Chapter 2

DIGITAL IMAGING
FUNDAMENTALS
 2

This lecture will cover:

 The human visual system.
 Image formation in the eye.
 Simultaneous contrast & optical illusion.
 Image acquisition & formation.
 Image sampling and quantization.
 Image representation .
 Spatial & intensity resolution.
 Some basic relationships between pixels.
 Mathematical tools used in image processing.
 Image transforms.
 Human Visual System
3

The best vision model we have!

Knowledge of how images form in the eye can help us
with processing digital images
 We will see the structure of the human visual system
 structure of human visual system
4

The lens focuses light from objects onto the retina

The retina is covered with
light receptors called
cones and rods.
Cones are (6-7 million)
concentrated at the central portion
of the retina and high sensetive to color.
Rods are (75-150 million)
distributed at retina surface
and are sensitive to low levels of illumination.
 Image Formation In The Eye
5

 In ordinary camera,the lens has a fixed focal

length,and foucsing at various distances is achieved
by varying the distance between the lens and the
image palne.
 But in human eye, Muscles within the eye can be
used to change the shape of the lens allowing us
focus on objects that are near or far away
 simultaneous contrast
6

 Is related to the fact that a region’s perceived

brightness does not depend on its intensity.
 Ex, All the inner squares have the same intensity,but
they appear progressively darker as the background
becomes lighter.
 Optical Illusions
7

 In which the eye fills in nonexisting information

or wrongly perceives geometrical properties of objects
 Image Acquisition
8

 to acquire a digital image

 Images are generated by the combination of
illumination source and reflection of energy by the
objects in that scene.
 A simple image formation model
9

 Images are two dimensional functions of the form

f(x,y) which mean the value or amplitude of ‘f’ at
spatial coordinates (x,y) .
 F(x,y) characterized by i(x,y) and r(x,y) as following
f(x,y)= i(x,y) r(x,y)
Where,
0<i(x,y) <∞
and 0<r(x,y) <1
 Reflectance is bounded by 0 (total absorption) and
1(total reflectance).
 Image Sampling And Quantisation
10

 The ouput of most sensors is a continuous waveform

so to create a digital image we need to do two
processes :
1. Image sampling : is to digitizing the image
coordinates.
2. Image quantization: is to digitizing the image
amplitude(intensity level).
 11

(a) Continuous image projected onto a sensor array.

(b) Result of sampling and quantization (digitized
image).

(a) (b)
 Representing digital images
12

 Image representation: to convert the input data to

a form suitable for computer processing.
13
 Dynamic range
14

 It can be defined as the ratio between maximum

measurable intensity level and minimum detectable
intensity level in an image.
 As a rule, the upper limit is
determined by the saturation and
the lower limit by noise.
 Image contrast : is the difference in
intensity between the highest and
lowest intensity level in image.
 Spatial and intensity resolution
15

 Spatial resolution is:

• a measure of the smallest detail in an image.
• line pairs or dots per unit distance(dpi).
• Note: to discriminate between two images we need to know dpi and
size of the image.

1024 256 64
 16

 intensity resolution:
• is the smallest discernible change in intensity level.
• the more intensity levels used, the finer the level of detail
discernable in an image
• Intensity level resolution is usually given in terms of the number of
bits used to store each intensity level
Number of Intensity
Number of Bits Examples
Levels
1 2 0, 1
2 4 00, 01, 10, 11
4 16 0000, 0101, 1111
8 256 00110011, 01010101
16 65,536 1010101010101010
 Ex,
17
256 grey levels (8 bits per 128 grey levels (7 bpp) 64 grey levels (6 bpp) 32 grey levels (5 bpp)
pixel)

16 grey levels (4 bpp) 8 grey levels (3 bpp) 4 grey levels (2 bpp) 2 grey levels (1 bpp)
 Some basic relationships between pixels
18

 Neighbors of pixel
A pixel ‘p’ at coordinates(x,y) has
1. four horizontal and vertical neighbors called 4-
neighbors N4(p):
(x+1,y),(x-1,y),(x,y+1),(x,y-1)

1. Four diagonal neighbors of ‘p’ called ND(p) :

(x+1,y+1),(x+1,y-1),(x-1,y+1),(x-1,y-1)

1. N4(p) and Nd(p) together are constructing N8(p).

 Adjacency
19

•Let v be the set of intensity values used to define adjacency.

•In binary image V={1} if we refer to adjacency of pixels with values 1.
•But in grey image ,the adjacency of pixels with values from 0 till 255 ,so set V
could be any subset from 0 till 255.
•Three types of adjacency :
Distance measures
20
 Mathematical tools used in digital image
processing
21

 Consider the following 2 * 2 images

 In this book we use array product,meaning that the

division is between corresponding pixels.
 Linear versus non linear operations
22

 Consider a general operator H that produce an

output g(x,y) from an input image f(x,y)

 So H is said to be linear operator if

 Where are constants and are

images of same size.
 Linear example
23

 The first step come from the fact that summtion is

distributive.so the summtion is alinear operator.
 Non linear example
24

 Consider the following two images

 Arithmetic operations
25

 Four arithmetic operation are denoted as:

 Images in arithmetic operations must be of the same

size.
 Averaging noisy images for noise reduction
26

 as
 Using multiplication in masking(ROI)
27

 Region of interest operation simply consists of :

Multiplying agiven image by a mask that has 1s in the ROI and 0s
elsewhere.
 28

 Note:
In arthmetic operations such as difference between
two 8-bit images can range from a min. Of -255 and
a max. Of 255 ,and the values of sum two images can
range from 0 to 510.
So we need to set all negative values to 0 and set to 255
all values exceed this limit.
 Set and logical operations
29

 Basic set operations

 If a is an element of A ,we write
 If a is not an element of A , we write
 The set with no elemnts called null set or empty set and is denoted
by the symbol
 If every elemnt in set A is an element in set B then A is said to be
subset of B
 The union of two sets A and B denoted as
 The set of elements belongs to either A , B or both
 Two sets A and B are said to be disjoint or mutually exclusive if
they have no common elemnts
 Complement set of A is all elemnts not in A
 Diffrence of two sets denoted as A-B
Logical operations
30
 Spatial operations
31

 Spatial operations are performed directly on the

pixels of a given image.
 Point : the output value at a specific coordinate is dependent only
on the input value at that same coordinate.
 Local : the output value at a specific coordinate is dependent on
the input values in the neighborhood of that same coordinate
 Global : the output value at a specific coordinate is dependent on
all the values in the input image
32
 Image transforms
33

 Most of image processing approches works directly

on spatial domain .but in some cases it is better to
work on transform domain and applying the inverse
transform to return to spatial domain .
 A 2-D linear transform can be expressed as:

 Where f(x,y) is the input image and r(x,y,u,v) is the

forward transformation kernel.
 34

 We can recover f(x,y) using the inverse transform

T(u,v).

 Where s(x,y,u,v) is called inverse transformation

kernel.
Probabilistic methods
35
 36

Chapter 3
 Image Enhancement
37

 is the process of manipulating images so that the

result is more suitable than the original for aspecific
application. As following:
 Highlighting interesting detail in images.
 Removing noise from images.
 Making images more visually appealing.

 Enhancement are problem oriented techniques.

 Ex, the method for enhancing X-ray images may be
not suitable for enhancing satellite images .
 Spatial Domain Image Enhancement

Most spatial domain enhancement operations can be

reduced to the form
g (x, y) = T[ f (x, y)] Origin x
where f (x, y) is the
input image, g (x, y) is
the processed image
and T is some
operator defined over (x, y)

some neighbourhood
of (x, y)

y Image f (x, y)
 Point Processing

The simplest spatial domain operations occur when

the neighbourhood is simply the pixel itself
Point processing operations take the form
s =T(r)

where s refers to the processed image pixel value and

r refers to the original image pixel value .
 Intensity Transformations
40

 Values of r lower than k are compresed by the

transformation function into a narrow range of s,
toward black. The opposite is true for values greater
than k .this is called contrast stretching.
 This transformation
Function is called
Thresholding function.
 Basic intensity transformation functions
41

 There are three basic Types of functions:

 Linear(negative and Identity transformation).
 Logarithmic (Log and invers-log tranformation).
 Power-low(nth-power and nth-root transformation).
 image negatives

The negative of the image with intensity level in the

range [0,L-1] is given by the expression:
S=L-1-r

 it is suitable for enhancing white or gray details

embeded in black regions of images.

Origina Negative
s = 1.0 - r
l Image Image
 Logarithmic transformation
43

 General form:
 s = c * log(1 + r)
The log transformation maps a narrow range of low
input grey level values into a wider range of output
values .(brights images).
The inverse log transformation performs the opposite
transformation .
 Logarithmic Transformations (cont…)
44
Log functions are particularly useful when the input
grey level values may have an extremely large range
of values
In the following example the Fourier transform of
an image is put through a log transform to reveal
more detail

s = log(1 + r)
 Power-law (gamma) transformation
45

 It has the following form:

 s=c*rγ
 Power-law curves with fracional values of map a
γ

narrow range of dark input values into a wider range

of output values, with the opposite being true for
higher values off input levels.
 Power Law Example (cont…)
46

The images to the

right show a
s = r 0.6
magnetic resonance
(MR) image of a
fractured human
spine

s = r 0.4
Different curves
highlight different
detail
 Gamma Correction
47

Many of you might be familiar with gamma correction

of computer monitors
Problem is that
display devices do
not respond linearly
to different
intensities
Can be corrected
using a log
transform
 Piecewise Linear Transformation Functions
48

 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 device .
 so the function is single
valued and monotonically increasing..
 If ,so the function is linear
that produce no change in intensity level.
 ,the function
is thresholding that creates a binary image.
 Intensity level slicing
49

Highlights a specific range of grey levels.

 To display in one value(white) all the values

in the range of interest and in another(black)

all other intensities.
 Brightens or darkens the desired range of

intenities but leaves all other intensity levels

in the image unchanged.
 Bit Plane Slicing
50

Often by isolating particular bits of the pixel values in

an image we can highlight interesting aspects of that
image
 Higher-order bits usually contain most of the significant visual
information
 Lower-order bits contain
subtle details