6CS3-01/6IT3-01: Digital Image Processing

UNIT 2
Image Transformation & Filtering

Presented by: Ms. Yashika Saini, Assistant Professor

Department of Computer Science & Engineering
AIET, Jaipur

1
TOPICS COVERED:
Intensity transform functions
Histogram processing
Spatial filtering
Fourier transforms and its properties
Frequency domain filters
Colour models
Pseudo colouring
Colour transforms
Basics of Wavelet Transforms

2
 Introduction
• To enhance a particular image we have to go
through image processing operations.

• There are principally two image processing

domains i.e. Spatial domain & Frequency
domain.

3
 Spatial domain
• The term spatial domain refers to the image plane
itself, and
• Image processing in this domain are based on
direct manipulation of pixels in an image.

4
 Frequency domain
• In this, first we transform an image into the
transform domain (Fourier transform), doing the
processing there, and obtaining the inverse
transform to bring the results back into the spatial
domain.

5
 Categories of spatial domain
Two principle categories of spatial processing are:
1. Intensity transformations &
2. Spatial filtering.
• Intensity transformations operate on single pixels
of an image, principally for the purpose of
contrast manipulation and image thresholding.
• Spatial filtering deals with performing operations,
such as image sharpening, by working in a
neighborhood of every pixel in an image.
6
7
 Intensity transformation
• It is represented by an expression
S= T(r)
Where, r = input intensity of the pixel before
transformation.
S= output intensity of the pixel after transformation.
T is transform operator.
Applications:
1. Contrast stretching and
2. Image thresholding.

8
 1. Contrast stretching
• Contrast stretching is used to enhance the quality of image.
• If the value of r<k, it will be mapped to narrow range of
intensity levels towards the dark region.
• If the value of r>k, it will be mapped to narrow range of values
towards the brighter region.

9
 2. Image thresholding
• Thresholding is based on the value of k.
• If r<k, it is assigned with black intensity level.
• If r>k, it is assigned with white intensity level.
Thus, produces binary image.

10
 Spatial Filtering
General expression for spatial filtering is given by,
g(x, y) = T[ f(x, y)]
Where, f(x, y) is input image
g(x, y) is output image
T is transform operator

11
 Spatial Filtering
• This neighborhood is called as spatial filter or
kernel or template or mask or window.
• If the origin of the neighborhood is at the border
of the image, part of the neighborhood will reside
outside the image will be ignored or to pad the
image with border of 0’s.

12
 Spatial Operations
• These are the functions that create new spatial
data from specified input data.
• Spatial operation is performed on the pixels of a
given image.
There are 3 broad categories of spatial operations:

13
 Single- pixel operations
• Alter the values of an image pixels based on the
intensity.
• Example : NOT operation

14
 Single- pixel operations
• Single point operation on gray scale images.

15
 Neighborhood operations
• Let Sxy denote the set of coordinates of a neighborhood
centered on an arbitrary point (x, y) in an image, f.
• Neighborhood processing generates a corresponding pixel at
the same coordinates in an output image, g,
• The value of this pixel is determined by a specified operation
involving the pixels in the input image with coordinates in
Sxy.

16
 Neighborhood operations
• Neighbor
• Padding

17
Neighborhood operations

18
Neighborhood operations

19
 Geometric spatial transformations
• A spatial transformation of an image is a geometric
transformation of the image coordinate system.
• In a spatial transformation each point (x, y) of image
is mapped to a point (v, w) in a new coordinate
system.
• The transformation of coordinates may be expressed
as
• Where (v, w) are pixel coordinates in the original
image and (x, y) are the corresponding pixel
coordinates in the transformed image.

20
 Geometric spatial transformations
• General form of spatial coordinate transformation
given by:

• This transformation can scale, rotate, translate or

sheer a set of coordinate points, depending on the
value chosen for the elements of matrix T.

21
Geometric spatial transformations

22
Intensity Transformation Functions
There are three types of transformation functions:

1. Linear (negative & identity) transformation.

2. Logarithmic (log & inverse log) transformation.

3. Power-Law(nth power & nth root) transformation.

23
 Linear Transformation
1.1 Image Negatives- The negative of an image with intensity levels
in the range [0, L-1] is obtained by using the negative
transformation. It is given by, S = L-1-r.
(If r =0, S = L-1 & If r = L-1, S=0)

24
 Linear Transformation
1.2 Identity transformation-
• The intensity level of the image after transmission is identical to
that of the original image.

• Each value of the input image is directly mapped to each other

value of output image.

25
 Log Transformations
2. Log Transformations- The general form of log transformation
is
S = c log (1+r)
Where, c is a constant, r ≥ 0.

• The shape of the log curve in Fig. 3.3 shows this

transformation maps a narrow range of low intensity values in
the input into a wider range of output levels.
• By using log transformation, we can compress or expand the
gray level.
• Log transformation produces high contrast images and Inverse
log will produce the low contrast image.

26
• We use this transformation to expand the values of dark
pixels in an image while compressing the higher level
values.

27
 Power-Law Transformations
3. Power-Law (Gamma) Transformations: Power law
transformation is given by,
Where C and γ are positive constants.
• We use gamma transformation where we need to expand or
compress darker region.

• If γ=1, No change in the image, it will be identity image.

• If γ>1, It provides compression of the brighter region.

• If γ<1, It provides expansion of the darker region.

28
Power-Law Transformations

29
Application of Gamma Transformations

30
Application of Gamma Transformations
(Expansion of darker intensity level)

31
Application of Gamma Transformations
(Compression of brighter intensity level)

32
Piecewise-Linear Transformation Function
Advantage: It is less complex, compared to other
transformation function.
Disadvantage: It requires more user inputs.
Three Piecewise transformation functions are:
1. Contrast stretching
2. Intensity level slicing
3. Bit-plane slicing.

33
 Contrast Stretching
• Low-contrast images can result from poor
illumination, lack of dynamic range in the
imaging sensor, or even the wrong setting of a
lens aperture during image acquisition.
• Contrast stretching is a process that expands
the range of intensity levels in an image so that
it spans the full intensity range of the recording
medium or display device.

34
 Contrast Stretching
• The locations of points (r1, s1) and (r2, s2) control the shape of the transformation
function.
• r1≤ r2 and s1 ≤ s2 is assumed so that the function is single valued and
monotonically increasing.
• Case 1: If r1 = s1 and r2 = s2, the transformation is a linear function that produces
no changes in intensity levels.
• Case 2: If r1 = r2, s1 = 0 and s2 = L - 1, the transformation becomes a thresholding
function that creates a binary image.
• Case 3: If (r1, s1) = (rmin, 0) and (r2, s2) = (rmax, L - 1), where rmin & rmax denote
the minimum and maximum intensity levels in the image, Contrast stretching is
obtained.

35
Contrast Stretching

36
 Intensity level slicing
• It is defined as Highlighting a specific range of
intensities in an image.
• Application include enhancing features such as
masses of water in satellite imagining and enhancing
flaws in X-ray images.
• Two approaches for implementing:
• One approach is to display in one value (say, white)
all the values in the range of interest.
• and in another (say, black) all other intensities.

37
Intensity level slicing

38
 Bit-plane slicing
• Highlight the contribution made to total image appearance by
specific bits.(We highlight the entire image appearance).
• Pixels are digital numbers composed of bits.
• Plane 1 contains least significant information of image & plane
8 contains the most significant information of image.

39
 Bit-plane slicing
• Fig 1 represent original image, Fig 2 represent bit plane 1
& fig 9 represent bit plane 8.

40
 Histogram Processing
• Histogram processing is used to improve the quality of the
image.
• Histogram shows the no. of pixels for each intensity value of
an given image.
• Histogram is a discrete function of any digital image having a
intensity level in [0, L-1].
• It is given by :

41
 Uses for Histogram Processing
• Image enhancements
• Image statistics
• Image compression
• Image segmentation
• Simple to calculate in software.

42
 Uses for Histogram Processing
• Four basic intensity characteristics and their
corresponding histograms
 Dark
 Light
 Low contrast
 High contrast
 Histogram commonly viewed in plot as

43
44
 Histogram Equalization
• Histogram equalization is used to enhance contrast.
• It is a technique used to obtain flat profile of
histogram by improving the quality of the image.

• To plot histogram equalization there are 2 conditions:

45
 Histogram Equalization
• Example of histogram equalization.

46
Histogram Equalization

47
 Histogram Equalization
• Let Pr (r) and Ps (s) be the probability density functions of r and s.
• From the basic of probability theory

• It shows that PDF of output intensity variable is obtained by the

PDF of input intensity and the transformation function used.
• This transformation function has the form

• Where w, is a dummy variable of integration.

• The right side of this equation is recognized as the cumulative
distribution function (CDF) of random variable r.

48
 Histogram Equalization
• Differentiating S= T (r), w. r. to r on both side

• Now, substitute eq. 3 in eq.1.

49
50
 Histogram Equalization
• For discrete values we deal with histogram values and
summations instead of probability density functions
and integrals.
• The probability of occurrence of intensity level in a
digital image is given by

• where MN is the total number of pixels in the image,

nk is the number of pixels that have intensity value rk.

51
 Histogram Equalization
• A plot of pr (rk) versus rk is commonly referred to as a
histogram.
• The discrete form of the transformation is

• The transformation T (rk) in this equation is called a

histogram equalization or histogram linearization
transformation.

52
 Histogram Equalization: Example
• Perform Histogram equalization on 3-bit image (L=8) of size
64 × 64 pixels (MN = 4096) has the intensity distribution
shown in following table.

Original image

53
 Histogram Equalization: Example
• The discrete form of the transformation is given by

54
Histogram Equalization: Example

55
Histogram Equalization: Example

56
 Histogram Matching
• Histogram equalization determines a transformation
function that seeks to produce an output image.

• It is useful to specify the shape of the histogram.

• So the method to generate a processed image which

has a specified histogram is called histogram
matching on histogram specification.

57
Histogram Matching

58
Histogram Matching

59
 Histogram Matching: Procedure
• Obtain pr(r) from the input image and then obtain the values of s
r
s  ( L  1)  pr ( w)dw
0

• Use the specified PDF and obtain the transformation function G(z)
z
G ( z )  ( L  1)  pz (t )dt  s
0

• Mapping from s to z

z  G 1 ( s )

60
 Histogram Matching: Example
Assuming continuous intensity values, suppose that an image
has the intensity PDF
 2r
 , for 0  r  L -1
pr (r )   ( L  1) 2

 0,
 otherwise
Find the transformation function that will produce an image
whose intensity PDF is
 3z 2
 , for 0  z  ( L -1)
pz ( z )   ( L  1) 3

 0,
 otherwise

61
 Histogram Matching: Example
Find the histogram equalization transformation for the input image
r r 2w r2
s  T (r )  ( L  1)  pr ( w)dw  ( L  1)  dw 
0 ( L  1) 2
0
L 1
Find the histogram equalization transformation for the specified
histogram
z z 3t 2 z3
G ( z )  ( L  1)  pz (t )dt  ( L  1)  dt  s
0 0 ( L  1)3 ( L  1) 2

We know that, G(z)= s. Thus the transformation function

1/3
1/3  2 r
2
 2 1/3
z  ( L  1) s 
2
 ( L  1)   ( L  1)r 
 L  1 

62
 Histogram Matching: Example
• For discrete intensity values: The transformation function
for discrete intensity values is given by,
k
( L  1) k
sk  T (rk )  ( L  1) pr (rj )   nj
j 0 MN j 0
• Use the specified PDF and obtain the transformation
function G(zq), round the value to the integer range [0, L-1].
q
G ( zq )  ( L  1) pz ( zi )  sk
i 0
• Mapping from sk to zq

1
z q  G ( sk )
63
 Local Histogram Processing
• In histogram equalization & matching, we
apply transformation function on the entire
image. So, this type of approach is called
Global.
• In local histogram we apply transformation
function on the neighborhood of the pixel.
From pixel to pixel we perform this operation.

64
Local Histogram Processing

65
Using Histogram Statistics for Image Enhancement

• The statistic obtained from an image histogram are

used for image enhancement.
• Let r denote a discrete random variable in the range
(0, L -1) and P(r) be the normalized histogram
corresponding to r.
• The nth moment of r about its mean is defined as

• Here m is the mean average intensity value of r and

66
 Histogram Statistics for Image Enhancement

• The intensity variance is given by

• We can directly estimate the mean & variance without

computing the histogram.
• These estimations are known as sample mean and sample
variance and we have sample mean.

67
 Histogram Statistics for Image Enhancement

• Let (x, y) denote the coordinates of any pixel in a given image,

and let Sxy denote a neighborhood of specified size, centered
on (x, y).
• The mean value of the pixels in this neighborhood is

• Where psxy, is the histogram of the pixel in region Sxy.

• The variance of the pixels in the neighborhood is

68
 Histogram Statistics : Example
Consider the following 2-bit image of size 5*5:

• The pixels are represented by 2 bits; therefore, L=4

and the intensity levels are in the range [0, 3].
• The total number of pixels is 25.

69
 Histogram Statistics : Example
Case 1 : When histogram of image is known

70
 Histogram Statistics : Example
Case 2: when histogram of image is not known

71
 Spatial Filtering
• Spatial filtering is a tool used in image processing for a
broad spectrum of applications.
• A spatial filter consists of (a) a neighborhood, and (b) a
predefined operation.
• Linear spatial filtering of an image of size M×N with a
filter of size m×n is given by:

• Where a & b are positive integers and

• x & y are varied so that each pixel in w visits every pixel
in f.

72
 Spatial Filtering
• Figure shows the mechanics of linear spatial filtering
using a 3×3 neighborhood.
• At any point (x, y) in the image, the response, g(x, y),
of the filter is the sum of products of the filter
coefficients and the image pixels encompassed by the
filter:

73
74
 Spatial Correlation & Convolution
• Correlation is the process of moving a filter mask
over the image and computing the sum of products at
each location.
• In convolution the mask or filter is first rotated by
180º.

75
76
77
78
General Expression for Correlation

79
General Expression for Convolution

80
 Vector Representation of Linear Filtering

• The characteristic response, R, of a mask either for

correlation or convolution is expressed as

• Where, Ws are the coefficients of an m×n filter.

• Zs are corresponding image intensities.

81
• As an example, a general mask 3×3 with
coefficients labeled in fig:

82
Classification of Spatial Filters
SPATIAL
FILTERS

SMOOTHING SHARPENING
FILTER SPATIAL FILTER

LINEAR NON-LINAER
FILTERS FILTERS

AVERAGING MEDIAN MIN & MAX

FILTERS FILTERS FILTERS
83
 Smoothing Spatial Filter
• Smoothing filters are used for blurring & for
noise reduction.
• Sometimes also used to extract the object
from the image & to remove small objects
present in the image.
• And used to bridge the small gaps in lines or
curves.

84
Smoothing filters – Example

Input image Smoothed image

85
 Smoothing Linear Filters
• These are also called as averaging filter.
• Linear filter is simply the average of the pixels
contained in the neighborhood of the filter
mask.
• This operation is performed pixel to pixel, by
shifting its origin.
Drawback: It blur the edges of the image, which
is the main feature in an image.

86
 Smoothing Linear Filters: Averaging
• Response of the averaging filter

87
 Smoothing Linear Filters: Averaging
• Mask size determines degree of smoothing (i.e., loss
of detail).

88
 Smoothing Linear Filters: Averaging

• Example: extract largest, brightest objects.

15 x 15 averaging Image thresholding

89
 Smoothing Nonlinear filters
• Order-Statistic (Nonlinear) Filters.
• Replacing the value of the center pixel with the
value determined by the ranking result.
• E.g., median filter, max filter, min filter.
• Median Filter: Replace each pixel by the median
in a neighborhood around the pixel.
• The size of the neighborhood controls the amount
of smoothing.

90
Median filtering

91
 Median filtering
• Median filters are popular because for certain types of random
noise, they provide noise reduction capabilities.
• Very effective for removing ‘salt and pepper’ noise (i.e.,
random occurrences of black and white pixels).

92
 MAX & MIN filters
• Max Filters: The highest value is assigned to the
center pixel, which is useful for finding the
brightest points in an image.
• The response is given by:

• Min Filters: Smallest pixel is assigned to the

center pixel.

93
 Sharpening Spatial Filter
• Used to highlight the transitions in intensity.
• Useful for highlighting fine details.
• Enhance the blurred image and edges.
• It is accomplished by spatial differentiation.
• Applications: Electronics printing, medical
imaging and autonomous guidance in military
systems.

94
Sharpening Filters - Example
Input image Sharpened image

95
 Basics of Sharpening
• To understand the concept various derivatives are used.
• For a first order derivative
1. It must be zero in areas of constant intensity.
2. It must be non-zero at the onset of an intensity step or ramp
and
3. It must be non-zero along ramps.

The first order derivative of 1-D function is given by

96
• For a second order derivative
1. It must be zero in constant areas.
2. It must be non-zero at the onset and end of intensity
ramp or step.
3. It must be zero along ramps of constant slope.

The second order derivative of 1-D function is given by

97
First & Second order derivative

98
 Second Order Derivative for Image Sharpening:
(The Laplacian Operator)

• Here Isotropic filters are preferred which are

rotational invariant.

• It means if we are rotating the image and then

applying the filter will gives the same result as
applying the filter to that image first and than rotating
the result.

99
• The laplacian operator for a function f(x,y) is given as
……….(1)
• In the x-direction
……….(2)
• In the y-direction
……….(3)

• Substitute eq. 3 & 2 in eq. 1.

• General laplacian operator for image sharpening is

Where f(x,y) is an input image and g(x,y) is resultant sharpened

image. C is constant.

100
2D image f(x,y) Filter Mask

101
Image Smoothing & Sharpening using frequency
domain

102
 Frequency domain
• In this, first we transform an image into the
transform domain (Fourier transform), doing the
processing there, and obtaining the inverse
transform to bring the results back into the spatial
domain.

103
 2-D Discrete Fourier Transform & Its Inverse

2-D discrete Fourier transform (DFT):

• Where f(x, y) is a digital image of size MxN.

• u and v are frequency domain variables.
• u = 0, 1, 2,....,M-1 & v = 0, 1, 2,...,N-1.

104
 2-D Discrete Fourier Transform & Its Inverse

If the transform F(u, v), than we can obtain f(x, y) by

using the inverse discrete Fourier transform(IDFT):

For x = 0, 1, 2,....,M-1 & y = 0, 1, 2,...,N-1.

105
 Properties of 2-D discrete Fourier Transform

1. Relationships Between Spatial & Frequency Intervals

The separations between the corresponding discrete,
frequency domain variables are given by:

• F(t, z) is continuous function is sampled to form a

digital image f(x, y), consisting of M×N samples in t
& z direction.
• ∆t & ∆z denote the separations between samples.

106
2. Translation and Rotation
Fourier transform pair satisfies the following translation
property as

• Multiplying f(x, y) by the exponential shown shifts

the origin of the DFT to (u0, v0).
• Multiplying f(u, v) by the negative of that exponential
shifts the origin of f(x, y) to (x 0, y0).
This represents translation property.

107
Using the polar coordinates

Substituting this in equation, will obtain

• It indicates that rotation f(x, y) by an angle θ0 rotates

F(u,v) by the same angle.
• Conversely, rotating F(u, v) rotates f(x, y) by the
same angle.

108
3. Periodicity
• The Fourier transform & its inverse are
infinitely periodic in the u and v directions, i.e.

• Where k1 and k2 are integers.

109
4. Symmetry Properties
Any real & complex function are the sum of even & odd
parts.
……………1

where the even and odd parts are defined as

…………..2

……………..3

Substitute eq. 2 & 3 in eq.1 gives the identity

w(x, y)Ξ w(x, y)

110
5. 2 D convolution
Circular convolution of 2-D DFT is given as

Here x= 0,1,2,……. M-1 & y= 0, 1, 2,…..N-1.

• The 2-D convolution theorem is given by

111
 Image Smoothing using Frequency Domain filters

• Smoothing (blurring) is achieved in the frequency

domain by high-frequency attenuation; i.e. by low
pass filtering.
• Three types of low pass filters:
1. Ideal low pass filtering.
2. Butterworth low pass filters.
3. Gaussian low pass filters.
• These three categories cover the range from very
sharp (ideal) to very smooth (Gaussian) filtering.

112
 Image Smoothing using Frequency Domain filters

• The Butterworth filter has a parameter called the

filter order.
• For high order values, the Butterworth filter
approaches the ideal filter.
• For lower order values, the Butterworth filter is
more like a Gaussian filter.
• Thus, the Butterworth filter may be viewed as
providing a transition between two “extremes.”

113
1. Ideal low pass filters:
• A 2-D lowpass filter that passes without attenuation
all frequencies within a circle of radius D0 from the
origin and “cuts off” all frequencies outside this circle
is called an ideal lowpass filter (ILPF).
• The transfer function of ILPF for 2D image in
frequency domain can be represented as:

• Where D0 is a positive constant and D(u,v) is the

distance from the center of the frequency rectangle.

• Where u=0,1,…..P-1 & v=0,1,…..Q-1. P & Q are

padded size.
114
115
2. Butterworth Low pass Filters:
• The transfer function of a Butterworth lowpass filter
(BLPF) of order n & distance D0 can be represented
as:

• Its cut off depends on the order of the filter.

• For high order values, the Butterworth filter
approaches the ideal filter.
• For lower order values, the Butterworth filter is more
like a Gaussian filter.

116
• If n=3 order shows the ideal LPF. Its almost gives the
sharp transition.
• If n=1 order BLPF gives very smooth transition.

117
118
3. Gaussian Low pass Filters
• The transfer function of a Gaussian lowpass filter
(GLPF) for 2-D images in frequency domain is
represented as:

• D(u,v) is the distance from the center of the

frequency rectangle.
• If σ = D0, than

• Where D0 is the cutoff frequency.

• GLPF gives very smooth transaction in filtering
operation.

119
120
Image Sharpening Using Frequency Domain Filters

• Image sharpening can be achieved in the frequency domain by

highpass filtering.
• It attenuates the low-frequency components without disturbing
high-frequency information in the Fourier transform.
• H(u,v) is discrete function of size P×Q & discrete frequency
variables are u=0,1,…..P-1 & v=0,1,…..Q-1.
• A highpass filter is obtained from a given lowpass filter using
the equation

• Where HLP (u, v) is the transfer function of the lowpass filter.

121
Image Sharpening Using Frequency Domain Filters

So, HPF can be achieved by using LPF by using

three high pass filters:
1. Ideal Highpass Filters.
2. Butterworth Highpass Filters.
3. Gaussian Highpass Filters.

122
 1. Ideal Highpass Filters:
A 2-D ideal highpass filter (IHPF) is defined as

• Where D0 is the cutoff frequency & D(u,v) is the

distance from the center of the frequency rectangle.

Fig: Perspective plot, image representation, and cross section of a typical ideal highpass Filter.

123
2. Butterworth Highpass Filters:
A 2-D Butterworth highpass filter (BHPF) of order
n and cutoff frequency D0 is defined as

• It is used to control the sharpening. Depending on

the order it provides the sharpening effect.

Fig: Perspective plot, image representation, and cross section of a Butterworth highpass Filter.
124
3. Gaussian Highpass Filters:
The transfer function of a Gaussian highpass filter
(GLPF) is represented by:

• It provides very high sharpening effect than other

filters.

Fig: Perspective plot, image representation, and cross section of a Gaussian highpass Filter.
125
Color Models

126
RGB Color Model

127
128
129
130
131
CMY Color Model

132
133
134
HSI Color Model

135
136
137
138
139
140
Color Transform

141
 Converting colors from RGB to HSI
• Given an image in RGB color format, the H component of
each RGB pixel is obtained using this equation:

142
Converting colors from HSI to RGB

143
 Pseudo Coloring
• Pseudo color processing is a technique that maps
each of gray levels of black & white image into
an assigned color.
• Images are grayscale which are assigned color
based on the intensities values.
• This colored image, when displayed can make
the identification of certain features easier for
the observer.

144
 Pseudo Image Processing
• Basic concept of the grey level to color
transformation.
• Performing three independent transformations on the
grey level of any input pixel.

145
 Wavelet
• Wavelets are the foundation of representing
images in various degrees of resolution. It is
used for image data compression.
• A Wavelet is a waveform of an effectively
limited duration that has an average value of
zero.

146
 What is a wavelet?
• A function that “waves” above and below the x-axis
with the following properties:
– Varying frequency
– Limited duration
– Zero average value
• A wavelet is a waveform of effectively limited
duration that has an average value of zero.

147
148
 Wavelets vs. Fourier Transform
• In Fourier transform (FT) we represent a signal in
terms of sinusoids
• FT provides a signal which is localized only in the
frequency domain
• It does not give any information of the signal in the
time domain.
• Basis functions of the wavelet transform (WT) are
small waves located in different times.
• Therefore, the WT is localized in both time and
frequency.

149
 Wavelet Transform
• Wavelet transform is an efficient tool to
represent an image.
• Wavelet transform allows multi resolution
analysis of an image.
• Aim of transform is to extract relevant
information from image.
• Wavelet transform provides a time frequency
representation of signal.

150
 Types of Wavelets
• There are many different wavelets, for
example:

Haar Morlet Daubechies

151
 Haar Wavelet
• It is bipolar step function. Expression for Haar
wavelet is given by

152
 Morelet Wavelet
• The morelet wavelet is obtained by multiplying
the fourier basis with a gaussian window.

153
 Daubechies wavelet
• They are a family of orthogonal
wavelets defining a discrete wavelet
transform and characterized by a maximal
number of vanishing moments for some
given support.
• With each wavelet type of this class, there is a
scaling function (called the father wavelet)
which generates an orthogonal multiresolution
analysis.

154
 Continuous Wavelet Transform
• The continuous wavelet transform of one
dimensional signal x(t) is given by

• Thus continuous wavelet transform is a function

of two variables a & b. Here a is scaling
parameter & b is shifting parameter.

155
156
 Discrete Wavelet Transform
• It is obtained by filtering the signal through a series
of digital filters at different scales.
• Scaling operation is done by changing the resolution
of signal by sampling.
• DWT can be computed by either convolution or
shifting based procedures.
• DWT uses a finite set of wavelets i.e. defined at a
particular set of scales and locations.

157
 Wavelets Applications
• Noise filtering
• Image compression
– Special case: fingerprint compression
• Image fusion
• Recognition
• Image matching and retrieval

158
 Assignment Questions
Q1 What do you understand by histogram? Explain its various
specifications also explain histogram equalization and matching.
Q2 Explain various basic intensity transfer function?
Q3 Describe the mechanism of the spatial filtering. Also Explain
how smoothing is done by filters in an image.
Q4 What is Fourier transform? Explain its properties.
Q5 Explain Image Smoothing & Sharpening Using Frequency
Domain Filters.

159