Chapter 3

MULTI-RESOLUTION TECHNIQUES: THEORETICAL ASPECTS

The most commonly used image compression methods consist of three steps Source
Encoder (Linear Transform) Quantization and Entropy Coding. The first step, i.e. the
transform, may be accomplished using Wavelets. A signal can be processed by any of the
available mathematical tool called as transform (An algorithm or procedure that changes
one group of data into another group of data). The transformation is used to accomplish
the following: 1. to create a representation for the data in which there is less correlation
among the co-coefficient values and it is often referred as decorrelating the data. The
purpose of decorrelation is to reduce the redundancy, which enables us to quantize each
coefficient independently of the others. 2. To have a representation in which it is possible
to quantize different co-ordinates with different precision [1]-[7]. The mathematical
transforms used to process the signal like FT (Fourier Transform), STFT (Short Term
Fourier Transform), DCT (Discrete Cosine transform), Laplace Transform, Z Transform,
Hilbert Transform and Wavelets etc. [8]]-[15], are normally applied to the signals to
obtain information from that signal which is not readily available in the raw signal. Here
the raw signals in practice are time domain. When we plot Time Domain signals we
obtain a time amplitude representation of the signal. This representation is not always the
best representation of the signal for most signals processing application, because most of
the information of signal is hidden in the frequency content. The frequency spectrum of

55

signal shows what frequencies exist in the signal (frequency is something to do with the
change in rate). We can measure the frequency contents of signal using Fourier Transform
(A family of mathematical techniques based on decomposing signals into sinusoids).
Fourier Transform does not indicate the time in which time these frequency components
exist in the signal. Hence these signals are called the Stationary signals The theory of
Wavelets starts with the concepts of Multiresolution Analysis (MRA) [7],[16]-[25] ; the
detailed theory of Multiresolution is found at [5][[16][26][27]. Before dealing with
Multiresolution techniques let us first see the mathematical background for the
transforms.
3.1 Vector spaces:
Consider an example of 2-D vector space is

Figure 3.1: Vector Space

The above vector shown in Figure 3.1 can be represented in many ways like.

In terms of magnitude and direction.

In terms of weighted sum of the vector in the x & y directions.

As an array whose components are the coefficients of the unit vectors?

As shown in Figure3.1 Ux = 4, Uy = 3. Therefore V= 4Ux+3Uy.

Thus the Vector V has magnitude 5 and angle 36.86 degrees.

V= 4Ux+3Uy(3.1)
4
V
3
We can view the second representation of Vector as a decomposition of Vector V into
simpler building blocks. i.e. any Vector in 2-D can be decomposed exactly the same way

56

as said in above equation 3.1. We can get the coefficients of Vector through a product
called the Dot product or Inner product.
3.1.1 Dot or Inner product:
Let us consider two vectors, Vector A and Vector B
Such that
A1
A
A2

B1
B ......(3.2)
B 2

The inner product between vector A and Vector B defined as

A.B= A1.B1+A2.B2.....(3.3)
Two vectors are said to be orthogonal if their inner product is zero.
A set of vectors is said to be orthogonal if each vector in the set is orthogonal to
every other vector in the set.
The inner product between a vector and a unit vector from an orthogonal basis set
will give us the coefficient corresponding to the unit vector we can write Ux and Uy as
1
Ux
0

0
Uy ..(3.4)
1

Therefore the coefficients of Ux can be obtained by

A.Ux = A1.1+A2.0=A1 ...(3.5)
and the coefficient of Uy can be obtained by
B.Uy = A1.0+A2.1=A2...(3.6)
3.1.2 Vector Space:
In order to handle 2D, 3DnD we need to generalize the above concepts. A vector space
consists of set of elements called vectors addition, multiplication, etc., furthermore, the
results of these operations are also elements of the vector space.
Vector addition:
The Vector addition can be obtained by the point wise addition of the components of
two vectors.
For. Ex. Consider vector A and Vector B

57

 A1
B1

A A2
B B 2 ..(3.7)
A3
B3
A1 B1
A B A2 B 2 .(3.8)
A3 B3

Scalar Multiplication:
We mean that the multiplication of vector with real or complex numbers.
Suppose Vector V is a vector space X, Y, Z are the vector elements: and are the
two scalar vectors then the following axioms are satisfied
X+Y = Y+X (Commutative)
(X+Y)+Z=X+(Y+Z) and ( ) X = (X) (Associative)
Their exist element in V that X+ = X for all X in is called the additive
identity
(X+Y)= X+ Y and ( + ) X = X + X (Distributive)
1.X = X and 0.X =
For every X in V there exit a (-X) such that X+(-X)= .

3.1.3 Subspace:
A subspace S of a Vector V is a subset of V whose members satisfy all the axioms of the
vector space and has the additional property that if X & Y are in S and is a scalar, then
X+Y and X is also in S.

3.1.4 Basis:
One way with which we can generate a subspace is by taking linear combinations of a set
of vectors. If this set of vectors is linearly independent then the set is called a basis for the
subspace. A set of vectors {X1, X2} is said to be linearly independent if no vectors of
the set can be written as a linear combination of the other vectors in the set.

58

We now have sufficient information in hand to begin looking at some of well-known

techniques for representing functions of time. Basically, the important ideas that we
should remember are:
- Vectors are not simply points in two or Three Dimensional space. Infarct, functions
of time can be viewed as an element in a vector space.
- All members of vector space can be represented as linear or weighted combinations
of the basis vector.
- If a basis set is orthonormal (A basis vector has a unit magnitude and are orthogonal
they are known as orthonormal basis set) the weights or coefficient can be obtained by
taking the inner product.
We can use the above concepts to represent periodic functions as linear combinations of
Sines and Cosines. [5]

3.2 Fourier series

The representation of Periodic functions in terms of series of sines and cosines was
discovered by Joseph Fourier is called as Fourier series.
Any periodic functions can be represented as the sum of smooth, well-behaved sines and
cosines, which are given a periodic function f (t) with period of T
i.e. f(t) = f(t+nT) ..(3.9)
Where
n= 1,2,3..
We can represent the Fourier series f(t) as

f (t ) a 0

an cos n 0 t
n 1

n 1

bn sin 0 t ...(3.10)

Where
w0 2 / T

an and bn are the coefficients of f(t)

A more useful form of the Fourier series representation from this work point of view is
the exponential form of Fourier series,

f (t )

Cne

jn 0 t

... (3.11)

59

We can easily move between the exponential and trigonometric representation by using
Eulers identity

cos

j sin

.....(3.12)

Where
j 1

We can find the coefficients of Cn by taking the inner product of f(t) and the basis vector
ejnw0t
T
jn 0 t
1 / T f (t ).e jn 0 t ....(3.13)
i.e. Cn f (t ), e
0

i.e. the fourier representation Cn

of function f(t)
n

The basis functions are sinusoids that differ from each other in how fast they fluctuate in
a given time interval. The basis vector e2jw0t fluctuate twice as fast as the basis vector ejw0t.
The coefficients of the basis Vectors Cn

gives us measure of the different amounts

of fluctuations. A fluctuation of this sort is usually measured in terms of frequency. For

example a frequency 1Hz denotes the completion of one period in one second, 2 Hz
denotes the completion of two periods on one second, 3 Hz denotes the completion of 3

periods in one second and so on. Thus the coefficient Cn

provides the frequency

n

profile. The amount of signal change is obtained by the rate in which the signal changes
at w0/2, 2w0/2, and 3w0/2 and so on. This information cannot be obtained by

looking time representation f (t). On the other hand the use of Cn

representation
n

tells us little about how the signal changes with time. The ability to view the same signal
in different ways helps us to understand the nature of the signal and also we need to
develop tools for manipulation of the signal. Thus Fourier series provides us with
frequency representation of periodic signals. However many of the signals we will be
dealing within day-to-day activities is a periodic (not periodic). Fortunately, the Fourier
series concepts can be extended to non-periodic signals also.
For ex. Stationary (periodic) signal

60

Figure 3.2: Stationary or Periodic signal

3.2.1 Fourier Transform
Fourier Transform (FT) is a standard tool for analyzing a given function f (t). We can
find the FT of given function as F ( )

F ( ) f (t )e jt dt . (3.14)

Fourier transform is suitable for Processing stationary signals or periodic signals, and not
suitable for non-periodic signals or Non stationary signals. For a Non-stationary signal
we can assume that some portion of the signal is stationary. Let us consider a nonstationary signal as shown in figure-3.2.1

61

Figure-3.2.1: Non stationary signal

Through FT we can obtain frequency components of the signal. But FT doesnt infer the
time instant of appearance of a particular frequency component in the frequency
spectrum. We may some times need the time instant of appearance of a particular
frequency component. One way to obtain this information is through the Short Term
Fourier Transform (STFT). In STFT we can consider some part of the signal to be
stationary.

3.2.2 Short Term Fourier Transform (STFT)

Thus a revised version of FT called Short Time Fourier Transform (STFT) is used for the
analysis of non-stationary signals. There is only minor difference between STFT and FT.
In STFT, the signal is divided into small enough segments, where these segments
(portions) of the signal can be assumed to be stationary. For this purpose, a window
function is chosen. The width of this window must be equal to the segment of the
signal where its stationary is valid.

62

In order to reduce the boundary effect we make the windowing ( ) each piece of before
we take the Fourier Transform. For Ex. 5khz may appear at T, 10 kHz appear at 2T, 20
kHz may appear at 3T. Thus we obtain an analysis of function of both time and frequency

Figure 3.2.2(a): Windowing of signal

i.e. If the window shape is given by g(t), then the STFT is given by

F ( , T )

f (t ) g * (t T ) e

j t

dt (3.15)

Figure 3.2.2(b) shows the time-frequency plane of a windowed Fourier transforms,

where the window is a square wave. It also shows what the basis function (sine or cosine)
looks alike as the frequency increases. The square wave window fits the sine or cosine
function to a window of a particular width. Because this same window is used for all
frequencies, the resolution of the analysis at every point in the plane is identical.

Figure 3.2.2(b) Time-frequency plane of a windowed Fourier Transform

63

3.2.3 Problem with STFT:

The problem with STFT is fixed window size. Consider a non-stationary signal as shown
in figure 3.7. The window size should be at least t 0 to get at least one cycle of the low
frequency component. However the window size t0 is greater means that we will not able
to localize the high frequency. i.e. we are unable to fix the window size because we do
not know the exact time frequency representation of a signal this is called the Heisenberg
Uncertainty principle i.e. If we use window of infinite length we get the FT, which
gives the perfect frequency resolution, but no time information. Furthermore, in order to
obtain the stationary signal, we have to have a short window, in which the signal is
stationary. The narrower Window

gives the good time resolution but poor frequency

resolution. The Wider Window gives the good frequency resolution but poor time
resolution. [16] This representation problem is depicted in the figures below. The Figure
3.2.3 (a) is the STFT of the above signal for a narrow window, which has very good time
resolution but relatively poor frequency resolution.

Figure 3.2.3 (a) STFT with Narrow Window

The Figure 3.2.3 (b) shows STFT for a much wider window and in this case the time
resolution degrades and frequency resolution increases.

64

Figure 3.2.3 (b): STFT with Wide Window

Thus it can be seen that fixed windowed Fourier transform is not a suitable tool as there
is a resolution trade off in terms of time and frequency. This above problem is called
resolution problem.
Now let us consider discrete STFT in terms of basis expansion and for the moment just
take a one interval

F ( m ,0 )

f (t ) g * te

jm 0 t

dt .....(3.16)

The basis function are g(t), g(t) e j 0t , g(t) e j 2 0t , and so on, the first three basis functions
are shown in figure 3.8. We can see that window is of constant size and with this window
we have sinusoids with an increasing number of cycles. Let us assume for instant number
of cycles are fixed but the size of the window keeps changing as shown in figure 3.12.

Figure 3.12: Windowing function

65

It clearly shows that the lower frequency function covers the long interval time, while
higher frequency covers the short time interval. This is called the frequency problem and
which is happening with STFT. Now STFT faces the problem of resolution, then question
arises about the use of proper kind of window it is because the narrower Window gives
the good time resolution but poor frequency resolution. The Wider Window gives the
good frequency resolution but poor time resolution. The problem is of course choosing
the window function (window size)[16][26][29][30][31].
3.3 Multi Resolution Analysis(MRA)
Although the time and frequency resolution problem are result of physical
phenomenon (the Heisenberg Uncertainty Principle) and exists regardless of the
transform used, it is possible to analyze any signal by using an approach called the Multi
Resolution Analysis (MRA). MRA, as the name implies, analyzes the signal at different
frequencies with different resolutions. Every spectral component is not resolved equally
as was the case in the STFT.
MRA is designed to give good time resolution and poor frequency resolution at
high frequencies and good frequency resolution and poor time resolution at low
frequency. This approach makes sense especially when signal at hand has high frequency
components for short durations and low frequency components for long durations for
example signal like non-stationary signal as shown in figure 3.3(a) and the time
frequency resolution as shown in figure 3.3(b). Fortunately, the signals that are
encountered in practical applications, in particular in the case of audio, are often of this
type.

Figure 3.3(a) Non-Stationary signals.

66

Figure 3.3(b) Resolution of Time & Frequency

Wavelet transforms (WT) overcome the aforementioned resolution problem. Every box
in Figure 3.3(c) has a constant area and therefore represents equal portions of the time frequency plane, but different proportions are allocated to time and frequency. The short
high-frequency basis functions and the long low-frequency ones enable detailed time and
frequency analysis to be performed at the same time. In short, the wavelet transform is
well localized in frequency and time. The data can also be processed on different scales
or resolutions unlike the fixed resolution of the STFT.

Figure 3.3(c) Time-frequency plane of a Wavelet Transform

The idea behind the multi resolution analysis is fairly simple. Let us define a function

(t ) called scaling function. By taking linear combinations of the scaling function and its
translates we can generate a larger function.

f (t ) ak (t k ) ...(3.17)

67

Where a and k are real numbers (scales) which we call as coefficients of (t k ) s

The scaling function has the property that any function that can be represented by the
scaling function can also be represented by the dilated versions of the scaling functions
For example
One of the simple scaling functions is the haar scaling function
t1
(t) 10,,0otherwise
...(3.18)

Then f(t) can be any piecewise continuous function that is constant in the interval [k,k+1]
for all k
Let us define (t ) (t k ) the set of all function that can be obtained using linear
combination of the set {k (t )}
f (t ) akk (t ) ..(3.19)
k

As the scaling function has the property that any function that can be represented exactly
by an expansion at some resolution j can also be represented by dilations of the scaling
function at resolution j+1. In particular, this means that the scaling function itself can be
represented by its dilations at higher resolutions.
The final equations of MRA is (t ) hk 2 (2t k ) ..(3.20)
Now this equation is useful for defining the wavelet transform, before going to see about
the wavelet transform we begin with What is scaling. The parameter scale is used in
wavelet analysis, as we use it to scale the map. In case of map scale, high scale
corresponds to a global view (i.e. non-detailed) and low scales corresponds a detailed
view. Similarly in terms of frequency Low frequencies correspond to global information
of signal (high scales). Whereas high frequencies correspond to the detailed information
of a hidden pattern in the signal (low scales). Scaling is a mathematical operations either
it dilates or compresses a signal. Larger scales correspond to dilated signals and small
scales correspond to compressed signals. In terms of mathematical functions if f(t) is a
given function f(st) (where s is scaling) corresponds to a contracted (compressed) version
of f(t) if s<1 and to an expanded (dilated) version of f(t) if s>1.[32][33]

68

3.4 Wavelet Transform (WT):

Wavelet means small wave. Wavelet analysis is about analyzing signal with short
durations finite energy function. They transform the signal under investigation into
another representation, which presents signal in more useful form. This transformation of
the signal is called the Wavelet transform[44],[45][46].

3.4.1 The Continuous Wavelet Transform (CWT)

The Continuous Wavelet Transform (CWT) is similar to the STFT because it also
involves the multiplication of the signal with some function (in this case, the wavelet),
and the transform is computed separately for different segments of the signal. The
difference between the CWT and the STFT is that the width of the window is changed as
the transform is computed for each spectral component and it can be written as.

( s, ) f (t ) s, (t )dt ..... (3.21)

*

Where * denotes the complex conjugation. The Equation (3.21) shows how function f (t)
is decomposed (sampled) into a set of basis functions s, (t ) called the wavelets. The
variables s and are the two new dimensions scale and translations. And the inverse
wavelet transform is given by
f (t ) ( s, )s , (t ) dds .. (3.22)

The wavelets are generated from a single basis wavelet (t ) , so called the mother
wavelet, by scaling transform

s , 1 / s * (t / s ) ..(3.23)
Where s is the scale factor and is translation factor and the factor

s is for energy

normalization across the different scales. For the function to be wavelet it should be
limited. For a given scaling parameter s we translate the wavelet by varying the
parameter . We define the wavelet transform as:
w( s, ) f (t ) *1 / s * (t / s )dt ..(3.24)
t

69

According to Equation (3.24) for every (s, ), we have wavelet transform coefficient
representing how much the scaled wavelet is similar to the function at location t= /s.
However, the Continuous wavelet Transform (CWT) still has three properties, which are
difficult to use:
The CWT (Equation 3.21) is calculated continuously shifting a continuously
scalable function over a signal and calculating the correlation between them and it
will be near an orthogonal basis and obtained wavelets coefficients will therefore
be highly redundant.
Even without the redundancy of the CWT we still have an infinite number of
wavelets in the wavelet transform.
Most of the functions of wavelet transforms have no analytical solutions and they
can be calculated only numerically by an optical analog computer [7][17]-[20].
The CWT in this form is not suitable for computer implementation. So, fast algorithms
are needed to be able to exploit the power of the wavelet transform and it is fact that the
existence of these fast algorithms that have put wavelet transforms where they are today
[7]. As mentioned in CWT it has two parameters (s and ) therefore contains high
amount redundant data. Instead of varying the parameters, this work has analyzed the
signal with a small number of scales with varying number of translations at each scale.
This is called the Discrete Wavelet Transform (DWT). This is achieved by modifying the
CWT (Continuous Wavelet Transform) through sampling specific wavelets coefficients

3.4.2 The Discrete Wavelet Transform (DWT)

DWT is obtained via s 2 j where j and k are integers representing the set of discrete
translations and discrete dilations.

j , k (t ) 1 / s 0 j . (t k 0 s 0 j / s 0 j
i.e.

j , k (t ) s0

j/2

( s0 j t ko) (3.25)

70

That which is the function of j and k, we denote it as ( j , k ) , which defines the resolution
of the DWT in both time and frequency. If j & k are integers and s 0 >1 is fixed then it is
dilation step. The translation factor 0 depends on the dilation step. The effect of
discretising the wavelet is that the time scale is now sampled at discrete intervals. We
usually choose s 0 =2 so that the sampling of frequency axis corresponds to dyadic
sampling. For the translation usually we choose 0 =1. So, that we also have dyadic
sampling of this time axis. DWT results the series of Wavelet coefficients and it is
referred as Wavelet Decomposition.
Thus redundancy can be removed by making the discrete wavelets orthogonal and by
making the Discrete Wavelets orthonormal. Dilation and translation made orthogonal to
their own dilations and translations i.e.

j, k

(t ) * j , k (t ) dt

1
0

if j=m & k=n, 0 other wise.(3.26)

We can get the inverse wavelet transform for discrete wavelets by summing the
orthogonal wavelet basis function, weighted by the wavelet transform coefficients
f (t ) ( j , k ) j , k (t ) ..(3.27)
j ,k

Every time wavelets do not need the orthogonality because in some application the
redundancy can help to reduce the sensitivity to noise; this gives the disadvantage of
discrete wavelet (DWT). [17]-[20]

3.4.3 Structure of Wavelet

Wavelet transform is a pair of filters. The way in which we compute the wavelet
transform by recursively averaging and differentiating coefficients is called the filter
bank [34],[35] where one is a low pass filter (lpf) and the other one is a high pass filter
(hpf). Each of the filters is downsampled by two. Each of those two output signals can be
further transformed. Similarly, this process can be repeated recursively several times,
resulting in a tree structure called the decomposition tree. Figure shows the Filter bank
structure for one level decomposition and reconstruction of two dimensional signals.
Wavelet transform can be used to analyze or decompose signals and an image operation
through wavelet is called decomposition [29][30][31]. The same components can be
assembled back into the original signal without loss of information called reconstruction

71

or synthesis Figure 3.4.3(a) Shows Wavelet Decomposition at Level 1, Figure 3.4.3(b)

shows the reconstruction and Figure 3.4.3(c) shows way in which decomposition of the
image taking place in four directions. Figure 3.4.3(d)-3.4.3(h) shows wavelet
decomposition from 1to 5 level for cameraman image.

Figure3.4.3 Shows Wavelet Decomposition at Level

Figure- 3.4.3(b )shows the Wavelet reconstruction at Level 1

Figure 3.4.3(c) Representation of Image Decomposition at level 1.

72

Figure-3.4.3(d) Wavelet Decomposition at Level 1

Figure-3.4.3(e) Wavelet Decomposition at Level 2

Figure-3.4.3(f) Wavelet Decomposition at Level 3

73

Figure-3.4.3(g) Wavelet Decomposition at Level 4

Figure-3.4.3(h) Wavelet Decomposition at Level 5

The structure of Wavelet can be represented as a four channel perfect reconstruction of
filter bank. Now each filter is 2D with subscript indicating the type of filter (HPF or LPF)
for separation of horizontal and vertical components [36]-[40]. The resulting fourtransform components consist of all possible combinations of high and low pass filtering
in the two directions. By using these filters in one stage an image can be decomposed into
four bands as shown in Figure-3.4.3(c) There are three types of detail of images for each
74

resolution Diagonal (HH), Vertical (LH) and Horizontal (HL). The operations can be
repeated on the low low (LL) i.e. on approximation band using the second identical filters
bank [41]. The decomposition process can be iterated, with successive approximations
being decomposed. However, in practice, more than one decomposition level is
performed on the image data. Successive iterations are performed on the low pass
coefficients (approximation) from the previous stage to reduce further the number of low
pass coefficients. Since the low pass coefficients contain most of the original signal
energy, this iteration process yields better energy compaction. The quality of compressed
image depends on the number of decomposition. Compression of an image can be
obtained by ignoring all coefficients less than the threshold value. If we use
decomposition iteration, it will be more successful in resolving DWT coefficient because
Human Visual System (HVS) is less sensitive to removal of smaller details.
Decomposition iterations depend on the filter order. Higher order does not imply better
image quality because of the length of the wavelet filter; this becomes a limiting factor
for decomposition. Usually, Five levels of decompositions are used in current wavelet
based image compression [8],[42]-[46]. The maximum levels of Decomposition of any
image can be determined by using the formula given in equation (3.28)
Maximum Levels of Decomposition = log2 xmax ....(3.28)
Where xmax is the maximum size of given image

3.4.4 Wavelet properties:

To achieve a high compression rate, it is often necessary to choose the best wavelet filter
bank and decomposition level, which will play a crucial role in compressing the images.
The selection of wavelet filters plays a crucial part in achieving an effective coding

75

performance, because there is no filter that performs the best for all images [47]. The
existing Compression system used the biorthogonal wavelet filters instead of orthogonal.
Because orthogonal filters have a property of energy preservation whereas, biorthogonal
filters lack it [48][49]. The choice of optimal wavelets has several criteria. The main
criteria are:
1.

Orthonormality

2.

Filter Length

3.

Vanishing order or moment

4.

Smoothness

5.

Filter magnitude response

6.

Group delay

7.

Decomposition level

8.

Regularity

These are discussed in [29],[33],[48]-[56]. Orthogonal Filters lead to orthogonal wavelet

basis functions; therefore the resulting wavelet transform is energy preserving; this
implies the Mean Square Error (MSE) introduced during the quantization of the DWT
coefficients which is equal to the MSE in the reconstructed signal. In biorthogonal
wavelet the basis functions are not orthogonal; therefore they do not preserve the energy
but conserve it. The efficiency of a transform depends on how much energy compaction
is provided by the transform. Wavelet Filter can be used to analyze or decompose signals
and images called decomposition. The same components can be assembled back into the
original signal without loss of information, which is called reconstruction or synthesis.
Shorter synthesis basis functions are desirable for minimizing distortion that affects the
subjective quality of the image. Longer filters are responsible for ringing noise in the
reconstructed image at low bit rates [50]. Each wavelet family is parameterized by an
integer N called the filter order, which is proportional to the length of the filter. The
length of the filter is related to the degree of the smoothness of the wavelet and which can
affect the coding performance. This relation is different for different wavelet families
[57] and non-smoothness basis function introduces artificial discontinuities that are
reflected as spurious artifacts in the reconstructed images. Higher filter order gives more
energy and increases the complexity of calculating the DWT coefficients, while lower

76

order preserves the energy. i.e. it preserves the important edge information. Therefore, we
must take care of wavelets in image compression application concerning that good
balance between filter orders, degree of smoothness and, computational complexity [58].
These properties depend on the image contents. Vanishing order is the measure of
compaction property of the wavelet and it corresponds to the number of zeros in the LL
sub band [50][59]. Filter Response is another critical property that affects the subjective
quality of the reconstructed image. The filter responses approach the ideal rectangular
response with the increase in the number of zeros and these numbers of zeros also
correspond to vanishing order of the wavelet. Group Delay Difference (GDD) measures
the deviation in-group delay of the orthogonal wavelets from the linear phase group
delay. A non-GDD introduces a phase distortion that affects encoding and decoding by
altering the DWT Subband structure. The decomposition process can be iterated with
successive approximations being decomposed. However, in practice more than one
decomposition level is performed on the image data. Successive iterations are performed
on approximation coefficients; this successive iterations process yields better energy
compaction. The quality of the compressed image depends on the number of
decompositions, and these decomposition iterations depend on the filter order. Higher
order does not imply better image quality because of the length of wavelet filter. This
becomes a limiting factor for decomposition. Usually five levels of decomposition are
used in wavelet-based compression [8][50][60][61]. Regularity is one of the properties of
the wavelet; greater regularity often does not improve the visual quality.

77

Table indicating the properties of wavelet

3.5 Wavelet Packet (WP)

In addition to the above properties of wavelet transform, wavelet packets provide more
flexible decomposition at any node of the decomposition by allowing decomposition at
any node of the decomposition tree and also to obtain the best decomposition tree [62].
The benefit of the wavelet packets over the wavelet decomposition comes from the
ability of the wavelet packets to better represent high frequency (this is how image may
contains a noise) content, and high frequency oscillating signals in particular. This allows
wavelet packet to perform significantly better than wavelets for compression of images
with large amount of texture (like fingerprint images, Barbara images etc) and also points
out that the perceived image quality is significantly improved using wavelet packets
instead of wavelets especially in the textured regions of the images [63][64]. The major
difference between wavelet and wavelet packet is: The filter design associated with the
wavelet analysis method involves iterating the low-pass and high -pass filtering and
downsampling procedure only on the low-pass branch of the previous. While, wavelet
packet (WP) provides an extension of the octave band wavelet decomposition to full tree
decomposition i.e. the high pass output of each branch is also filtered and downsampled

78

up to maximum number of decomposition this is one of the key differences between the
wavelet and wavelet packet [65][66].
Figure 3.5 shows the Wavelet Packet (WP) presenting on extension of the octave band
wavelet decomposition to full tree decomposition

Figure 3.5(a) shows the Wavelet decomposition to full tree decomposition

Figure 3.5(b) Wavelet packet decomposition at level 1 (DB1) of Woman image

Figure 3.5(c) Wavelet packet decomposition at level 2 (DB1) of Woman image

79

3.6 Multiwavelets (MW)

3.6.1 Motivation behind Multiwavelets for Compression
Algorithms based on wavelets have been shown to work well in image compression. For
better performance in image compression, wavelets transforms require filters which can
combine some of desirable properties such as orthogonality, Symmetry. However, the
design possibilities for wavelets are limited because they cannot simultaneously posses
all the desirable properties. In contrast to the limitations of wavelets, Multiwavelets are
able to posses the best for all these properties simultaneously due to the extra freedom in
the design of multi-filters. Another desirable feature of any transform used in image
compression is the amount of energy compaction achieved. A filter with good energy
compaction property can decorrelate a fairly uniform input signal into a small number of
scaling coefficients containing most of the energy and large number of sparse wavelet
coefficients. This becomes important factor during quantization. Therefore better
performance is obtained when the wavelet coefficients have values clustered about the
zero with little variance to avoid as much as quantization noise as possible. Thus
Multiwavelets have the potential to offer better reconstructive quality of the same bit rate
and can achieve better level of performance than the wavelets with a similar
computational complexity [67]-[72]
Multiwavelets are similar to the wavelets but have some important difference i.e.
Multiwavelets are characterized with several scaling functions and associated wavelet
functions [73],[74].
Let the scaling functions be denoted in vector form as (t) =[1(t), 2(t), , L(t)]T,
where (t) is called the multi-scaling function, T denotes the vector transpose and j(t) is
the jth scaling function. Likewise, let the wavelets be denoted as
(t) = [1(t), 2(t), . . . , L(t)]T, where j(t) is the jth wavelet function. Then, the
dilation and wavelet equations for Multiwavelets take the following forms, respectively:

(t ) 2 H k ( 2t k )
k

(t ) 2 Gk (2t k )
k

. (3.29)

80

The low-pass filter H and the high-pass filters G are N N matrix filters, instead of
scalars. In theory, N could be as large as possible, but in practice it is usually chosen to
be two. This degree freedom can be used to incorporate useful properties of the
Multiwavelet such as orthogonality, symmetry, and high order approximation [68]. For
single scaling and wavelet functions extend to the matrix version. The resulting two
channels, 2 2 matrix filter bank operates on two input data streams, filtering them into
four output streams. Each of these streams is then downsampled by a factor of two. This
procedure is as shown in Figure-3.6.1(a)

Figure-3.6.1(a) Multiwavelets Filters with repeated input

3.6.2. Decomposition of Multiwavelets.
In scalar wavelet, during the single level of decomposition the image data is replaced
with four blocks corresponding to the sub bands representing either low-pass or high-pass
filtering in each direction (Horizontal, Vertical, Diagonal and Approximation).
Multiwavelets have two channels, so that there will be two sets of scaling and two sets of
wavelets coefficients. 2-D Multiwavelet decomposition has the 16 subband intermediate
image (single level decomposition) as shown in figure-3.6.2(a).
L1L1

L1L2

L1H1

L1H2

L2L1

L2L2

L2H1

L2H2

H1L1

H1L2

H1H1

H1H2

H2L1

H2L2

H2H1

H2H2

Figure-3.6.2(a): Single Level Multiwavelets Decomposition

81

As Scalar wavelet transforms give a single quarter sized low-pass sub band from the
original larger sub band. In Multiwavelets, the multilevel decompositions are performed
in the same. Multiwavelet iterates on the low-pass coefficients from the previous
decomposition; the quarter image of Low-pass coefficients is actually a 2x2 sub bands.
L1L1

L1L2

L2L1

L2L2

The next step decomposes the Low-Low-Pass sub matrix in the similar manner which
gives the Second level of Decomposition and as shown in Figure-3.6.2(b). An N number
of Decomposition of a 2D image will produce the 4(3N+1) subband (sub
images)[73][74].
L1L1

L1L2

L1H1

L1H2

L2L1

L2L2

L2H1

L2H2

H1L1

H1L2

H1H1

H1H2

H2L1

H2L2

H2H1

H2H2

H1L1
H2L1

H1L2
H2L2

L1H1

L1H2

L2H1

L2H2

H1H1

H1H2

H2H1

H2H2

Figure-3.6.2(b): Second Level Multiwavelets Decomposition

Figure 3.6.2(c): Multiwavelets Decomposition of Cameraman image at Level 1

82

Figure 3.6.2(d). Multiwavelets Decomposition of Cameraman image at Level 2

3.7 Stationary Wavelet Transform (SWT)

Wavelet transform is superior approach time-frequency analysis tools because its time
scale width of the window can be stretched to match the original signal, especially in
image processing analysis. Particularly it has a better performance when it is applied to
non-stationary signal analysis such as noise removal and transient detection. Discrete
wavelet transform (DWT) provides the useful information for texture analysis in the
fingerprint image.

DWT is Multiresolution analysis, which is a non-redundant

decomposition due to non-invariance in time. The Stationary Wavelet Transform was

proposed to make the decomposition time invariant [75] [76].

In order to preserve the invariance by translation, the down sampling operation must be
suppressed and the decomposition obtained in redundant form, which is to be referred as
Stationary Wavelet Transform, depicted in figure 3.7(a). The redundancy of this
transform facilitates the identification of salient features in a signal, especially for
recognizing the noises. This approach gives good result for images but not for signal
analysis.

83

DWT

Wavelet Coefficient

DWT

Translation

Wavelet Coefficient

Input Image

Translation

Figure 3.7: Stationary Wavelet Transform

SWT has similar tree structure implementation without any sub-sampling. This balance
of Perfect Reconstruction (PR) is preserved through level dependent zero padding
interpolation of respective low pass and high pass filters in the filter bank structure. In
fingerprint image decomposition, separate the variables x, y of fingerprint image and
have the following wavelets [77].
Vertical wavelet (LH):

1 ( x , y ) ( x ) ( y ) ..(3.30)

Horizontal Wavelet (HL): 2 ( x , y ) ( x ) ( y ) ..(3.31)

Diagonal Wavelet (HH): 3 ( x , y ) ( x ) ( y ) ...(3.32)
Where, is the wavelet function and is the scaling function. The detailed signals
contained in the three sub-images as follows:

w1j 1 ( k x , k y )

g (lx)h(ly)c

lx ly

2
j 1

w (k x , k y )

j ,k 2 j

(lx, ly ) .......(3.33)

h(lx)g(ly)c

j ,k 2 j

(lx, ly) . (3.34)

lx ly

w 3j 1 ( k x , k y )

g (lx ) g (ly )c

lx ly

j ,k 2 j

(lx , ly ) ..(3.35)

84

3.8 Summary:
This Chapter describes the various multi-resolution techniques used in transform coding
by providing mathematical background for the transforms followed by the limitations of
time-frequency limitations of the STFT. It emphasis the theory and practical
implementation of wavelets i.e. Structure, properties required for the compression and
decompression of the image. This chapter also describes the limitations of wavelets and
how these are overcome through wavelet packet and Multiwavelet while discussing the
additional properties of the wavelet. Finally describes the stationary wavelet transform
(SWT). The experiments are carried out through these multi-resolution techniques
compression and de-nosing pertaining to the various applications in presence/absence of
noise discussed in Chapter 4.

85

References:
01

Raghuveer M Rao, Ajit S Bopardikar, Wavelet Transforms : Introduction to

Theory and Applications, LPE Pearson Education, 2005

02

R.C. Gonzalez, R.E. Woods and Steven L Eddins, Digital Image Processing
using MATLAB, Second Edition, Pearson Education, 2004

03

R.C. Gonzalez, R.E. Woods, Digital Image Processing, Second Edition,

Pearson Education, 2004.

04

David Salomon, Data Compression the Complete Reference, 2nd Ed. Springer

05

Khalid Sayood Introduction to Data Compression", Second Edition, Morgan

Kaufman Publisher, 2003.

06

Subhasis Saha, 21 Jan 2001, Image Compression from DCT to Wavelet: A

review, http://www.acm.org/crossroads/xrds6-3/sahaimgcoding.html

07

K.P.Soman, K.Ramchandran "Insight into Wavelets from Theory to Practice",

PHI New Delhi-2004

08

Michael B. Martin Applications of Multiwavelets to image Compression,

M.S., Thesis, June 1999.

09

Yun Q. Shi and Huifang Sun. Image and Video Compression for Multimedia
Engineering: Fundamentals, Algorithms, and Standards. CRC Press LLC, Boca
Raton FL, first edition, 2000.

10

M. B. Martin and A. E. Bell. New image compression techniques using

Multiwavelets and Multiwavelet packets. to appear in IEEE Transactions on
Image Processing, March 2001

11

M. Saenz, R. Oktem, K. Egiazarian, , and E. Delp, Color image wavelet

compression using vector morphology. Proc. of the European Signal Processing
Conference, September 5-8 2000

12

V. Strela. Multiwavelets: Theory and Applications. PhD thesis, Massachusetts

Institute of Technology, 1996

13

M. Vetterli and J. Kovacevic. Wavelets and Subband Coding. Prentice Hall,

Englewood Cliffs NJ, 1995

14

J. Y. Tham, L. Shen, S. L. Lee, and H. H. Tan. A general approach for analysis

and application of Discrete Multiwavelet transforms. IEEE Trans. on Signal
Processing, 48(2):457464, February 2000.

86

15

Chaur-Chin Chen, On the Selection of Image Compression Algorithms, IEEE,

14th International Conference on Pattern Recognition (ICPR'98) - Volume 2 p.
1500, 1998.

16

Wavelet
Tutorial
Part
I,
II&III
by
Robi
http://users.rowan.edu/~polikar/WAVELETS/Wttutorial.html

17

Simon Haykin & Barry Van Veen Signals and Systems Library of congress
catalog in publication Data 1999

18

C.Valens,
"A
really
C.Valens@mindless.com

19

C.Valens, "A really

C.Valens@mindless.com

20

Naoki Saito, Frequently Asked questions on Wavelets, University of

California,
USA, April 4-2004.

21

Martin J. Mohlenkamp, "A Tutorial on Wavelets Thesis and Applications",

University
of Colorado at Boulder, Dept, of Applied Mathematics,
MJM@colorado.edu

22

I. Daubechies Book review Ten Lecture on Wavelets, IEEE Geosciences &

Remote Sensing Society Newsletter 1992

23

David Bethel, Optimization of still Image Compression Techniques, Ph. D.

Thesis, University of Bath, 1997

24

Panchamkumar D Shukla, Complex wavelet Transform and their applications,

Ph. D. Thesis, University Strathclyde, UK, 2003.

25

Yinpeng Jin, Elsa Angelini and Andrew Laine, Wavelets in Medical Image
Processing: De-nosing, Segmentation and Registration, in Handbook of Medical

Friendly

Friendly

Guide

Guide

to

to

Polikar,

Wavelets",

Wavelets",

1999,

1999-2004,

Image Analysis, Segmentation and Registration Models, Vol. I.

26

S. G. Mallat, Theory of Multiresolution signal Decomposition: The Wavelet

Representation, IEEE Trans. On Pattern Analysis and Machine Intelligence
Vol.11.
No. 7 July-1989.

27

Andrew P began, Wavelet based image compression using human visual

System Models, MS thesis, VPI and State University, Virginia, 2001

28

Amar Graps, An Introduction to wavelets, IEEE computational Science and

Engineering. Vol 2. Number 2. 1995.

29

Wavelet Toolbox users Guide, www.mathworks.com

87

30

Ali Reza, From Fourier Transform to Wavelet Transform basic concepts,

spire Lab,
UWM oct-27-1999

31

T.K.Sarkar,C.Su, R Adve, M. Salazar-Palma L Carcia-Castilla, Rafeal R. Boix,

"A
Tutorial on wavelets from an Electrical Engineering Perspective Part-1
Discrete Wavelet
Techniques", IEEE Antenna and Propagation Magazine
Vol.40. No.5 October-1998

32

I Daubechies, Ten lectures on wavelet, SIAM, Philadelphia PA, 1st edition,

1992

33

S. Mallat, A Wavelet Tour of Signal Processing, Academic Press, 1999

34

Geeta S. Rao, "Wavelet Analysis and Applications" New Age International

Publishers, 2004

35

M Vetterli and C Merely, Wavelets and Filter Banks: Theory and Design,
IEEE Transaction on Signal Processing Vol. 40 PP 2207-2232

36

J. Kovacevic and M. Vetterli, Non separable Multidimensional perfect

reconstruction filter banks and wavelet bases for Rn, IEEE Trans. on
Information Theory, vol. 38(2), pp. 533535, Mar. 1992

37

J. Kovacevic andW. Sweldens, Wavelet families of increasing order in arbitrary

dimensions, IEEE Trans. on Image Processing, 1999

38

E. Simoncelli and E. Adelson, Non-separable extensions of quadrature mirror

filters to multiple dimensions, Proceedings of the IEEE, vol. 78, pp. 652664,
Apr. 1990

39

M. Antonini, M. Barlaud, P. Mathieu, and I. Daubechies. Image coding using

wavelet transform. IEEE Transactions on Image Processing, 1(2):205220,
1992

40

Image Processing Toolbox for use with Matlab. The Mathworks Inc., Natick
MA, second edition, 1997

41

Sonja Grigic, Mislav Grigic, Branko Zovko-Cihlar, Performance Analysis of

Image
Compression using Wavelets", IEEE Transactions on Industrial
electronics PP.682- 695, Vol.48. N0. 3, June 2001

88

42

Sonja Grigic, Mislav Grigic, Bronka zovko Optimal Decomposition for

Wavelet
Image Compression", First International Workshop on Image and
Signal Processing And Analysis, June 14-15, 2000, Pula, Coatia

43

S. Mallat, A theory for Multiresolution signal decomposition: The wavelet

representation, IEEE Trans. Pattern Anal. Machine Intelligence ., vol. 11, pp.
674 693, July 1989.

44

A. N. Skodras, C. A. Christopoulos, and T. Ebrahimi. Jpeg2000: The upcoming

still image compression standard. Proc. of the 11th Portugese Conference on
Pattern Recognition, pages 359366, May 2000

45

Imgeun Lee and Jongsik Kim et al. Wavelet transform image coding using
human visual system. IEEE Asia-Pacific Conference on Circuits and Systems,
pages 619623, 1994

46

M. Miloslavski and Yo-Sung Ho. Zero tree wavelet image coding based on the
human visual system model. IEEE Asia-Pacific Conference on Circuits and
Systems, pages 5760, 1998

47

Subhasis Saha & Rao Vemuri Analysis based adaptive Wavelet filter selection
in
Lossy image coding schemes, ISCAS-2000-IEEE International
Symposium on
Circuits and systems May , Geneva, Switzerland, 2000

48

Bryan E. Usevitch A Tutorial on Modern Lossy Wavelet Image Compression:

Foundations of JPEG 2000, IEEE Signal Processing Magazine, 2001.

49

John D. Villasenar, Benjamin Belzer and Judy Liao, Wavelet Filter Evaluation
for image compression, IEEE Transactions on Image Processing, Vol. 4. No. 8,
1995.

50

Satyabrata Rout Orthogonal Vs Biorthogonal Wavelets for Image

Compression, MS-Thesis, Virgina Polytechnic Institute and State University,
Virgina, August, 2003.

51

John D Villasenor, Benjamin Belzer, and Judy Liao, Wavelet filter evaluation
for image compression, IEEE Trans. On Image Processing, Vol. 4(8) 10531060, 1995.

52

Michael Lightstone, Eric Majani and Sanjit K. Mitra, Low bit-rate design
considerations for wavelet based image coding, Multidimensional system and
signal processing, Vol. 8(1-2) PP:11-128, 1997.

89

53

J. E. Odegard and C. S. Burrus, Discrete finite variation: a new measure of

smoothness for the design of wavelet basis, Proc. Of Int. Conference on
Acoustic, speech and signal Processing, Vol. 3, PP: 1467-1470, 1996.

54

Olivier Rioul, Regular Wavelets: a discrete-time approach, IEEE Tran. On

Signal Processing, Vol. 41(12) PP 3572-3579, 1993.

55

Gillbert Strang and Truong Nguyen, Wavelets and Filter Banks WellesleyCambridge Press, Wellesley MA, 1st Edition, 1996.
Michael Unser and Thierry Blu, Mathematical Properties of the JPEG2000
wavelet filters, IEEE Tran. On Image Processing, 2003

56

57

Yinfen Low and Rosli Besar Wavelet based Medical Image Compression using
EZW, Proceedings of 4th National Conference on Telecommunication
Technology, Shah Alam, Malaysia IEEE, PP. No 203-206

58

Timo Eirola, Sobolev characterization of solutions of dilation equations

SIAM, J. Math Anal, Vol 23 PP 1015-1030, 1992

59

Zhuhan Jiang Xiling Guo, Wavelet of Vanishing moments and minimal filter
and the Application to image Processing IEEE, 2001.

60

Kwo-Jyr Wong et. al. Image Compression with Fully Decomposed Wavelet
Transform, University of Sothern California, Los Angeles, CA IEEE, 1992.

61

M. Babel, O. Deforges, Lossless and Lossy minimal redundancy pyramidal

decomposition for scalable image compression technique IEEE Proc. of
International Conference on Multimedia and Expo - Volume 3 (ICME '03),
2003.

62

Karen Lees, Image Compression Using Wavelets, Report of MS, (2002)

63

Deepti Gupta, Shital Mutha Image Compression Using Wavelet Packet,

IEEE (2003).

64

Michael B. Martin and A.E. Bell Multiwavelet packet Image compression:

Theory And results IEEE, DSP Workshop Oct 15th 18th, (2000)

65

Zixing, Xiong,Kannan Ramchandran, Michel T Orchad Wavelet Packet Image

Coding Space Frequency Quantization, IEEE Transactions on Image
Processing, (2001) Vol 7, No.6.

66

Wavelet Toolbox users Guide, www.mathworks.com

90

67

Michael B. Martin and A.E. Bell, New Image Compression Techniques Using
Multiwavelets and Multiwavelets Packets, IEEE Transactions on Image
Processing Vol. 10. No. 4. PP 500-510, 2001.

68

Michael B. Martin and A.E. Bell, Multiwavelets Packet Image Compression:

Theory and Results, IEEE DSP Workshop October 15th-18th, 2000.

69

Mariantonia Contronie et. al., Image Compression through Embedded

Multiwavelet Transform Coding, IEEE Transactions on Image Processing
Vol.9.No.2. Feb-2000.
Kai Bao and Xiang Genxia, Image Compression Using a New Discrete
Multiwavelet Transform and a new embedded Vector Quantization , IEEE
Transactions on Circuits and Systems for Video Technology Vol.10. No. 6.
September, 2000.

70

71

L.R.Iyer and A.E. Bell, Improving Image Compression Performance with

Balanced Multiwavelets, IEEE Transactions 2001.

72

Sudhakar R. and Jayaram S., Fingerprint Compression using Multiwavelets,

International Journal of Signal Processing Vol.2. No. 2, 2005.

73

Gerlind Plonka and Vasily Sterla, From Wavelets to Multiwavelets, Morten

Daehlen, Tom Lycle, Larry L Schumakar, PP :375-399, 1998 Vanderbilt
University Press, Nashville, TN.

74

V. Sterela , P. N Heller, P. Topiwala and C-hall, The Applications of

Multiwavelets Filter banks to image Processing, IEEE Trans Image Processing
Vol.8.April, 1999.

75 J. C. Pesquet, H. Krim, and H. Carfantan, "Time-invariant orthonormal wavelet

representations," IEEE Trans. Signal Processing, vol. 44, pp. 19641970, Aug.
1996
76

Warin Chumsamrong and et al, "Using Stationary Wavelet Transform in the

classification of SAR Images", 19th Asian Conf. on Remote Sensing, 16-20
1998

77 X. H. Wang, Robert S. H. Istepanian, and Yong Hua Song, "Microarray Image

Enhancement by Denoising Using Stationary Wavelet Transform" IEEE Tran. on
Nanobioscience, Vol. 2, No. 4, 2003

91