Digital

Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

Chapter 2: Digital Image Fundamentals

2.1 Elements of Visual Perception

2.2 Light and the Electromagnetic Spectrum
2.3 Image Sensing and Acquisition
2.4 Image Sampling and Quantization
2.5 Some Basic Relationships between Pixels
2.6 Linear and Nonlinear Operations

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1 Elements of Visual Perception

In this section, we are interested in learning

• Mechanics and parameters related to how images are formed
in the eye.
• Physical limitations of Human Vision
• Resolution
• Ability to adapt to changes in illumination

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

1
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.1 Structure of the Human Eye

• The eye is nearly a sphere

• average diameter of approximately 20 mm
• Three membranes :-
- the outer cover
+ cornea (tough, transparent tissue)
+ sclera (opaque membrane)
- the choroid (below the sclera)
+ ciliary body
+ iris diaphragm (pupil)
contracts or expands
to control the amount of light
- the retina (the innermost
membrane of the eye)
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.1 Structure of the Human Eye

• When the eye is properly focused, light from an object outside

the eye is imaged on the retina.
• There are two classes of receptors
- Cones : 6-7 million, located primarily in the fovea,
highly sensitive to color, one cone - one nerve end,
Cone vision, photopic or bright-light vision
- Rods : 75-150 million, distributed over the retina surface,
sensitive to low level of illumination,
several rods - one nerve end, scotopic or dim-light vision
Objects that appear brightly colored in daylight when seen by
moonlight appear as colorless forms because only the rods are
stimulated.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

2
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.1 Structure of the Human Eye

• Blind spot (the absence of receptors in this area)

• fovea : a circular indentation in the retina of
about 1.5 mm in diameter
• The density of cones in the fovea is approximately
150,000 elements per mm2

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.2 Image Formation in the Eye

The focal length varies from approximately 17 mm to about 14 mm,

as the refractive power of the lens increases from its minimum to its
maximum.
Object farther away than about 3 m Æ lowest refractive power
Æ focal length 17 mm
Nearby object Æ most strongly refractive power
Æ focal length 14 mm
It is eas to calculate the size of the retinal image of any object.
15/100 = h/17
h = 2.55 mm

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

3
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Subjective brightness
(intensity as perceived by
the human visual system)
is a logarithmic function
of the light intensity incident on the eye.
Photopic vision : range 106
Transition from scotopic to photopic vision :
range fom 0.001 to 0.1 millilambert -3 to -1 mL in the log scale
Visual system cannot operate over such a total range simultaneously.
Æ brightness adaptation
Brightness adaptation level (the current sensitivity level of the visual
system) at level Ba is Bb to Ba Somkiat Wangsiripitak
© 2002 R. C. Gonzalez & R. E. Woods

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Brightness discrimination : the ability of the eye to discriminate

between changes in light intensity
at any specific adaptation level
A flat, uniformly illuminated diffuser (such as opaque glass), I
A short-duration flash, ∆I
When ∆I is strong enough, the subject will give a response of “yes”
all the time.
The increment of illumination discriminable 50% of the time, ∆Ic
Weber ratio, ∆Ic/I

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

4
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Small value of ∆Ic/I :

Æ a small percentage change in intensity is discriminable
Æ good brightness discrimination
The curve in Fig.2.6 shows that brightness discrimination is poor at
low levels of illumination.
The two branches in the curve :- the function of rods and cones.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Mach bands

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

5
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Simultaneous contrast

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.1.3 Brightness Adaptation and Discrimination

Optical illusions

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

6
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.2 Light and the Electromagnetic Spectrum

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.2 Light and the Electromagnetic Spectrum

Visible band of EM : about 0.43 µm (violet) to about 0.79 µm (red)

Achromatic or monochromatic light: void of color
Æ intensity or gray level
Chromatic light
Æ radiance: total amount of energy that flows from the light
source, measured in watts (W)
Æ luminance: amount of energy an observer perceives from a light
source, measured in lumens (lm)
Æ brightness: subjective descriptor of light perception, impossible
to measure
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

7
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3 Image Sensing and Acquisition

The three principal

sensor arrangements
used to transform
illumination energy into
digital images.
Incoming energy is transformed into
a voltage.
A digital quantity is obtained from each
sensor by digitizing its response voltage.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.1 Image Acquisition Using a Single Sensor

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

8
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.2 Image Acquisition Using a Sensor Strips

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.3 Image Acquisition Using a Sensor Arrays

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

9
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.4 A Simple Image Formation Model

0 < f(x, y) < ∞ (2.3-1)

f(x, y) = i(x, y) r(x, y) (2.3-2)
illumination : 0 < i(x, y) < ∞ (2.3-3)
reflectance : 0 < r(x, y) < 1 (2.3-4)
total absorption total reflectance
When images are formed via transmission of the illumination
through a medium, such as a chest X-ray, we would deal with a
transmissivity instead of a reflectivity function.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.4 A Simple Image Formation Model

A clear day : 90,000 lm/m2 (sun, on the surface of the Earth)

A cloudy day : 10,000 lm/m2 (sun, on the surface of the Earth)
A clear evening : 0.1 lm/m2 (moon , on the surface of the Earth)
A commercial office : 1,000 lm/m2
r(x, y) for black velvet : 0.01
for stainless steel : 0.65
for flat-white wall paint : 0.80
for silver-plated metal : 0.90
for snow : 0.93

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

10
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.3.4 A Simple Image Formation Model

[Lmin, Lmax] Æ gray scale

Lmin = iminrmin
Lmax = imaxrmax
Common practice is to shift this interval numerically to the interval
[0, L-1]
0 Æ black
L-1 Æ white
All intermediate values are shades of gray varying from black to
white.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4 Image Sampling and Quantization

To create a digital image, we need to convert the continuous

sensed data into digital form.
Æ sampling and quantization

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

11
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.1 Basic Concepts in Sampling and Quantization

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.1 Basic Concepts in Sampling and Quantization

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

12
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.2 Representing Digital Images

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.2 Representing Digital Images

M x N digital image in the matrix form :

f(0, 0) f(0, 1) … f(0, N-1)
f(x, y) = f(1, 0) f(1, 1) … f(1, N-1) (2.4-1)

f(M-1, 0) f(M-1, 1) … f(M-1, N-1)

A more traditional matrix notation
a0, 0 a0, 1 … a0, N-1
A= a1, 0 a1, 1 … a1, N-1 (2.4-2)

aM-1, 0 aM-1, 1 … aM-1, N-1

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

13
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.2 Representing Digital Images

Due to processing, storage, and sampling hardware considerations,

the number of gray levels typically is an integer power of 2.
L = 2k (2.4-3)
The discrete gray levels are equally spaced and are integers in the
interval [0, L-1].
Dynamic range : the range of values spanned by the gray scale
High dynamic range : gray levels span a significant portion of
the gray scale Æ high contrast
Low dynamic range : Æ low contrast Æ dull, washed out gray look

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.2 Representing Digital Images

The number of bits required to store a digitized image :

b=M*N*k (2.4-4)
When M = N,
b = N2 k (2.4-5)
Image with 2k gray levels Æ “k-bit image”

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

14
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

Spatial resolution
Their size differences make it difficult to see the effects resulting
from a reduction in the number of samples.

The simplest way to compare these

effects is to bring all the subsampled
images up to size 1024 x 1024 by row
and column pixel replication.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

checkerboard

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

15
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

Gray-level resolution
In this example, we
keep the number of
samples constant and
reduce the number of
gray levels from 256
to 2, in integer
powers of 2.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

False contouring
Caused by the use
of an insufficient
number of gray
levels in smooth
areas of digital
image. Somkiat Wangsiripitak
© 2002 R. C. Gonzalez & R. E. Woods

16
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

Huang [1965] attempted to quantify experimentally the effects on

image quality produced by varying N and k simultaneously.
Subjective test
Images similar to those shown in Fig. 2.22 were used.

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.3 Spatial and Gray-Level Resolution

Isopreference curves
(Nk-plane)
Each point in the Nk-plane
represents an image having values of N and k
equal to the coordinates of that point.
Points lying on an isopreference curve correspond to
images of equal subjective quality.
Isopreference curves tended to shift right and upward, means larger
values for N and k, which implies better picture quality.
Isopreference curves tend to become more vertical as the detail in
the image increases. Æ only a few gray levels may be needed.
A decrease in k tends to increase the apparent contrast of anSomkiat
image. Wangsiripitak
© 2002 R. C. Gonzalez & R. E. Woods

17
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.4 Aliasing and Moiré Patterns

Shannon sampling theorem :

If the function is sampled at a rate equal to or greater than twice
its highest frequency, it is possible to recover completely the
original function from its samples.
undersampled Æ aliasing corrupts
the sampled image.
The corruption is in the form of
additional frequency components
being introduced into the sampled
function :
Æ Aliased frequencies

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.5 Zooming and Shrinking Digital Images

Zooming requires two steps:

1. the creation of new pixel locations
2. the assignment of gray levels to those new locations
Æ interpolation
Pixel replication : applicable when increase the size of an image an
integer number of times.
Nearest neighbor interpolation : Assign the gray level of the closest
pixel to the new pixel in the grid.
Æ fast, but it produces a checkerboard effect.
Bilinear interpolation : using the four nearest neighbors of a point

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

18
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.4.5 Zooming and Shrinking Digital Images

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5 Some Basic Relationships Between Pixels

2.5.1 Neighbors of a Pixel

A pixel p at coordinates (x, y).

4-neighbors of p : four horizontal and vertical neighbors
N4(p) : (x+1, y), (x-1, y), (x, y+1), (x, y-1)
Diagonal neighbors of p
ND(p) : (x+1, y+1), (x-1, y-1), (x-1, y+1), (x-1, y-1)

8-neighbors of p : N4(p) together with ND(p)

N8(p) : (x+1, y), (x-1, y), (x, y+1), (x, y-1),
(x+1, y+1), (x-1, y-1), (x-1, y+1), (x-1, y-1)

Some of neighbors of p lie outside the image if (x,y) is on the

border of the image.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

19
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.2 Adjacency, Connectivity,

Regions, and Boundaries

To establish if two pixels are connected, it must be determined

1. if they are neighbors and
2. if their gray levels satisfy a specified criterion of similarity.
Ex.
In a binary image with values 0 and 1, two pixels may be 4-neighbors,
but they are said to be connected only if they have the same value.
V : the set of gray-level values used to define adjacency.
In binary image, V = {1} if we are referring to adjacency of pixels with
value 1.
In a gray-scale image, set V typically contains more elements.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.2 Adjacency, Connectivity,

Regions, and Boundaries

4-adjacency : Two pixels p and q with values from V are 4-adjacency

if q is in the set N4(p).
8-adjacency : Two pixels p and q with values from V are 8-adjacency
if q is in the set N8(p).
m-adjacency (mixed adjacency) : Two pixels p and q with values
from V are m-adjacency if
(i) q is in N4(p), or
(ii) q is in ND(p) and the set N4(p) ∩ N4(q) has no
pixels whose values are from V.

m-adjacency is introduced to eliminate the ambiguities that often

arise when 8-adjacency is used.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

20
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.2 Adjacency, Connectivity,

Regions, and Boundaries

Fig. 2.26 : for V = {1}

Two image subsets S1 and S2 are adjacent if some pixel in S1 is adjacent

to some pixel in S2.
A path from pixel p with coordinates (x,y) to pixel q with coordinates
(s,t) is a sequence of distinct pixels with coordinates
(x0 , y0), (x1 , y1), …, (xn , yn)
where (x0 , y0) = (x, y) , (xn , yn) = (s, t) and pixels (xi , yi) and (xi-1 , yi-1)
are adjacent for 1 ≤ i ≤ n.
Æ n is the length of a path.
if (x0 , y0) = (xn , yn), the path is a closed path. Somkiat Wangsiripitak
© 2002 R. C. Gonzalez & R. E. Woods

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.2 Adjacency, Connectivity,

Regions, and Boundaries
4-path, 8-path, and m-path can be defined depending on the type of
adjacency specified.
Two pixels p and q are said to be connected in subset S if there exists
a path between them consisting entirely of pixels in S.
A connected component of S : the set of pixels that are connected in S
A connected set : the set that only has one connected component in it.
A subset of pixels R is a region of the image if R is a connected set.
The boundary (border, contour) of a region R is the set of
pixels in the region that have one or more neighbors that are not in R.
If R happened to be an entire image, then its boundary is the
set of pixels in the first and last rows and columns of the image.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

21
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.3 Distance Measures

For pixels p, q, and z, with coordinates (x,y), (s,t), and (v,w),

respectively, D is a distance function or metric if
(a) D(p,q) ≥ 0 ( D(p,q) = 0 if p = q ),
(b) D(p,q) = D(q,p), and
(c) D(p,z) ≤ D(p,q) + D(q,z).
The Euclidean distance between p and q is
[ ]
1
De ( p, q ) = ( x − s ) + ( y − t ) (2.5-1)
2 2 2

A D4 distance (city-block distance) is

D4 ( p, q ) = x − s + y − t (2.5-2) 2
2 1 2
The pixels with D4 = 1 are the 4-neighbors of (x,y). 2 1 0 1 2
2 1 2
2
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.3 Distance Measures

A D8 distance (chessboard distance) is 2 2 2 2 2

D8 ( p, q) = max( x − s , y − t ) (2.5-3) 2 1 1 1 2
2 1 0 1 2
The pixels with D8 = 1 are the 8-neighbors of (x,y). 2 1 1 1 2
2 2 2 2 2
D4 and D8 distances are independent of any paths that might exist
between the points.
Dm distance between two points is defined as the shortest m-path
between the points.
Ex. p3 p4 when V = {1}
p1 p2 p, p2, and p4 have value 1
p

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

22
 Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.5.3 Distance Measures

0 1
0 1
1
Dm distance between p and p4 is 2.
0 1 1 1
1 1 0 1
1 1
Dm distance between p and p4 is 3.
1 1
1 1
1
Dm distance between p and p4 is 4.
© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

Digital
Digital Image
Image Processing,
Processing, 2nd
2nd ed.
ed. www.imageprocessingbook.com

2.6 Linear and Nonlinear Operations

H : an operator whose input and output are images.

H is a linear operator if, for any two images f and g and any two
scalars a and b,
H (af + bg ) = aH ( f ) + bH ( g ) (2.6-1)
Ex.
An operator whose function is to compute the sum of K images is a
linear operator.
An operator that computes the absolute value of the difference of two
images is not.
Æ nonlinear

© 2002 R. C. Gonzalez & R. E. Woods Somkiat Wangsiripitak

23