Note that
e.g.,

width: 1/2

V0 V1
=

1D Haar Wavelets (contd)

V j represents all the 2j-pixel images
Functions having 2j equal-sized constant pieces over
interval [0,1).
Examples:

width: 1/2j

Note that V j 1 V j

e.g.,

Vj

Vj

1D Haar Wavelets (contd)

V0, V1, ..., V j are nested
i.e.,

V j V j 1

VJ fine details

V1
V0 coarse details

Define a basis for Vj

Mother scaling function:
1
0

Lets define a basis for V j :

i ( x ) ji ( x )
note alternative notation:

Define a basis for Vj (contd)

Define a basis for Wj (contd)

Mother wavelet function:
1
-1
0

Lets define a basis ji for Wj :

note alternative notation:

i ( x ) ji ( x )

1/2

Define a basis for Wj

Note
that the dot
product
between basis
functions in Vj
and Wj is zero!

Define a basis for Wj (contd)

Basis functions ji of W j
Basis functions ji of V j

form a basis in V j+1

Define a basis for Wj (contd)

V3 = V2 + W2

Define a basis for Wj (contd)

V2 = V1 + W1

Define a basis for Wj (contd)

V1 = V0 + W0

Example - Revisited

f(x)=

V2

Example (contd)
f(x)=
2,0(x)
V2

2,1(x)
2,2(x)
2,3(x)

Example (contd)
(divide by 2 for normalization)

V1and W1
V2=V1+W1

1,0(x)
1,1(x)
1,0(x)
1,1(x)

Example (contd)

Example (contd)
(divide by 2 for normalization)

V0 ,W0 and W1
V2=V1+W1=V0+W0+W1

0,0(x)
0,0(x)

1,0 (x)
1,1(x)

Example

Example (contd)

Filter banks
The lower resolution coefficients can be calculated
from the higher resolution coefficients by a treestructured algorithm (i.e., filter bank).

Subband
encoding
(analysis)

h0(-n) is a lowpass filter and h1(-n) is a highpass filter

Example (revisited)
[9 7 3 5]
low-pass,
down-sampling

(9+7)/2

high-pass,
down-sampling

(3+5)/2

(9-7)/2 (3-5)/2

V1 basis functions

Filter banks (contd)

Next level:

Example (revisited)
[9 7 3 5]

low-pass,
down-sampling

high-pass,
down-sampling

V1 basis functions

(8+4)/2

(8-4)/2

Convention for illustrating

1D Haar wavelet decomposition
x

x
average
detail

re-arrange:

V1 basis functions
re-arrange:

Examples of lowpass/highpass
(analysis) filters

Haar

h0
h1

h0
Daubechies
h1

Filter banks (contd)

The higher resolution coefficients can be calculated from
the lower resolution coefficients using a similar structure.

Subband
encoding
(synthesis)

Filter banks (contd)

Next level:

Examples of lowpass/highpass
(synthesis) filters
Haar (same as
for analysis):

g0
g1
g0

Daubechies
+

g1

2D Haar Wavelet Transform

The 2D Haar wavelet decomposition can be computed
using 1D Haar wavelet decompositions.
i.e., 2D Haar wavelet basis is separable

Two decompositions (i.e., correspond to different

basis functions):
Standard decomposition
Non-standard decomposition

Standard Haar wavelet decomposition

Steps
(1) Compute 1D Haar wavelet decomposition of each row of
the original pixel values.
(2) Compute 1D Haar wavelet decomposition of each
column of the row-transformed pixels.

Standard Haar wavelet decomposition

(contd)
average
detail

(1) row-wise Haar decomposition:

re-arrange terms

xxx
xxx

xxx

...

x
x
.
x

Standard Haar wavelet decomposition

(contd)
average
detail

(1) row-wise Haar decomposition:

row-transformed result

Standard Haar wavelet decomposition

(contd)
average
detail

(2) column-wise Haar decomposition:

row-transformed result

column-transformed result

Example

row-transformed result

re-arrange terms

Example (contd)

column-transformed result

2D Haar basis for standard decomposition

To construct the standard 2D Haar wavelet basis, consider
all possible outer products of 1D basis functions.
0,0(x)

Example:
V2=V0+W0+W1

0,0(x)
1,0(x)
1,1(x)

2D Haar basis for standard decomposition

To construct the standard 2D Haar wavelet basis, consider
all possible outer products of 1D basis functions.

00(x), 00(x)

00(x), 00(x)

i j ( x) ji ( x)

10(x), 00(x)

i j ( x) ji ( x)

2D Haar basis of standard decomposition

V2

i j ( x) ji ( x)
i j ( x) ji ( x)

Non-standard Haar wavelet decomposition

Alternates between operations on rows and columns.
(1) Perform one level decomposition in each row (i.e., one
step of horizontal pairwise averaging and differencing).
(2) Perform one level decomposition in each column from
step 1 (i.e., one step of vertical pairwise averaging and
differencing).
(3) Repeat the process on the quadrant containing averages
only (i.e., in both directions).

Non-standard Haar wavelet decomposition

(contd)
one level, horizontal
Haar decomposition:
xxx
xxx

xxx

...

x
x
.
x

one level, vertical

Haar decomposition:

Non-standard Haar wavelet decomposition

(contd)
re-arrange terms

one level, horizontal

Haar decomposition
on green quadrant

one level, vertical

Haar decomposition
on green quadrant

Example

re-arrange terms

Example (contd)

2D Haar basis for non-standard decomposition

Defined through 2D scaling and wavelet functions:

( x, y ) ( x) ( y )

000 ( x, y ) ( x, y )

( x, y ) ( x) ( y )
( x, y ) ( x) ( y )
( x, y ) ( x) ( y )

kfj ( x, y ) 2 j (2 j x k , 2 j y f )
kfj ( x, y ) 2 j (2 j x k , 2 j y f )
kfj ( x, y ) 2 j (2 j x k , 2 j y f )

2D Haar basis for non-standard decomposition

(contd)

i j ( x) ji ( x)

V2

i j ( x) ji ( x)

2D Haar basis for non-standard decomposition

(contd)
Three sets of detail coefficients (i.e., subband encoding)

000 ( x, y ) ( x, y ) LL: average LL

( x, y ) 2 (2 x k , 2 y f )
j
kf

kfj ( x, y ) 2 j (2 j x k , 2 j y f )

Detail coefficients
LH: intensity variations along
columns (horizontal edges)

HL: intensity variations along

kfj ( x, y ) 2 j (2 j x k , 2 j y f ) HH:

rows (vertical edges)

intensity variations along
diagonals

2D DWT using filter banks (analysis)

LL

LH

HL

HH

2D DWT using filter banks (analysis)

LL

LH

HL

HH

LL

LL
H LH
V HL
D HH

The wavelet transform is applied again on the LL sub-image

2D DWT using filter banks (analysis)

2D Inverse DWT using filter banks

(synthesis)

H LH
V HL
D HH

Wavelets Applications
Noise filtering
Image compression
Fingerprint compression

Image fusion
Recognition
G. Bebis, A. Gyaourova, S. Singh, and I. Pavlidis, "Face Recognition by
Fusing Thermal Infrared and Visible Imagery", Image and Vision
Computing, vol. 24, no. 7, pp. 727-742, 2006.

Image matching and retrieval

Charles E. Jacobs Adam Finkelstein David H. Salesin, "Fast
Multiresolution Image Querying", SIGRAPH, 1995.

Fast Multiresolution Image Querying

painted

low resolution

target

queries
Charles E. Jacobs Adam Finkelstein David H. Salesin, "Fast Multiresolution
Image Querying", SIGRAPH, 1995.