ZERO

TO

HERO

DIGITAL IMAGE PROCESSING

USING MATLAB
 ZERO TO HERO

DIGITAL IMAGE PROCESSING USING

MATLAB

A RSATH N ATHEEM S

Copyright © 2017, 2018 by futurebme.com

2
WHY I WROTE THIS BOOK
I wrote this book because Image processing is fresh and interesting topic
for Research work. This book is all about that how to develop theory and project
based notion about Image processing. This book contain plenty of programming
illustration that are analyzed can also be beneficial for learners and under
graduates pupils. This book starts from very basic Knowledge and gradually
cover all of the improvement topics of a Image processing with MATLAB
examples.

3
WHY YOU SHOULD READ THIS BOOK
This book will help you learn all about digital image processing
Importance, and necessity of image processing stems from application areas the
first being the Improvement of data for individual interpretation and the second
being that the Processing of a spectacle data for an machine perception. Digital
image processing includes a assortment of applications such as remote sensing,
image and information storage for transmission in acoustic imaging, medical
imaging, business applications , Forensic sciences and industrial automation.
Images are helpful in tracking of earth resources mapping, and forecast of urban
populations, agricultural crops, climate forecasting, flooding and fire control.
Space imaging applications include comprehension and analyzation of objects
contained in images obtained from deep space-probe missions. There are also
medical programs such as processing of X-Rays, Ultrasonic scanning, Electron
micrographs, Magnetic Resonance Imaging, Nuclear Magnetic Resonance
Imaging, etc.. In addition to the aforementioned applications, digital image
processing is being used to solve a variety of issues. Even unrelated, these
problems commonly require methods effective at improving information. The
Image processing Procedures like restoration and Image enhancement are used to
procedure images that were degraded or blurred. Powerful uses of image
processing concepts are observed in defense astronomy, biology, medical and
industrial applications. As per Medical Imaging is concerned almost all of the
pictures could be utilized in the discovery of tumors or for viewing the patients.
The current key field of use of digital image processing (DIP) methods is in
solving the issue of machine vision so as to attain superior results

4
 TABLE OF CONTENTS

CHAPTER TITLE PAGE

NO NO
Introduction 7

1 Basic Morphological Operation 18

MATLAB program for dilation, erosion, opening, 23

closing and their properties step by step explanation

2 Image Segmentation 36

Thresholding 37

Region, Edge based segmentation 41

MATLAB Source Code for Image Segmentation 55

Output for Image Segmentation 60

Application, Conclusion 65

3 Intensity Transformation 67

Theoretical Concepts: Introduction 67

MATLAB program for Image Intensity 74

Transformation
4 Histogram Equalization 84

Introduction:Practical approach 84

MATLAB Source Code for Histogram Equalization 93

Conclusion 104

5 Spatial Intensity Resolution 105

5
 MATLAB Source Code For Spatial Intensity 111
Resolution

6 Enhancement in Spatial Filter 119

MATLAB Source Code For Image Enhancement in 127

Spatial Filtering

7 Enhancement in Frequency Filter 134

MATLAB Source Code For Image Enhancement in 142

Frequency Filtering

8 Color Image Processing 149

MATLAB Source Code For Color Image 164

Processing

9 DFT (Discrete Fourier Transform) Analysis 180

MATLAB Source Code For DFT Analysis

10 Basic Thresholding Function

MATLAB Source Code For Basic Thresholding

Function

11 Sampling and Quantization

MATLAB Source Code For Sampling and

Quantization

12 Image Transformation

MATLAB Source Code For Image Transformation

6
 INTRODUCTION

What is MATLAB?

• MATLAB = Matrix Laboratory

• “MATLAB is a high-level language and interactive environment that
allows you to achieve computationally intensive tasks faster than with
old-style programming languages like C, C++ and Fortran.”
(www.mathworks.com)
• MATLAB is an interactive, understood language that is designed for
fast numerical matrix calculations

MATLAB window components:

7
Workspace

Displays all the defined variables

Command Window

To perform commands in the MATLAB environment

Command History

Displays record of the commands used

File Editor Window

Define your functions

MATLAB Help

8
• MATLAB Help is an very powerful assistance to learning MATLAB

• Help not only contains the theoretic background, but also shows demos
for implementation

• MATLAB Help can be unlocked by using the HELP pull-down menu

9
 • Some command description can be found by typing the command in the
search field

• As shown above, the command to take square root (SQRT) is examined

• We can also use MATLAB Help from the command window as shown

More about the Workspace

• who, whos – present variables in the workspace

• save – save workspace variables to *.mat file

• load – load variables from *.mat file

• clear – clear workspace variables

Matrices in MATLAB

Matrix is the main MATLAB data type

How to build a matrix?

➢ A=[1 2 3; 4 5 6; 7 8 9];
➢ Creates matrix A of size 3 x 3

Special matrices:

• zeros(n,m), ones(n,m), eye(n,m), rand(), randn()

• Numbers are always double (64 bits) unless you specify a
dissimilar data type

10
Basic Operations on Matrices

• All operators in MATLAB are defined on matrices: +, -, *, /, ^,

sqrt, sin, cos, etc.

• Element-wise operators defined by a previous dot: .*, ./, .^

• sum(A) – columns sums vector

• size(A) – size vector

• sum(sum(A)) – sum of all the elements

Variable Name in Matlab

✓ Variable naming rules

- must be unique in the first 63 characters

- Necessity begin with a letter

- may not contain blank spaces or additional types of punctuation

- may contain any grouping of letters, digits, and underscores

- are case-sensitive

- should not use Matlab keyword

✓ Pre-defined variable names

pi

Logical Operators

• ==, <, >, (not equal) ~=, (not) ~

• find(‘condition’) – Returns directories of A’s elements that satisfy

the condition

11
 • Example:

>>A=[7 3 5; 6 2 1], Idx=find(A<4)

A=

735

621

Idx=

Flow Control

MATLAB has five flow control concepts:

1. if statement

2. for loop

3. while loop

4. switch statement

5. break statement

12
Scripts and Functions

There are two kinds of M-files:

1. Scripts, which don’t accept input opinions or return production

arguments. They operate on data in the workspace
2. Functions, which can accept input opinions and return
production arguments. Interior variables are local to the
function

• Example:

A file called STAT.M:

function [mean, stdev]=stat(x)

%STAT Interesting statistics.

n=length(x);

mean=sum(x)/n;

stdev=sqrt(sum((x-mean).^2)/n);

Defines a novel function called STAT that computes the mean and
standard deviation of a vector. Function name and file name should
be the SAME!

Visualization and Graphics

• plot(x,y),plot(x,sin(x)) – plot 1D function

• figure, figure(k) – open a fresh figure

• hold on, hold off – refreshing

• axis([xmin xmax ymin ymax]) – change axes

• title(‘figure titile’) – add title to figure

13
 • mesh(x_ax,y_ax,z_mat) – view surface

• contour(z_mat) – view z as topo map

• subplot(3,1,2) – find several plots in figure

Saving your Effort

Save mysession

% creates mysession.mat with all variables

Save mysession a b

% save only variables a and b

Clear all

% clear all variables

Clear a b

% clear variables a and b

Load my session

% load session

14
What is the Image Processing Toolbox?

• The Image Processing Toolbox is a group of functions that extend the

abilities of the MATLAB’s numeric scheming environment. The toolbox
supports a wide range of image processing operations, together with:

▪ Image analysis and enhancement

▪ Geometric operations
▪ Linear filtering and filter design
▪ Neighborhood and block operations
▪ Transforms
▪ Region of interest operations
▪ Binary image operations

YOU CANNOT USE THE IMAGE PROCESSING TOOLBOX FOR

HOMEWORK OR FINAL PROJECT

Images in MATLAB

MATLAB can import/export some image formats:

▪ JPEG (Joint Photographic Experts Group)

▪ GIF (Graphics Interchange Files)

▪ PNG (Portable Network Graphics)

▪ TIFF (Tagged Image File Format)

▪ BMP (Microsoft Windows Bitmap)

▪ HDF (Hierarchical Data Format)

▪ PCX (Paintbrush)

▪ XWD (X Window Dump)

15
 ▪ raw-data and other types of image data

▪ Naturally switch images to double to achieve any processing and

convert back to unsigned integer

Data types in MATLAB

▪ Single (32-bit single-precision floating point)

▪ Double (64-bit double-precision floating point)
▪ Int16 (16-bit signed integer)
▪ Int32 (32-bit signed integer)
▪ Int8 (8-bit signed integer)
▪ Uint8 (8-bit unsigned integer)
▪ Uint16 (16-bit unsigned integer)
▪ Uint32 (32-bit unsigned integer)

Images in MATLAB

16
Image Import and Export

Read and write images in Matlab

img = imread('apple.jpg');

dim = size(img);

figure;

imshow(img);

imwrite(img, 'output.bmp', 'bmp');

Alternatives to imshow

imagesc(I)

imtool(I)

image(I)

Image Display

• imshow - display image

• image - create and display image object

• imagesc - scale and display as image

17
 CHAPTER 1.

BASIC MORPHOLOGICAL OPERATION

Introduction

Image Processing is a technique to perform some functions on an image,

in order to have an improved image or to extract various interesting facts from it.
Image Morphology is an essential tool in image processing. It is actually the learn
shapes of object present in the image and extraction of image features. Image
features are essential for object recognition. The basic morphological operations
include Erosion and Dilation, Opening and Closing are also morphological
operators. These operators are deciding on as basic operations in image
processing algorithms.

DILATION
Dilation operator can be used to binary and grey scale images. The purpose
of this operator is to expand the window and shrinks background. It slowly
increases the boundaries of the region, while the small holes existing in the image
become smaller. It increases the illumination of the object.

Fig.1.Example: Dilation

18
 Fig.1.1. Applied Structuring Element

EROSION
Erosion is significant operation. The purpose of erosion operators is to
shrinks the window and grows background. Erosion is used to make an object
shorter by eliminating is outside region of pixels. After implementing the erosion
operator on the image, the image gets darker. This particular operator will take
the image and structuring element as inputs and thins the subject.

Fig.1.2.Example: Erosion

19
 Fig.1.3. Applied Structuring Element

OPENING
Opening operation is mixed of dilation and erosion operations. If A and are
two sets of pixels, then in the opening, 1st erode A by B then dilate the result by
B. Opening is the union of all B objects

Totally contained in A.

• Similar to Erosion
• Spot and noise removal
• Less destructive
• Erosion next dilation
• The similar structuring element for both operations.

Input:

• Binary Image
• Structuring Element.

20
 Fig.1.4.Example: Opening

➢ Choose the structuring element (SE) and move it around inside each
highlight area.
➢ All highlight pixels which cannot be accessed by the structuring element
without lapping over the edge of the feature object will be eroded away
➢ All pixels which can be protected by the SE with the SE being completely
within the highlight region will be preserved.
➢ Opening is idempotent, repeating application has no further impact

21
CLOSING
Closing operation is a dilation operation adopted by an erosion
operation. Closing is actually the group of points, which the intersection of
object B about them with object A is not vacant.

Fig.1.5. Example: Closing

➢ Choose the structuring element (SE) and move it around outside each
highlight area.

➢ All background pixels which can be protected by the SE with the

structuring element to be entirely within the background area will be
preserved.

➢ All background pixels which can't be achieved by the structuring element

without lapping over the edge of the window object will be turned into a
foreground.

➢ Opening is idempotent: repeating application has no additional effects

22
MATLAB PROGRAM FOR DILATION, EROSION, OPENING, CLOSING
AND THEIR PROPERTIES

SOURCE CODE:

clc;

clear all;

close all;

a=imread('C:\Users\natheem\Desktop\images\India_satellite.png');

% Read an input image

b=strel('line',11,90);

% strel represents morphological structuring element

figure(1);

imshow(a);

title('Original image');

c=imdilate(a,b);

figure(2);

imshow(c);

title('Dilate image');
23
d=strel('disk',11);

e=imerode(a,d);

figure(3);

imshow(e);

title('Erode image')

c=imopen(a,b);

figure(4);

imshow(c);

title('Open image');

d=strel('disk',11);

e=imclose(a,d);

figure(5);

imshow(e);

title('Close image')

%Properties

h=[0 1 0;1 1 1;0 1 0];

24
i1=imdilate(a,h);

figure(6);

imshow(i1);

title('Dilate image pro1'); % pro=properties

i2=imerode(a,h);

figure(7);

imshow(i2);

title('Erode image ');

i3=a-i2;

figure(8);

imshow(i3);

title('pro3');

i4=i1-a;

figure(9);

imshow(i4);

title('pro4');

25
i5=i1-i2;

figure(10);

imshow(i5);

title('pro5');

26
STEP BY STEP EXPLANATION

STEP 1: Type Program on MATLAB comment window

STEP 2: Save program & Run Program Click ‘Add to path’

27
STEP 3: You will get Corresponding Output for Morphological Operation

28
RESULT:

Fig.1.Orignal Image

Fig.2.Dialating Image

29
Fig.3.Eroded Image

Fig.4.Opening Image

30
 Fig.5.Closing Image

Fig.6.Dialate Image property 1

31
Fig.7.Erode Image property 2

Fig.8. property 3

32
 Fig.9. property 1

Fig.5.Dialate Image property 1

33
FUTURE SCOPE

The Morphological Image Processing can be further applied to a wide

range of problems including

➢ Medical image analysis:

• Tumour detection,
• Regurgitation,
• Measurement of size and shape of interior organs.

➢ Robotics:
Recognition and understanding of objects in a scene, motion control
and implementation through visual feedback

➢ Radar imaging:
• Target detection and identification.
• Selection of Structuring element for object classification over
morphology is still interesting to this technique and has been chosen
to be the main direction of the future work.

APPLICATIONS
Morphology is applied as a technique for image transformation. It
has been utilized for extraction of edges and detection of the attribute
objects in mobile photogrammetric techniques to making maps from
images taken from a car, named mobile mapping systems Morphology is
used primarily for decrease an region of interest and extracting particular
objects like street signals. Features of morphology are also used in
discovering sewer pipelines problems. Architectural commemorations as

34
 well as industrial things have edges and areas which can be probably
detected by use of mathematical morphology functions.

CONCLUSION

The handling of image is faster as well as more cost effective.

Morphological image processing characterized an image processing
technique which compact with the form of features in an image. in this
chapter use of morphological operators are described with morphological
algorithms. This section highlighted the Morphological operations like as
Dilation, Erosion, and Opening, Closing and morphological processes like
Boundary Extraction,

Thickening, thinning, Noise elimination and pruning which are

really helpful process or implement any image. Most Application areas of
image processing tend to be Medical imaging, Industrial automated,
Biometrics, Cinematography, Armed Application. Image Processing
applications are existing in all area.

35
 CHAPTER 2.

IMAGE SEGMENTATION

What is segmentation?

Image segmentation is the process of partitioning an image into a collection

of connected sets of pixels. The goal of segmentation is to simplify change the
representation of an image into something that is more meaningful and easier to
analyses. Image segmentation is typically used to locate objects and boundaries
like lines, curves in images. It is the process of assigning a label to every pixel in
an image.

Three main techniques to do.

1. Thresholding

2. Region based segmentation

3. Edge based segmentation

36
 THRESHOLDING

Thresholding is finding histogram of gray level intensity.

Types

1. Basic Global Thresholding

2. Multiple Threshold
3. Variable Thresholding
4. Otsu’s Technique

Basic Global Thresholding

• Primarily Segment image use:

g ( x, y)   01 if f(x,y)T
if f(x,y)T

• Calculate the average intensity m1 and m2 for the pixels

37
• Calculate a fresh threshold:

1
T (m 1  m 2 )
2

• Until the variance between values of T is minor than a predefined parameter.

38
Otsu’s Technique

• Based on a very simple idea: Find the threshold that minimizes the
weighted within-class variance.
• This turns out to be the same as maximizing the between-class variance.
• Operates directly on the gray level histogram.

Multiple Threshold

As Otsu’s technique, it takes extra area and k*

Disadvantage:

It becomes too complex when amount of region more than two or three.

39
Variable Thresholding

1) Image partitioning

It is work once the objects of interest and the background inhabit areas of
sensibly similar size. If not, it will fail.

2) Variable thresholding based on local image properties

3) By moving average

It debated is based on computing a moving average along scan

appearances of an image.

40
 REGION BASED SEGMENTATION

Region Segmentation is procedure of finding region, but not finding edge

1. Region Growing
2. Cheng-Jin Kuo`s method is used
3. Data Grouping (Clustering Technique)
4. Partitional clustering

Algorithm:

• Select a random pixels

• Usage 8-connected and threshold to decide
• Repeat a and b until nearly points are classified.

41
 Reproduction of region growing (90% pixels )

Threshold/second: 20/4.7 seconds.

42
 Data Clustering

• Using centroid to signify the huge amounts of clusters

• Partitional clustering, we have to choose the number of clustering we
want first before we begin the process.
• Hierarchical clustering, we can alteration the number of cluster anytime
throughout process if we need.

Hierarchical clustering

1) Algorithm of hierarchical accumulation (built)

• Understand every solo data as a cluster Ci
• Find out Ci, Cj for the distance is the shortest.
• Repeat the steps until satisfies our request.
• d(a,b) as the space between data a and b
2) Algorithm of hierarchical division (break up )：

Diameter of cluster

D(Ci )  max(d (a,b)) , for a  Ci , b  Ci

43
Partitional clustering

Choose the numbers of the cluster (k-means)

44
ADVANTAGES & DISADVANTAGES

Hierarchical algorithm

Advantages

• Result is reliable
• Idea is simple

Disadvantages

• It is consuming, so is not suitable for a big database.

Partitional algorithms

Advantages

• Numbers of cluster is fixed, so the concept is also simple.

• Calculating speed is fast.

Disadvantages

• Determine the number of clusters.

• Initial problem

45
 EDGE BASED SEGMENTATION

Edge-Based Segmentation

Edge-Based Segmentation is by mask to notice edge in image by

convolution.

1. Basic Edge Detection

2. SRHLT - Short response Hilbert transform
3. Watersheds
4. The Marr Hildreth edge detector (LoG)

Basic Edge Detection

46
Gradient

 Gradient is to find edge strength and direction at site (x,y) of image.

 f 
 g x   x 
f  grad( f )      
 g y   f 
 y 

 The magnitude of vector, meant as M(x,y)

mag(f )  g x  g y

 The direction of the gradient vector is assumed by the angle

gy
 ( x, y)  tan 1
 
 g x 

Roberts

Robert edge detection, the perpendicular and horizontal edges carry out
separately and then place together for resulting edge detection. The Roberts edge
detector uses the next masks to approximate digitally the first derivatives as
differences between adjacent pixels.

47
Prewitt operator

Prewitt operator edge detection masks are the one of the oldest and
greatest understood methods of detecting edges in images The Prewitt edge
detector uses the following mask to approximate digitally the first derivatives Gx
and Gy.

Sobel operator

The sobel edge detector calculates the gradient by using the discrete
differences between rows and columns of a 3X3 neighborhood. The sobel
operator is based on convolving the image with a minor, divisible, and number
valued filter.

G( x, y)   1
0
if R ( x , y ) T
otherwise

48
Canny Edge Detection

Canny edge detection is a multistage algorithm to detect a extensive range

of edges in images. This detector finds edges by observing for local maxima of
the gradient off (x, y). The gradient is calculated using the derivative of a
Gaussian filter. The technique uses two thresholds to detect strong and weak
edges and includes the weak edges in the output only if they are connected to
strong edges

49
The Marr-Hildreth edge detector (LoG)

This is second-order deviation, known as Laplacian.

Filter the input image with an n*n Gaussian low pass filter. 99.7% of the
capacity under a 2-D Gaussian surface lies between around the mean.

50
 Short response Hilbert Transform (SRHLT)

1) Hilbert transform

1
g H ( )  h( x) * g ( x), where h( x) 
x

GH ( f )  H ( f )G ( f ) H ( f )   j sgn( f )

2) Short response Hilbert transform (SRHLT)

g H    hb  x   g  x  , where hb  x   b csch( bx)

GH ( f )  H b ( f )G( f ) where GH ( f )  FT [ g H ( )],

G( f )  FT [ g ( )], H b ( f )   j tanh( f / b).

51
 Choose a suitable
HLT value differentiation
b0 b

52
Watersheds

Algorithm:

53
Markers

Outside markers:

• Points along the watershed line along maximum points.

Interior markers:

• All points in connected component have the similar intensity.

• Points in area form a connected component
• That is surrounded higher points.

MATLAB CODE FOR IRIS SEGMENTATION

Iris segmentation is very significant for an iris recognition system. If the
iris regions were not properly segmented, there would possibly exist four kinds
of noises in segmented iris areas: eyelashes, eyelids, reflections and pupil, which
will outcome in poor recognition performance. This paper proposes a new noise-
removing method, Founded on the fusion of edge and region data.

THE ENTIRE PROCEDURE INCLUDES 3 STEPS:

1) Irregular localization and normalization,

2) Edge data extraction based on phase congruency, and
3) The infusion of edge and region data

54
MATLAB GUI for Iris Segmentation

MATLAB CODE FOR IRIS SEGMENTATION

Here you can download this code from this link

https://www.pantechsolutions.net/image-processing-projects/matlab-
code-for-iris-segmentation

MATLAB code for K-means Segmentation

https://www.dropbox.com/s/um4zhhb5qhfj9kt/matlab_code_for_k_means
_segmentation.rar?dl=0bb

55
MATLAB SOURCE CODE FOR IMAGE SEGMENTATION

Flow Chart

%IMAGE SEGMENTATION

clc;

clear all;

close all;

%edge detection

a=imread('autumn.tif');

b=rgb2gray(a);

56
figure(1);

imshow(b);

title('original image');

%roberts operator

c=edge(b,'roberts');

figure(2);

imshow(c);

title('roberts image');

%sobel operator

d=edge(b,'sobel');

figure(3);

imshow(d);

title('sobel image');

%prewitt operator

e=edge(b,'prewitt');

figure(4);

imshow(e);

57
title('prewitt image');

%log operator

f=edge(b,'log');

figure(5);

imshow(f);

title('log image');

%canny operator

g=edge(b,'canny');

figure(6);

imshow(g);

title('canny image');

%point detection

h=[0,1,0;1,-4,1;0,1,0];

i=imfilter(b,h,'symmetric','conv');

figure(7);

imshow(i);

title('laplacian point filtered image');

58
%line detection

j=im2bw(b);

w=(1/3)*[1,0,-1;1,0,-1;1,0,-1];

k=imfilter(double(j),w);

k(k<1)=0;

figure(8);

imshow(k);

title('line detection image');

59
OUTPUT FOR IMAGE SEGMENTATION

Fig.1 Original Image

Fig.2. Gray Scale image

60
Fig.3. Robert image

Fig.4. Sobel Image

61
Fig.5. Prewitt Image

62
 Fig.6. Log image

Fig.7. Canny Image

63
Fig.8. Laplacian Point Filtering Image

Fig.9. line detection Image

64
APPLICATION

Real-world applications of image segmentation are:

1. Machine vision
2. Deep Learning
3. Medical imaging, including volume rendered images from CT (Computed
Tomography) and MRI (Magnetic Resonance Imaging).
4. Measure tissue volumes
5. Surgery planning
6. Intra-surgery navigation
7. Locate tumors and pathologies
8. Virtual surgery simulation
9. Diagnosis, learning of anatomical structure
10.Pedestrian recognition
11.Object recognition
12.Face detection
13.Face recognition
14.Recognition Tasks
15.Fingerprint recognition
16.Iris recognition
17.Traffic controller schemes
18.Brake light detection
19.Video investigation
20.Find objects in satellite images like roads, forests, crops.

65
CONCLUSION
Since edge detection is the first step in object boundary extraction and
object recognition, it is significant to know the differences between different edge
detection operators. from this section an attempt is made to review the edge
detection methods which are based on discontinuity intensity levels. The relative
presentation of various edge detection techniques is carried out with two images
by using MATLAB. It have been experiential that that the Canny edge detector
produces greater accuracy in detection of object edges with greater entropy,
PSNR, MSE and implementation time compared with Sobel, Roberts, Prewitt,
Zero crossing and LOG.

66
 CHAPTER 3.

INTENSITY TRANSFORMATION

THEORETICAL CONCEPTS: INTRODUCTION

What is Intensity Transformation?

Intensity transformation is increase the contrast between certain Intensity

values, most important application of intensity transformation is Enhance
the low Quality image.

Image Enhancement Techniques

1. Spatial operates on Pixels

2. Frequency domain operates on Fourier transform of image

Spatial Domain Methods

• Spatial Domain Technique a operation (linear or non-linear) is performed
on the pixels in the neighborhood of coordinate (x,y) in the input image F,
giving enhanced image F’
• Neighborhood can be any shape but generally it is rectangular ( 3x3, 5x5,
9x9)
g(x,y) = T[f(x,y)]

67
 Grey Scale Manipulation

• Simplest form of window (1x1)

• Assume input gray scale values are in range [0, L-1] (in 8 bit images L =
256)
• Nth root Transformation S = c (r)n .

Intensity Transformation function

• Linear:
1. Negative
2. Identity
• Logarithmic:
1. Log
2. Inverse Log
• Power-Law:
1. nth power,
2. nth root.

68
Negative Image

S = (L – 1) – r

• L- Number of gray Level in image

• r- Pixel of input image

(Courtesy of G.E Medical system)

Fig.1. Original Mammogram image
Fig.2.Negative Image Obtained Using Negative Intensity transformation

Log Transformation

Compresses the dynamic range of images with big variations in pixel

values

• s = c log(1+r)
• c: constant

69
 a) Fourier Spectrum b) Result of apply log transformation

Power Law Transformation

• s = crγ for various values of  (c=1 in all case)
• C,  : positive constants
• Gamma correction

70
Contrast Stretching

To increase the dynamic range of the gray levels in the image is being
processed.

• The locations of r1, s1 and r2, s2 control the shape of the transformation
function.
• If r1= s1 and r2= s2 the transformation is a linear function and produces no
changes.
• If r1=r2, s1=0 and s2=L-1, the transformation becomes a thresholding
function that creates a binary image.
• Intermediate values of r1, s1 and r2, s2 produce various degrees of spread
in the gray levels of the output image, thus affecting its contrast.
• Commonly, r1≤r2 and s1≤s2 is assumed.

71
Histogram Processing
The histogram of a digital image by gray levels from 0 to L-1 is a discrete
function h(rk)=nk, anywhere:

a) rk is the kth gray level

b) nk is the # pixels in the image by that gray level
c) n is the total number of pixels in the image
d) k = 0, 1, 2, …, L-1
e) Normalized histogram: p(rk)=nk/n
f) sum of all constituents = 1

Histogram Equalization

k nj k
sk  T (rk )     pr (rj )
j 0 n j 0

Histogram equalization results are alike to contrast stretching

however offer the advantage of full automation, since HE automatically
determines a transformation function to yield a new image with a uniform
histogram.

72
Histogram Matching (or Specification)

• Histogram equalization does not allow collaborating image enhancement

and generates only one result: an approximation to a constant histogram.

• Sometimes however, we need to be able to specify particular histogram

forms capable of highlighting certain gray-level ranges.

Frequency Domain Methods

• We simply compute the Fourier transform of the image to be enhanced,
multiply the result by a filter and take the reverse transform to produce the
enhanced image.
• Low pass filtering involves the rejection of the high frequency components
in the image. It outcomes in blurring of the image

73
 PRACTICAL APPROACH:

MATLAB PROGRAM IMAGE INTENSITY TRANSFORMATION

clc;

clear all;

close all;

%Read Input Image

a=imread('C:\Users\natheem\Desktop\images\mri.jpg');

subplot(2,2,1);

imshow(a);

title('original image');

%negative image

b=255-a;

subplot(2,2,2);

imshow(b);

74
title('negative image');

%log transform

l=255;

d=l/log10(1+l);

e=d*log10(1+double(a));

f=uint8(e);

subplot(2,2,3);

imshow(f);

title('log transform');

%power law transform

gamma=1.1;

g=double(a).^gamma;

subplot(2,2,4);

imshow(uint8(g));

title('power law transform');

75
RESULT

Fig.2.1 Intensity Transformation using Subplots

76
MATLAB PROGRAM USING INDIVIDUAL FIGURE

clc;

clear all;

close all;

%Read Input Image

a=imread('C:\Users\natheem\Desktop\images\mri.jpg');

figure(1);

imshow(a);

title('original image');

%negative image

b=255-a;

figure(2);

imshow(b);

title('negative image');

%log transform

l=255;

d=l/log10(1+l);

77
e=d*log10(1+double(a));

f=uint8(e);

figure(3);

imshow(f);

title('log transform');

%power law transform

gamma=1.1;

g=double(a).^gamma;

figure(4);

imshow(uint8(g));

title('power law transform');

78
RESULT

Fig.2.1.1.Original Image

Fig.2.1.2.Negative Image

79
 Fig.2.1.3.Log Transformation

Fig.2.1.4. Power Law transformation

80
OUTPUT FOR X-RAY IMAGE USING SUBPLOT

a) Original Image
b) Negative Image
c) Log transformation
d) Power law transformation

81
OUTPUT FOR X-RAY IMAGE USING SEPARATE FIGURE

Fig.2.2.1 Original image

Fig.2.2.2 Negative Image

82
 Fig.2.2.3 Log Transformation

Fig.2.2.4 Power law Transformation

83
 CHAPTER 4.

HISTOGRAM EQUALIZATION

INTRODUCTION:

This application note defines a technique of Digital imaging processing

which allows medical images to have better contrast. This is achieved via the
histogram of the image, making use of a technique that allows the places with
minimal contrast to improvement increasing contrast by circulation out the most
regular intensity values. For example, in digital x-rays in what kind of colors
achieved are a palette of whites and blacks, unlike types of colors give the
physician an idea of the type of density that he or she is perceiving. Therefore
white components are most likely to indicate bone or water and black structures
mean air. Whenever pathologies are existing in an image, attempting to define
the region of the lesion or object of interest may be a challenge, because a variety
of structures are usually layered one over the other. like in the case of the chest
the heart, lungs, and blood vessels are so near with each other that contrast is
critical for accomplishing an accurate detection of problems.

84
Digital Image

A digital image is a binary symbol of a 2D image. The digital

exemplification is an array of picture elements called pixels. Each of these pixels
has a numerical value which in homochromatic images represents a grey level

Histogram of an image

Histogram is the approximation of the probability distribution of a

particular type of data. An image histogram is a type of histogram which offers a
graphical representation of the tonal distribution of the gray values in a digital
image. By viewing the image’s histogram, we can analyze the frequency of
appearance of the different gray levels contained in the image. In we can see an
image and its histogram.

85
 The histogram displays us that the image contains only a fraction of the
entire range of gray levels. In this case there are 256 gray levels and the
image only has values Between 50 to 100. Therefore this image has low
contrast.

What is a good Histogram?

A good histogram is that which covers all the expected values in the gray
scale used. This type of histogram proposes that the image has good
contrast and that particulars in the image may be observed more easily.

86
Methods for histogram equalization

• Histogram expansion
• Local area histogram equalization (LAHE)
• Cumulative histogram equalization
• Par sectioning
• Odd sectioning

87
Histogram Processing

The histogram of a digital image with gray levels from 0 to L-1 is a discrete
function h(rk)=nk, wherever:

a) rk is the kth gray level

b) nk is the # pixels in the image with that gray level
c) n is the total number of pixels in the image
d) k = 0, 1, 2, …, L-1
e) Normalized histogram: p(rk)=nk/n
f) sum of all constituents = 1

Histogram Equalization
k nj k
sk  T (rk )     pr (rj )
j 0 n j 0

Histogram equalization output are alike to contrast stretching however

offer the advantage of full automation, meanwhile HE automatically
determines a transformation function to produce a new image with a
uniform histogram.

88
Fig.1 Dark Image vs High contrast histogram image

89
Fig.2. Histogram Equalized image

90
Histogram Matching (or Specification)

• Histogram equalization does not allow cooperating image enhancement and

makes only one result: an approximation to a constant histogram.

• Sometimes still, we need to be able to specify particular histogram forms

capable of highlighting certain gray-level ranges.

91
 PRACTICAL APPROACH

FLOW CHART

92
MATLAB SOURCE CODE FOR HISTOGRAM EQUALIZATION

clc;

clear all;

close all;

%Read an input image

a=imread('C:\Users\nadeem\Desktop\palace.png');

%histogram processing

subplot(3,2,1);

imshow(a);

title('original image');

%converting to gray

subplot(3,2,3);

b=rgb2gray(a);

imshow(b);

title('grayscale image');

93
%equalisation of image

subplot(3,2,4);

c=histeq(b);

imshow(c);

title('equalised image');

subplot(3,2,5);

imhist(b);

title('histogram of grayscale image');

subplot(3,2,6);

imhist(c);

title('histogram of equalised image');

94
OUTPUT

Fig.1. Histogram Equalization with subplots

95
Fig.1 Original Image (Low contrast | Dark Image)

Fig.2 Corresponding Grayscale Image

96
Fig.4 Corresponding Histogram plot of grayscale image

Fig.3 Histogram Equalized image

97
Fig.5 Corresponding Histogram plot of Equalized image

98
HISTOGRAM EQUALIZATION USING COLOR IMAGE

clc;

clear all;

close all;

%Read Color Image

a=imread('C:\Users\natheem\Desktop\images\monky.jpg');

%histogram processing

subplot(3,3,1);

imshow(a);

title('Original image');

%converting to gray

subplot(3,3,2);

b=rgb2gray(a);

imshow(b);

title('Grayscale image');

%equalisation of image

99
 subplot(3,3,3);;

c=histeq(b);

imshow(c);

title('Equalised image');

subplot(3,3,5);

imhist(b);

title('Histogram of grayscale image');

subplot(3,3,6);

imhist(c);

title('Histogram of equalised image');

RESULT

100
OUTPUT WITH SEPARATE FIGURE

Fig.1. original Image

Fig.2.Grayscale image

101
 Fig.3.Histogram equalized image

Fig.4.Histogram plot of grayscale image

102
Fig.5. Histogram plot of Equalized image

103
CONCLUSION
Histogram equalization is a direct image processing technique often used
to achieve better quality images in black and white color balances in medical
applications such as X-rays, MRIs, and CT scans. All these images require high
definition and contrast of colors to determine the pathology that is being
experiential and reach a diagnosis. Though, in some type of images histogram
equalization can show noise hidden in the image after the processing is done.

104
 CHAPTER 5.

SPATIAL INTENSITY RESOLUTION

INTRODUCTION

Sampling is the principal factor Estimation of the spatial resolution of an

image. Commonly spatial resolution is the smallest perceptible detail in an
image, a widely used meaning of resolution is simply the smallest number of
discernible line pairs per unit distance; for estimation 100 line pairs/mm.

Gray level resolution: This refers to the smallest visible change in gray level.
The measurement of visible changes in gray level is a extremely subjective
procedure.

We have significant discretion concerning the number of Samples used to

generate a digital image. But this is not true for the amount of gray levels. Due
to hardware restraints, the number of gray levels is usually an integer power of
two. The most common value is 8 bits. It can vary depending on application.
When an actual portion of physical resolution relating pixels and level of detail
they resolution in the original scene are not necessary, it is not rare to refer to an
L-level digital image of size as consuming a spatial resolution of pixels and a
gray level resolution of L levels.

105
Functional Representation of Images

• Two-D function f(x,y), (x,y) pixel position. Positive and bounded

• Written as f(x,y)=i(x,y)r(x,y), i(x,y) illumination from light source, r(x,y)
reflectance (bounded between 0 and 1) based on material properties. E.g
r(x,y)=0.01 for black velvet, r(x,y) = 0.93 for snow.
• Intensity of monochrome image f(x,y) is synonymous with grey levels. By
convention grey level are from 0 to L-1.

106
 •

Spatial and Gray Level Resolution

• Spatial resolution is the lowest level of detail discernable in an image.

Number of line sets per millimeter, approximately 100 line pairs per
millimeter.
• Gray-level resolution is the lowest discernable change in gray level. Very
Subjective.

107
108
109
 Image Enhancement in Spatial Domain Find gray level transformation
function T(r) to obtains (x,y) =T(f(x,y)) processed image from input
image.

REASONS

1. Contrast enhancement

2. Image understanding

3. Visual improvement

Picture Reference:

Photos Reference from Digital Image Processing, Gonzalez and Woods,

Copyright 2002

110
MATLAB SOURCE CODE FOR SPATIAL INTENSITY RESOLUTION

clc;

clear all;

close all;

a=imread('C:\Users\natheem\Desktop\images\satelite.jpg');

subplot(3,2,1);

imshow(a);

title('original image');

subplot(3,2,2);

imshow(grayslice(a,128),gray(128));

title('128 graylevel image');

subplot(3,2,3);

imshow(grayslice(a,64),gray(64));

title('64 graylevel image');

subplot(3,2,4);

imshow(grayslice(a,32),gray(32));

111
 title('32 graylevel image');

subplot(3,2,5);

imshow(grayslice(a,16),gray(16));

title('16 graylevel image');

subplot(3,2,6);

imshow(grayslice(a,8),gray(8));

title('8 graylevel image');

Result

112
MATLAB PROGRAM USING INDIVIDUAL FIGURE

clc;

clear all;

close all;

a=imread('C:\Users\natheem\Desktop\images\satelite.jpg');

figure(1);

imshow(a);

title('original image');

%128 graylevel image

figure(2);

imshow(grayslice(a,128),gray(128));

title('128 graylevel image');

%64 graylevel image

figure(3);

imshow(grayslice(a,64),gray(64));

title('64 graylevel image');

%32 graylevel image

figure(4);

113
imshow(grayslice(a,32),gray(32));

title('32 graylevel image');

%16 graylevel imag

figure(5);

imshow(grayslice(a,16),gray(16));

title('16 graylevel image');

%8 graylevel image

figure(6);

imshow(grayslice(a,8),gray(8));

title('8 graylevel image');

114
Output

115
116
117
118
 CHAPTER 6.

ENHANCEMENT IN SPATIAL FILTER

Introduction

• Filter term in “Digital image processing” is stated to the subimage, filtering

is a technique for adapting or enhancing an image.
• There are others term to call subimage such as kernel, mask, template, or
window
• The value in a filter subimage are stated as coefficients, rather than pixels.

Basics of Spatial Filtering

Spatial filtering term is the filtering operations that are performed

directly on the pixels of an image, Spatial domain operation or filtering (the
processed value for the present pixel processed value for the present pixel
be contingent on both itself and nearby pixels). Henceforth Filtering is a
region operation, the value of any assumed pixel in the output image is
determined by applying some procedure to the values of the pixels in the
neighborhood of the consistent input pixel. A pixel's area is some set of
pixels, defined by their positions relative to that pixel. The concept of
filtering has its roots in the use of the Fourier transform for signal
processing in the so-called frequency domain.

119
Mechanics of spatial filtering

• The process contains simply of moving the filter mask from point to point
in an image.
• At each point (x,y) the response of the filter at that point is calculated using
a predefined relationship

Linear spatial filtering

Generally, linear filtering of an image f of size MxN with a filter mask of

size mxn is given by the expression

a b
g ( x, y)    w(s, t ) f ( x  s, y  t )
s   at   b

• The result is the sum of products of the mask coefficients with the
corresponding pixels directly under the mask
• The coefficient w(0,0) coincides with image value f(x,y), indicating that
the mask is centered at (x,y) when the computation of sum of products
takes place.

• For a mask of size mxn, we assume that m-2a+1 and n=2b+1, where a and
b are nonnegative integer. Then m and n are odd.

• The process of linear filtering similar to a frequency domain concept called

“convolution”

120
Nonlinear spatial filtering

• Nonlinear spatial filters also operate on neighborhoods, and the

mechanism of sliding a mask previous an image are the same as was just
outlined.
• The filtering operation is based conditionally on the values of the pixels
in the neighborhood under consideration

Smoothing Spatial Filters

• Smoothing filters are used for blurring and for noise reduction.
• Blurring is used in preprocessing stages, such as removal of small details
from an image prior to object extraction, and bridging of small gaps in
lines or curves
• Noise reduction can be accomplished by blurring

Type of smoothing filtering

There are two types of smoothing spatial filters

1. Smoothing Linear Filters

2. Order-Statistics Filters

121
Smoothing Linear Filters

• Linear spatial filter is just the average of the pixels contained in the
neighborhood of the filter mask.
• Sometimes it is called “averaging filters”.
• The idea is substituting the value of every pixel in an image by the
average of the gray levels in the neighborhood defined by the filter
mask.
• The general implementation for filtering an MxN image with a
weighted averaging filter of size mxn is given by the expression
•

a b

  w(s, t ) f ( x  s, y  t )
g ( x, y )  s   at   b
a b

  w(s, t )
s   at   b

Two 3x3 Smoothing Linear Filters

122
 5x5 Smoothing Linear Filters

Result of Smoothing Linear Filters

123
Order-Statistics Filters

Order-statistics filters are nonlinear spatial filters whose response is based

on ordering or ranking the pixels contained in the image area included by
the filter, and then swapping the value of the center pixel with the value
determined by the ranking result.

• Best-known as “median filter”

Process of Median filter

• Corp region of neighborhood

• Sort the values of the pixel in our region
• In the MxN mask the median is MxN div 2 +1

124
Result of median filter

Sharpening Spatial Filters

• The principal objective of sharpening is to highpoint fine detail in an

image or to enhance feature that has been blurred, either in error or as
a natural effect of a specific method of image acquisition.
• The image blurring is accomplished in the spatial domain by pixel
averaging in a neighborhood.
• Since averaging is similar to integration.
• Sharpening could be accomplished by spatial differentiation.

Definition for a first derivative

• A basic definition of the first-order derivative of a one-dimensional

function f(x) is

f
 f ( x  1)  f ( x)
• x
• Must be zero in flat segments
• Must be nonzero at the onset of a gray-level step or ramp; and
• Must be nonzero along ramps.

125
Definition for a second derivative

We define a second-order derivative as the difference

2 f
 f ( x  1)  f ( x  1)  2 f ( x).
x 2

• Must be zero in flat areas;

• Must be nonzero at the onset and end of a gray-level step or ramp;

• Must be zero along ramps of constant slope

Analyze

1st make thick edge and

2nd make thin edge

• The 1st-order derivative is nonzero along the whole ramp, while the 2nd-
order derivative is nonzero only at the start and end of the ramp.

• The reply at and around the point is much stronger for the 2nd- than for
the 1st-order derivative

126
MATLAB SOURCE CODE FOR IMAGE ENHANCEMENT IN SPATIAL
FILTERING

clc;

clear all;

close all;

a=imread('C:\Users\natheem\Desktop\images\mri.jpg');

b=double(a)+50;

subplot(3,3,1);

imshow(a);

title('original image');

subplot(3,3,2);

imshow(uint8(b));

title('enhanced image');

b1=double(a-70);

subplot(3,3,3);

imshow(uint8(b1));

title('brightness suppressed image');

e=a*.5;

f=a*.20;

127
subplot(3,3,4);

imshow(e);

title('increased in contrast');

subplot(3,3,5);

imshow(f);

title('decreased in contrast');

h1=1/9*ones(3,3);

h2=1/25*ones(5,5);

b1=conv2(a,h1,'same');

b2=conv2(a,h2,'same');

subplot(3,3,6);

imshow(uint8(b1));

title('output usins 3*3 mask');

subplot(3,3,7);

imshow(uint8(b2));

title('output using 5*5 mask');

128
Result

129
MATLAB Source Code for using separate figure

130
131
132
133
 CHAPTER 7.

ENHANCEMENT IN FREQUENCY FILTER

Enhancement in Frequency Domain Filtering

Spatial Domain

Frequency Domain

134
Major filter categories

Naturally, filters are categorized by examining their properties in the

frequency domain:

1) Low-pass
2) High-pass
3) Band-pass
4) Band-stop

Example

135
Low-pass filters (Smoothing Filters):

➢ Low-pass filters also known as called Smoothing Filters

➢ Preserve low frequencies - useful for noise suppression

frequency time
domain domai

Example:

High-pass filters (Sharpening Filters)

➢ High-pass filters also known as called Sharpening Filters

➢ Preserves high frequencies - useful for edge detection

136
Example

Band-Stop filters

137
Frequency Domain Methods

138
Case 1: h(u,v) is stated in the frequency domain.

Case 2: h(x,y) is stated in the spatial domain.

Frequency domain filtering:

STEPS

F(u,v) = R(u,v) + jI(u,v)

G(u,v)= F(u,v)H(u,v) = H(u,v) R(u,v) + jH(u,v)I(u,v)

139
140
Types of Low Pass (LP) Filters

1. Ideal low-pass filter (ILPF)

2. Butterworth low-pass filter (BLPF)
3. Gaussian low-pass filter (GLPF)

Types of High Pass (LP) Filters

1. Ideal high-pass filter (IHPF)

2. Butterworth high-pass filter (BHPF)

3. Gaussian high-pass filter (GHPF)

4. Difference of Gaussians

5. Unsharp Masking and High Boost filtering

141
MATLAB SOURCE CODE FOR IMAGE ENHANCEMENT IN FREQUENCY FILTERING

%LOW PASS FILTER

clc;

clear all;

close all;

a=imread('coins.png');

[m,n]=size(a);

mask=zeros(m,n);

for i=150:180

for j=210:240

mask (i,j)=i;

end

end

c=fftshift(mask);

b=fft2(a);

d=b.*c;

e=abs(ifft2(b));

figure(1);

subplot(3,3,1);

142
imshow(a);

title('original image');

subplot(3,3,2);

imshow(uint8(e));

title('low passs filtered image');

subplot(3,3,3);

imshow(mask);

subplot(3,3,4);

imshow(c);

title('mask after fftshift ooperation');

%HIGH PASS FILTER

a=imread('coins.png');

[m,n]=size(a);

mask=ones(m,n);

for i=150:180

for j=210:240

mask (i,j)=0;

end

end

c=fftshift(mask);

143
b=fft2(a);

d=b.*c;

e=abs(ifft2(b));

subplot(3,3,5);

imshow(a);

title('original image');

subplot(3,3,6);

imshow(uint8(e));

title('high pass filtered image');

subplot(3,3,7);

imshow(mask);

subplot(3,3,8);

imshow(c);

title('high pass filter mask');

%BAND PASS PASS FILTER

a=imread('coins.png');

[m,n]=size(a);

mask=zeros(m,n);

for i=160:170

for j=260:230

144
 mask (i,j)=0;

end

end

c=fftshift(mask);

b=fft2(a);

d=b.*c;

e=abs(ifft2(b));

figure(2);

subplot(2,2,1);

imshow(a);

title('original image');

subplot(2,2,2);

imshow(uint8(e));

title('band pass filtered image');

subplot(2,2,3);

imshow(uint8(mask));

subplot(2,2,4);

imshow(uint8(c));

title('mask after fftshift ooperation');

% BAND REJECT FILTER

145
a=imread('coins.png');

[m,n]=size(a);

mask=ones(m,n);

for i=150:180

for j=210:240

mask(i,j)=0;

end

end

c=fftshift(mask);

b=fft2(a);

d=b.*c;

e=abs(ifft2(b));

figure(3);

subplot(2,2,1);

imshow(a);

title('original image');

subplot(2,2,2);

imshow(uint8 (e));

title('band reject pass filtered image');

subplot(2,2,3);

146
 imshow(mask);

subplot(2,2,4);

imshow(c);

title('high pass filter mask');

Result

147
148
 CHAPTER 8.

COLOR IMAGE PROCESSING

Why use color in image processing?

Color is a powerful descriptor.

➢ Object identification and extraction

➢ Example: Face detection using skin colors

➢ Humans can discern thousands of color shades and intensities

➢ Human discern only two dozen shades of grays

Types of color image processing

1. Pseudo-color processing

2. Full color processing

149
 Full color processing

Images are acquired from full-color sensor or equipment

Pseudo-color processing

• In the past period, color sensors and processing hardware are not
available
• Colors are assigned to a range of monochrome intensities
•

Color fundamentals

1) Color models
2) Pseudo-color image processing
3) Color transformations
4) Smoothing and sharpening

1) COLOR FUNDAMENTALS

Physical phenomenon

Physical nature of color is identified

Psysio-psychological phenomenon

How human brain see and understand color?

150
Visible light

Chromatic light distance the electromagnetic spectrum (EM) from 400

to 700 nm

151
 The color that human perceive in an object = the light reflected from
the object

Radiance:

Total amount of energy that flow from the light source, measured in
watts (W)

Luminance:
Amount of energy an observer perceives from a light source,
measured in lumens far infrared light: high radiance, but 0 luminance

Brightness:
Subjective descriptor that is hard to measure, alike to the achromatic
view of intensity

• 6~7M Cones are the sensors in the eye

• 3 principal sensing categories in eyes
• Red light 65%, green light 33%, and blue light 2%

152
In 1931, CIE (International Commission on Illumination) defines
specific wavelength values to the

153
Primary colors

• B = 435.8 nm, G = 546.1 nm, R = 700 nm

• However, we know that no single color may be called red, green, or
blue
Secondary colors:
✓ G+B=Cyan,
✓ R+G=Yellow,
✓ R+B=Magenta

154
 Color TV

2) COLOR MODELS

Color model, color space, color system

a. Specify colors in a standard way
b. A coordinate system that each color is represented by a single point

155
RGB color model

Pixel depth:
The number of bits used to represent each pixel in RGB space
Full-color image:
24-bit RGB color image
(R, G, B) = (8 bits, 8 bits, 8 bits)

156
CMY model (+Black = CMYK)

✓ CMY: secondary colors of light, or primary colors of pigments

✓ Used to produce hardcopy output

HSI color model

✓ Will you describe a color using its R, G, B components?

✓ Human describe a color by its hue, saturation, and brightness

1. Hue : color attribute

2. Saturation: purity of color (white->0, primary color->1)

3. Brightness: achromatic notion of intensity

Fig. RGB -> HSI model

157
Pseudo-color image processing

• Assign colors to gray values based on a specified criterion

• For human visualization and understanding of gray-scale events
• Intensity slicing
• Gray level to color conversions

Application 1

Application 2

158
Application 3

159
Gray level to color transformation

Intensity slicing: piecewise linear transformation

General Gray level to color transformation

160
3) Color transformation

• Similar to gray scale transformation

g(x,y)=T[f(x,y)]

• Color transformation

RGB CMY (K)  HSI

➢ Theoretically, any transformation can be performed in any color model

➢ Practically, some operations are better suitable to specific color model

Example:

g(x,y)=k f(x,y), 0<k<1

1) HSI color space

• Intensity: s3 = k r3

• Note: transform to HSI requires complex operations

2) RGB color space

• For each R,G,B component: si = k ri

161
3) CMY color space

• For each C,M,Y component:

• si = k ri +(1-k)

162
Color image smoothing

Neighborhood processing

163
MATLAB SOURCE CODE FOR COLOR IMAGE PROCESSING

clc;

clear all;

close all;

a=imread('peppers.png');

subplot(3,3,1);

imshow(a);

title('original image');

g=rgb2gray(a);

subplot(3,3,2);

imshow(g);

title('gray conversion');

ntsc_img=rgb2ntsc(a);

subplot(3,3,3);

imshow(ntsc_img);

title('ntsc colour space conversion');

ycbcr_img=rgb2ycbcr(a);

subplot(3,3,4);

imshow(ycbcr_img);

164
title('ycbcr colour space conversion');

hsv_img=rgb2hsv(a);

subplot(3,3,5);

imshow(hsv_img);

title('hsv colour space conversion');

cmy_img=imcomplement(a);

subplot(3,3,6);

imshow(cmy_img);

title('cmy colour space conversion');

rgb=im2double(a);

r=rgb(:,:,1);

g=rgb(:,:,2);

b=rgb(:,:,3);

num=.5*((r-g)+(r-b));

den=sqrt(r-g).^2+(r-g).*(g-b);

theta=acos(num./(den +eps));

H=theta;

H(b>g)=2*pi-H(b>g);

H=H/(2*pi);

num=min(min(r,g),b);

165
den=r+g+b;

den(den==0)=eps;

S=1-3.*num./den;

H(S==0)=0;

I=(r+g+b)/3;

hsi=cat(3,H,S,I);

subplot(3,3,7);

imshow(hsi);

title('HSI');

figure(2);

subplot(2,3,1);

imshow(r);

title('red component');

subplot(2,3,2);

imshow(g);

title('green component');

subplot(2,3,3);

imshow(b);

title('blue component');

166
subplot(2,3,4);

imshow(H);

title('hue component');

subplot(2,3,5);

imshow(S);

title('saturation component');

subplot(2,3,6);

imshow(I);

title('brightness component');

167
OUTPUT

168
MATLAB PROGRAM USING INDIVIDUAL FIGURE

clc;

clear all;

close all;

a=imread('peppers.png');

figure(1);

imshow(a);

title('original image');

g=rgb2gray(a);

figure(2);

imshow(g);

title('gray conversion');

ntsc_img=rgb2ntsc(a);

figure(3);

imshow(ntsc_img);

title('ntsc colour space conversion');

ycbcr_img=rgb2ycbcr(a);

figure(4);

imshow(ycbcr_img);

169
title('ycbcr colour space conversion');

hsv_img=rgb2hsv(a);

figure(5);

imshow(hsv_img);

title('hsv colour space conversion');

cmy_img=imcomplement(a);

figure(6);

imshow(cmy_img);

title('cmy colour space conversion');

rgb=im2double(a);

r=rgb(:,:,1);

g=rgb(:,:,2);

b=rgb(:,:,3);

num=.5*((r-g)+(r-b));

den=sqrt(r-g).^2+(r-g).*(g-b);

theta=acos(num./(den +eps));

H=theta;

H(b>g)=2*pi-H(b>g);

H=H/(2*pi);

num=min(min(r,g),b);

170
den=r+g+b;

den(den==0)=eps;

S=1-3.*num./den;

H(S==0)=0;

I=(r+g+b)/3;

hsi=cat(3,H,S,I);

figure(7);

imshow(hsi);

title('HSI');

figure(8);

subplot(2,3,1);

imshow(r);

title('red component');

figure(9);

imshow(g);

title('green component');

figure(10);

imshow(b);

title('blue component');

171
figure(11);

imshow(H);

title('hue component');

figure(12);

imshow(S);

title('saturation component');

figure(13);

imshow(I);

title('brightness component');

172
OUTPUT

173
174
175
176
177
178
179
 CHAPTER 9.

DFT (DISCRETE FOURIER TRANSFORM) ANALYSIS

Fourier Transform

The Fourier Transform is a significant image processing tool which is

used to decompose an image into its sine and cosine components. The output
of the conversion represents the image in the Fourier or frequency domain,
though the input image is the spatial domain correspondent. Fourier domain
image, each point signifies a particular frequency contained in the spatial
domain image.

The Fourier Transform is used in a wide variety of applications, such as

image filtering, image rebuilding, image compression and Image investigation

Periodic Signals

A continuous-time signal x(t) is periodic if: x(t + T) = x(t)

Fundamental period,

T0, of x(t) is smallest T satisfying above equation.

Fundamental frequency:

f0 = 1/T0

Fundamental angular frequency:

ω0 = 2π/T0 = 2πf0

180
Periodicity in Biology and Medicine

181
Fourier analysis

Fourier series

Expansion of continuous function into weighted sum of sins and

cosines, or weighted sum of complex exponentials.

Fourier Transform

Maps one function to another: continuous-to-continuous

mapping, an integral transform.

Discrete Fourier Transform (DFT)

Approximation to Fourier integral. Maps discrete vector to

another discrete vector. Can be viewed as a matrix operator.

182
Fast Fourier Transform (FFT)

Special computational algorithm for DFT.

Discrete Fourier Transform (DFT):

• (u, v) are the frequency coordinates while (x, y) are the spatial
coordinates

• M, N are the number of spatial pixels along the x, y coordinates

183
Fast Fourier Transform (FFT)

• The FFT is a computationally effectual algorithm to compute the

Discrete Fourier Transform and its inverse.
• Assessing the sum above directly would take O(N2) arithmetic
operations.
• The FFT algorithm decreases the computational burden to O(N log
N) arithmetic operations.
• FFT needs the number of data points to be a power of 2 (usually 0
padding is used to make this true)
• FFT needs evenly-spaced time series

184
DFT in MATLAB

Let fbe a 2D image with dimension [M,N], then its 2D DFT can be
computed as follows:
Df = fft2(f,M,N);
• fft2 puts the zero-frequency component at the top-left corner.

• fftshift Shifts the zero-frequency component to the center. (Useful for

visualization.)
MATLAB Program
f = imread(‘saturn.tif’); f = double(f);

Df = fft2(f,size(f,1), size(f,2));

figure; imshow(log(abs(Df)),[ ]);

Df2 = fftshift(Df);

figure; imshow(log(abs(Df2)),[ ]);

Result

185
Examine the Fourier transform of a synthetic image

f = ones(10,20);

F = fft2(f, 500,500);

f1 = zeros(500,500);

f1(240:260,230:270) = 1;

subplot(2,2,1);

imshow(f1,[]);

S = abs(F);

subplot(2,2,2);

imshow(S,[]);

Fc = fftshift(F);

S1 = abs(Fc);

186
 subplot(2,2,3);

imshow(S1,[]);

S2 = log(1+S1);

subplot(2,2,4);
imshow(S2,[]);

Result

187
Fourier transform of natural images

f = imread(‘lenna.jpg’);

subplot(1,2,1);

imshow(f);

f = double(f);

F = fft2(f);

Fc = fftshift(F);

S = log(1+abs(Fc));

Subplot(1,2,2);

imshow(S,[]);

188
Result

a. Original image

b. corresponding DFT image

189
MATLAB SOURCE CODE FOR DFT ANALYSIS

clc;

clear all;

close all;

a=zeros(256);

[m,n]=size(a);

for i=120:145

for j=120:145

a(i,j)=225;

end;

end;

b=imrotate(a,45,'bilinear','crop');

a1=log(1+abs(fftshift(fft2(a))));

b1=log(1+abs(fftshift(fft2(b))));

subplot(2,2,1);

imshow(a);

title('Orignal Image');

subplot(2,2,2);

imshow(b);

title('Rotate Image');

190
 subplot(2,2,3);

imshow(mat2gray(a1));

title('orignal image spectrum');

subplot(2,2,4);

imshow(mat2gray(b1));

title('spectrum of rotate img');

RESULT

191
 CHAPTER 10.

BASIC THRESHOLDING FUNCTION

Thresholding

Thresholding is finding histogram of gray level intensity.

Types

a) Basic Global Thresholding

b) Multiple Threshold
c) Variable Thresholding
d) Otsu’s Method

Basic Global Thresholding

• Initially Segment image use:

g ( x, y)   01 if f(x,y)T
if f(x,y)T

• Compute the average intensity m1 and m2 for the pixels

192
• Compute a new threshold:

1
T (m 1  m 2 )
2

• Until the difference between values of T is smaller than a predefined

parameter.

193
Otsu’s Method

• Based on a very simple idea: Find the threshold that minimizes the
weighted within-class variance.
• This turns out to be the same as maximizing the between-class
variance.
• Operates directly on the gray level histogram.

Assumptions
• Histogram (and the image) are bimodal.
• No usage of spatial coherence, nor any other idea of object structure.
• Assumes stationary statistics, but can be modified to be locally
adaptive. (exercises)
• Assumes uniform illumination (indirectly), so the bimodal
illumination behavior arises from object appearance changes only.

194
Multiple Threshold

As Otsu’s method, it takes more area and k*

Disadvantage:

It becomes too complex when number of area more than two or three.

195
Variable Thresholding

• Image partitioning

• It is work when the objects of interest and the background occupy

regions of reasonably comparable size. If not, it will fail.
• Variable thresholding based on local image properties
• Using moving average
• It discussed is based on computing a moving average along scan lines
of an image.

196
197
MATLAB SOURCE CODE FOR BASIC THRESHOLDING
FUNCTION

clc;

clear all;

close all;

i=imread('cell.tif');

subplot(2,2,1);

imshow(i);

title('original image');

subplot(2,2,2);

im2bw(i,0.35);

title('mannual thershold image');

[counts,x]=imhist(i);

p=polyfit(x,counts,6);

y=polyval(p,x);

[v,ind]=sort(abs(diff(y)));

thresh=ind(3)./255;

198
subplot(2,2,3);

im2bw(i,thresh);

title(' polynomial thershold image');

level=graythresh(i);

subplot(2,2,4);

im2bw(i,level);

title('otsu method');

figure;

plot(x,counts);

hold on,plot(x,y,'r');

title('graph');

199
RESULT

200
MATLAB Code for Binary Thresholding Function & Histogram

clc;

clear all;

close all;

count=0;

a=imread('cell.tif');

t=mean2(a);

done=false;

while done

count=count+1;

g=f>t;

tnext=0.5*(mean(f(g)));

mean(a(~g));

done=abs(t-tnext)<0.5;

t=tnext;

end;

count=2;

t=125.3860;

g=im2bw(a,t/255);

201
 subplot(2,2,1);

imshow(a);

title('original image');

subplot(2,2,2);

imhist(a);

title('histogram image');

subplot(2,2,3);

imshow(g);

title('binary image');

OUTPUT

202
 CHAPTER 11.

SAMPLING AND QUANTIZATION

Image representation

An image is a function of the space.

Typically, a 2-D projection of the 3-D space is used, but the image can exist
in the 3-D space directly.

203
Digitalization

Digitalization of an analog signal involves two processes:

1. Sampling, and
2. Quantization.

Both processes correspond to a discretization of a quantity, but in

different fields.

Sampling

Sampling corresponds to a discretization of the space. That is, of the

domain of the function, into f : [1, . . . , N] × [1, . . . , M] −→ Rm.

204
Quantization

Quantization corresponds to a discretization of the intensity values.

That is, of the co-domain of the function.

After sampling and quantization, we get

f : [1, . . . , N] × [1, . . . , M] −→ [0, . . . , L].

205
MATLAB SOURCE CODE FOR SAMPLING AND QUANTIZATION

clc;

clear all;

close all;

a=imread('trees.tif');

f=ind2gray(a,gray(256));

f=f(1:256,1:256);

[N,M]=size(f);

m=7;

w=1/m;

F=fftshift(fft2(f));

for i=1:N

for j=1:N

r2=(i-round(N/2))^2+(j-round(N/2))^2;

if(r2>round(N/2*w)^2)F(i,j)=0;

end;

end;

end;

Id=real(ifft2(fftshift(F)));

206
Id=imresize(Id,[N/m,N/m],'nearest');

m=10;

[N,M]=size(f);

Iu=zeros(m*N,m*N);

for i=1:N

for j=1:N

Iu(m*(i-1)+1,m*(j-1)+1)=f(i,j);

end;

end;

[N,M]=size(Iu);

w=1/m;

F=fftshift(fft2(Iu));

for i=1:N

for j=1:N

r2=(i-round(N/2))^2+(j-round(N/2))^2;

if(r2>round(N/2*w)^2)F(i,j)=0;

end;

end;

end;

Iu=(m*m)*abs(ifft2(fftshift(F)));

207
figure;imshow(a);

title('original image');

figure;imshow(f);

title('gray image');

figure;imshow(Id);

title('Idown image');

figure;imshow(Iu);

title('Iup image');

208
209
210
 CHAPTER 12.

IMAGE TRANSFORMATION

Discrete Fourier Transform (DFT)

• The 2-D discrete Fourier transform of function f meant by F(u, v) is given

by F(u, v) =∑ ∑ for u=0, 1, 2… M-1 and v=0, 1, 2… N-1
• The values F(u,v) are the DFT constants of f(x,y),
• The zero-frequency coefficients, F(0,0) is often called the ‘DC
component’.
• The MATLAB functions fft, fft2 and fftn and their reverses ifft, ifft2 and
ifftn all use Fast Fourier Transform procedure to calculate DFT.
• DFT of a vector x of length n is another vector y of length n,

211
 2. Discrete Cosine Transform (DCT)

DCT signifies an image as a sum of sinusoids of variable

magnitude and frequencies. DCT has the property that, most of the
info about the image is focused in just a few coefficients of DCT, for
this aim DCT is used in compression.

Example:

DCT is at the heart of the world-wide standard lossy image

compression algorithm known as JP

212
 There are two ways to compute DCT using image processing
toolbox, first method is to use dct2 function, dct2 uses an FFT-based
algorithm for quick computation with big inputs, The 2nd method is to use
the DCT transform matrix, which is returned by the function dctmtx and
might be more effectual for small square inputs, such as 8-by-8 or 16-by-
16.

WALSH HADAMARD TRANSFORM

Walsh Hadamard transform is a non-sinusoidal orthogonal transform

method that decomposes a signal into set of basic functions. These
functions are Walsh functions which are rectangular or square waves with
+1 or -1.

Hadamard function:

• H=hadamard(n) returns hadamard matrix of order n.(n must be an integer

and n,n/12 or n/20 must be a power of 2).
• Hadamard matrices are matrices of 1’s and -1’s whose columns are
orthogonal.
• Where [n n]=size(H) and I=eye(n,n);{eye returns the identity
matrix)
• Different ordering scheme are used to store Walsh
function.That are: Sequency: coefficients in order of
increasing sequence value. Hadamard: coefficient in
hadamard order.
• Dyadic: coefficient in gray code order.

213
Application of Walsh hadamard transform:

✓ Speech processing,
✓ filtering and
✓ power spectrum analysis.

Walsh-hadamard transform has a fast version fwht (fast Walsh-hadamard

transform). Compare to fft, fwht requires less storage space and is faster to
calculate because it uses real addition and subtraction while fft requires
complex calculation. Inverse of fwht is ifwht.

Definition of IFWHT:

• Y=fwht(X) returns the coefficient of the discrete Walsh Hadamard transform.

• Y=fwht(X,n) returns the n-point discrete Walsh hadamard transform,nmust be
power of 2.X and n must be of same length.If X is longer than n,X is truncate
and if X is shorter than n,X is padded with zeros.
• Y=fwht(X,n,ordering) specifies the ordering to use for the returned Walsh-
hadamard transform coefficient.

Applications of Transforms
• feature extraction and representation
• Enhancement and Compression
• pattern recognition,e.g.,eigen faces
dimensionality reduction
• Analyse the principal (“dominating”) components

214
Radon transform

• The Radon transform of a two dimensional function, f(x,y) is the line

integral of f parallel to the y´-axis given by
• The radon function computes projections of an image matrix along
specified directions. It computes the line integral from multiple source
along parallel paths, or beams, in a certain direction. The beams are
spaced 1 pixel unit apart. To represent an image, radon function takes
multiple, parallel-beam projections of the image from different angles
by rotating the source around the center of the image

215
MATLAB SOURCE CODE FOR IMAGE TRANSFORMATION

clc;

clear all;

close all;

a=imread('mri.tif');

figure,imshow(uint8(a));

title('original image');

a=double(a);

[s1,s2]=size(a);

bs=8;

%waslh matrix

hadamardMatrix=hadamard(bs);

M=log2(bs)+1;

hadIdx=0:bs-1;

binhadIdx=fliplr(dec2bin(hadIdx,M))-'0';

binseqIdx=zeros(bs,M-1);

for k=M:-1:2

binseqIdx(:,k)=xor(binhadIdx(:,k),binhadIdx(:,k-1));

end;

216
seqIdx=binseqIdx*pow2((M-1:-1:0)');

walshMatrix=hadamardMatrix(seqIdx+1,:);

%walsh

temp_walsh=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=a((y:y+bs-1),(x:x+bs-1));

h=walshMatrix;

m=log2(bs);

t=(1/(2^m))*h*croppedImage*h';

temp_walsh((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

figure(2),subplot(1,2,1);

imshow(uint8(temp_walsh));

title('walsh transform image');

%inverse walsh

temp_inwalsh=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=temp_walsh((y:y+bs-1),(x:x+bs-1));

h=walshMatrix;
217
 m=log2(bs);

t=(1/(2^m))*h*croppedImage*h;

%t=getInwalshtransform(croppedImage,bs);

temp_inwalsh((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

subplot(1,2,2);

imshow(uint8(temp_inwalsh));

title('inverse walsh transform image');

%hadamard

temp=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=a((y:y+bs-1),(x:x+bs-1));

N=bs;

h=hadamard(N);

m=log2(N);

t=(1/(2^m))*h*croppedImage*h';

%t=getInwalshtransform(croppedImage,bs);

temp((y:y+bs-1),(x:x+bs-1))=t;

end;

end;
218
figure(3),subplot(1,2,1);

imshow(uint8(temp));

title('hadamard transform image');

%inverse hadamard

temp1=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=temp((y:y+bs-1),(x:x+bs-1));

h=hadamard(N);

m=log2(N);

t=(1/(2^m))*h'*croppedImage*h;

%t=getInvhadamardtransform(croppedImage,bs);

temp1((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

subplot(1,2,2);

imshow(uint8(temp1));

title('inverse hadamard transform image');

%dct

temp_dct=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1
219
 croppedImage=a((y:y+bs-1),(x:x+bs-1));

s=dctmtx(N);

t=s*croppedImage*s';

%t=getdcttransform(croppedImage,bs);

temp_dct((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

figure(4),subplot(1,2,1);

imshow(uint8(temp_dct));

title('dct transform image');

%inverse dct

temp_invdct=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=temp_dct((y:y+bs-1),(x:x+bs-1));

s=dctmtx(N);

t=s'*croppedImage*s;

%t=getInvdcttransform(croppedImage,bs);

temp_invdct((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

subplot(1,2,2);
220
imshow(uint8(temp_invdct));

title('inverse dct transform image');

%haar matrix

a=1/sqrt(N);

for i=1:N

H(1,i)=a;

end;

for k=1:N-1

p=fix(log2(k));

q=k-2^p+1;

t1=N/2^p;

sup=fix(q*t1);

mid=fix(sup-t1/2);

inft=fix(sup-t1);

t2=2^(p/2)*a;

for j=1:inft

H(k+1,j)=0;

end;

for j=inft+1:mid

H(k+1,j)=t2;

end;

for j=mid+1:sup
221
 h(k+1,j)=-t2;

end;

for j=sup+1:N

H(k+1,j)=0;

end;

end;

%haar

temp_haar=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=a((y:y+bs-1),(x:x+bs-1));

h=H;

m=log2(N);

t=h*croppedImage*h';

%t=gethaartransform(croppedImage,bs);

temp_haar((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

figure(5),subplot(1,2,1);

imshow(unit8(temp_haar));

title('haar transform image');

%inverse haar
222
temp_invhaar=double(zeros(size(a)));

for y=1:bs:s1-bs+1

for x=1:bs:s2-bs+1

croppedImage=temp_haar((y:y+bs-1),(x:x+bs-1));

h=H;

t=h'*croppedImage*h;

%t=getInvhaartransform(croppedImage,bs);

temp_invhaar((y:y+bs-1),(x:x+bs-1))=t;

end;

end;

subplot(1,2,2);

imshow(unit8(temp_invhaar));

title('inverse haar transform image');

223
OUTPUT FOR IMAGE TRANSFORMATION\

224
225
 APPENDIX

Top 100+ Image Processing Projects - Free Source Code and Abstracts

https://www.pantechsolutions.net/blog/category/image-video-
processing/amp/

https://www.pantechsolutions.net/blog/top-100-image-processing-projects-
free-source-code/amp/

Basic Image Processing Projects

https://electronicsforu.com/electronics-projects/software-projects-
ideas/image-processing-using-matlab

Official Image Processing References

https://in.mathworks.com/discovery/digital-image-processing.html

MATLAB Official Images Processing Toolbox Link

https://in.mathworks.com/products/image.html

Example Codes for Image Processing

https://in.mathworks.com/products/image/code-examples.html

If u don't have MATLAB software please Purchase from this link

https://in.mathworks.com/store

226
REFERENCE:

[1] Kenneth R. Castleman. Digital Image Processing. Prentice Hall, 1996.

[2] Ashley R. Clark and Colin N Eberhardt. Microscopy Techniques for Materials
Science. CRC Press, Boca Raton, Fl, 2002.

[3] Rafael Gonzalez and Richard E. Woods. Digital Image Processing. Addison-
Wesley, second edition, March, 2017.

[4] Robert M. Haralick and Linda G. Shapiro. Computer and Robot Vision. Addison-
Wesley, 1993.

[5] James D. Foley, Andries van Dam, Steven K. Feiner, John F. Hughes, and Richard
L. Phillips. Introduction to Computer Graphics. Addison-Wesley, 1994.

[6] Robert V. Hogg and Allen T. Craig. Introduction to Mathematical Statistics.

Prentice-Hall, f ifth edition, 1994.

[7] William K. Pratt. Digital Image Processing. John Wiley and Sons, second edition,
1991.

[8] Jae S. Lim. Two-Dimensional Signal and Image Processing. Prentice Hall, 1990.

[9] Majid Rabbani and Paul W. Jones. Digital Image Compression Techniques. SPIE
Optical Engineering Press, 1991.

[10] Jean Paul Serra. Image analysis and mathematical morphology. Academic Press,
1982.

[11] Steven Roman. Introduction to Coding and Information Theory. Springer-Verlag,

1997.

[12] Azriel Rosenfeld and Avinash C. Kak. Digital Picture Processing. Academic
Press, second edition, 1982.

227
[13] Melvin P. Siedband. Medical imaging systems. In John G. Webster, editor,
Medical instrumentation : application and design, pages 518–576. John Wiley and
Sons, 1998.

[14] Milan Sonka, Vaclav Hlavac, and Roger Boyle. Image Processing, Analysis and
Machine Vision. PWS Publishing, second edition, 1999.

[15] Dominic Welsh. Codes and Cryptography. Oxford University Press, 1989.

[16] Scott E. Umbaugh. Computer Vision and Image Processing: A Practical

Approach Using CVIPTools. Prentice-Hall, 1998.

[17] https://www.pantechsolutions.net/blog/category/image-video-processing/amp/

[18] https://www.pantechsolutions.net/blog/top-100-image-processing-projects-free-
source-code/amp/

[19] https://in.mathworks.com/discovery/digital-image-processing.html

228
 ABOUT THE AUTHOR

Arsath Natheem is an indian Biomedical Engineer and Youtuber who works primarily
in the field of Artificial intelligence, He is best known for his multimedia Presentation about
"How the Biomedical Engineers Save the life" at Velalar College of Engineering and
Technology in Tamilnadu, he was awarded the best project holder for IoT Based Voice
Recognition Robot and also presented his project at Adhiyamaan college of engineering and
Technology and won the first prize for his project. He participated project competition at
Madras institute of technology (MIT) in Chennai, now he pursuing final year BE Biomedical
Engineering at Velalar College of Engineering and Technology, He's field of interest is an
Artificial Intelligence it's specifically applicable for Medical Diagnosis.

229
ONE LAST THING…

If you enjoyed this book or found it useful I’d be very grateful if you’d post a short
review on Amazon. Your support really does make a difference and I read all the
reviews personally so I can get your feedback and make this book even better.

If you’d like to leave a review then all you need to do is click the review link on this
book’s page on Amazon here: http://amzn.to/2jXcHfL

Thanks again for your support!

230