0% found this document useful (0 votes)
5 views59 pages

Digital Image Processing Basics

Uploaded by

heba heba
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views59 pages

Digital Image Processing Basics

Uploaded by

heba heba
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

CoE4TN4

Image Processing

Chapter 2: Digital Image


Fundamentals
Digital Image Fundamentals

• Elements of visual perception


• Image sensing and acquisition
• Sampling and quantization
• Relationship between pixels
• Mathematical tools used in image processing

2
The Human Visual System (HVS)
• Why study the HVS?
– A true measure of image processing quality is how well
the image appears to the observer.

– The HVS is very complex and is not understood well in a


complete sense. However, many of its properties can be
identified and used to our advantage.

3
Structure of Human Eye
• Eye: Sphere, diameter of 20 mm
• Consists of 3 membranes:
1. Cornea and sclera
2. Choroid
3. Retina
• Cornea: transparent
• Sclera: opaque, connected to cornea
• Choroid: network of blood vessels
• In front choroid is connected to iris diaphragm
• Iris: contracts or expands to control amount of light
• Pupil: central opening of iris, 2 to 8 mm in diameter

4
Structure of Human Eye
• Lens:
– Focuses light on retina
– Contains 60% to 70% water
– Absorbs 8% of visible light
– High absorption in infrared and ultraviolet (can cause damage to eye)
• Retina: the inner most layer, covers the posteriori portion of
eye
• When eye is properly focused, light of an object is imaged on
the retina
• Light receptors are distributed over the surface of retina

5
• Retina contains light receptors: Cones & rods
– Cones:
• 6 to 7 million,
• located mainly in central part of retina
(fovea)
• Sensitive to color
• Can resolve fine details because each
one is connected to its nerve
• Cone vision: photopic or bright-light
– Rods:
• 75 to 150 million
• No color vision, responsible for low-
light vision
• Distributed a wide region on the
retina
• Rod vision: scotopic or dim-light

6
Human Eye
• Blind spot: a region of retina without receptors, optic nerves
go through this part
• Fovea: a circular area of about 1.5 mm in diameter
• A comparison between eye (fovea) and a CCD camera:
– Density of cones in fovea: 150,000 /mm2
– Number of cones: 337,000
– A medium resolution CCD chip has the same number of elements in a
5mm x 5mm area.

7
Image formation in the eye
• Lens is flexible
• Refraction of lens is controlled by its thickness
• Thickness is controlled by the tension of muscles connected to
the lens
• Focus on distance objects: lens is relatively flattened,
refractive power is minimum
• Focus on near objects: lens is thicker, refractive power is
maximum
• Perception takes place by excitation of receptors which
transform radiant energy into electrical impulses that are
decoded by the brain.

8
Brightness & Intensity
• The dynamic range of light intensity to which eye can adapt is
enormous - on the order of 1010- from the scotopic threshold
to the glare limit
• Brightness (intensity perceived by visual system) is a
logarithmic function of light intensity.
• HVS can not operate over the entire range simultaneously. It
accomplishes large variations due to brightness adaptation

9
Weber Experiment
• To characterize the intensity discrimination properties of eye
• ΔI starts at zero and is increased slowly
• The observer is asked to indicate when the circle on the
constant background becomes visible (just noticeable
difference).
• The ratio ΔI/I is called the Weber ratio.
• Procedure is repeated for different values of I.

10
Weber Experiment
• Small values of ΔI/I: good discrimination
• Large values of ΔI/I: poor discrimination
• Low levels of illumination: high Weber ratio: poor
discrimination
• In high levels of illumination, discrimination improves.

11
Intensity & Brightness
• Relationship between brightness and intensity is not a simple
function!

12
Intensity & Brightness
• Mach Band effect:
Although the shades are
constant, overshoot and
undershoot are observed near
the transition boundary.

13
Intensity & Brightness
• Simultaneous contrast:
perceived brightness does not
depend simply on intensity

14
Image sensing and acquisition
• If a sensor can be developed that is capable of detecting
energy radiated by a band of the EM spectrum, we can image
events in that band.
• Image is generated by energy of the illumination source
reflected (natural scenes) or transmitted through objects (X-
ray)
• A sensor detects the energy and converts it to electrical
signals
• Sensor should have a material that is responsive to the
particular type of energy being detected.

15
Image sensing and acquisition
• Three principle sensor arrangements:
1. Single imaging sensor
2. Line sensor
3. Array sensor

16
Single sensor
• Most familiar sensor of this type is photodiode
• In order to generate a 2-D image using a single sensor, there
has to be relative displacement in both x and y directions
between the sensor and the area to be imaged

17
Sensor strips
• Sensor elements are arranged in a line
• Strip provides imaging in one direction, and motion provides
imaging in the other direction
• Used in scanners and airborne imaging
• Airborne imaging: imaging system is mounted on the aircraft
which flies at a constant altitude over the area to be imaged

18
Sensor strips
• Sensor strips mounted in a ring configuration are used in
medical and industrial imaging to obtain cross-sectional
images of 3-D objects (CAT)
• Output of the sensors must be processed by reconstruction
algorithms to transform the sensed data into meaningful cross-
sectional images.

19
Sensor arrays
• This is the arrangement used on digital cameras
• Typical sensor for these cameras is the CCD array (Charge
Coupled Devices)
• Since the sensor is two dimensional a complete image can be
obtained
• Motion is not necessary

20
A simple image model
• The amount of light that enters the eye depends
on:
1. The amount of source illumination incident on the
scene, i(x,y)
2. The amount of illumination reflected by the objects
in the scene, r(x,y)

(x,y): coordinates
Total absorption: r(x,y)=0
Total reflection: r(x,y)=1

21
Sampling & Quantization
• Computer processing: image f(x,y) must be digitized both
spatially and in amplitude
• Digitization in spatial coordinates: sampling
• Digitization in amplitude: quantization
• Image: [f(i,j)]NxM
• What should be the values of N, M and the number of gray
levels G?
• Normally: N=2n , M=2m , G=2k

22
Sampling & Quantization

23
Sampling & Quantization

• Number of bits required to store image: N x M x k


• The more the values of N,M and G: the better approximation
of a continuous image
• Storage and processing requirements increase as well

24
Spatial and Intensity Resolution
• Spatial resolution: smallest discernable detail in an image
• Expressed in dots (pixels) per unit distance
• In printing: dots per inch (dpi)
– Newspaper: 75 dpi
– Magazines: 133 dpi
• Intensity resolution: smallest discernible change in intensity
level
• Number of intensity levels is usually a power of 2
• 8 and 16 bits are common

25
Effects of Reducing Spatial Resolution

26
Effects of Reducing Gray Levels

27
Effects of Reducing Gray Levels

28
Effects of Reducing Gray Levels
• Effects of Reducing Gray Levels:
• appearance of fine ridge-like structures in areas of
smooth gray levels
• This effect is called false contouring

29
Image Interpolation
• Image interpolation: used extensively in tasks such as
zooming, shrinkage, rotation, geometric transforms
• Interpolation: estimate values at unknown locations
• Zooming: increasing the resolution of an image
• Shrinkage: decreasing the resolution of an image
• Example of zooming: we have an image of 500x500 pixels
and we want to enlarge it to 750x750
• Zooming has two steps: creation of new pixel locations and
the assignment of gray levels to those locations

30
Zooming
• A simple way of zooming which works for increasing the size
of an image by integer numbers is pixel replication
• Visualize assignment in zooming: the enlarged image is
placed on the original image
• Gray level of each pixel in the enlarged image is set to the
gray-level of its nearest pixel in the original image

31
Zooming
• A more sophisticated way of accomplishing gray-level
assignment is bilinear interpolation
– v(x,y)=ax+by+cxy+d
• The four coefficients are determined from the four equations
in four known (four nearest neighbors of the point (x,y))
• Next level of complexity: bicubic

32
Shrinkage
• Shrinkage by an integer number can be done by deleting some
of the rows and columns of the image
• Shrinkage by an noninteger factor can be done as the inverse
of zooming

33
Shrinkage and Zooming

34
Relationship between pixels
0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 1 1 1 1
0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 1 0 0 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 1 0 0 0 0 0 0
0 0 0 0 0 1 0 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 1 0 0

35
Relationship between pixels
• Neighbors
• Adjacency
• Path
• Connectivity
• Region
• Boundary
• Distance

36
Basic relationships between pixels
• A pixel p at coordinates (x,y) has four horizontal and vertical
neighbors:
N4(P)={(x+1,y), (x-1,y),(x,y+1),(x,y-1)}
• The four diagonal neighbors of P
ND(P)={(x+1,y+1), (x-1,y-1),(x-1,y+1),(x+1,y-1)}
• The eight point neighbors of P
N8(P)=N4(P)U ND(P)

37
Adjacency
• Two pixels are adjacent if they are neighbors and their gray
levels are similar
• V: set of gray levels
• Similar gray level means that the gray levels of both pixels
belong to set V
• Exp:
– Binary images: V={1}
– Gray level image: V={32,33, …,63,64}

38
Adjacency
• 4-adjacency: Two pixels p and q with values from V are 4-
adjacent if q is in N4(p)
• 8-adjacency: Two pixels p and q with values from V are 8-
adjacent if q is in N8(p)
• 4-adjacency: broken paths
• 8-adjacency: multiple paths

0 1 1 0 1 1 0 1 1
0 1 0 0 1 0 0 1 0
0 0 1 0 0 1 0 0 1

39
Adjacency
• m-adjacency: Two pixels p and q with values from V are m-
adjacent if:
q is in N4(p) or
q is in ND(p) and the intersection of N4(p) and N4(q) has no pixels with
values in V.

q
q
0 1 1 0 1 1 0 1 1
0 1 0 0 1 0 0 1 0
0 0 1 0 0 1 0 0 1

p p

40
Path
• 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 points (xi,yi) and (xi-1,yi-1) are adjacent for

• n is the length of the path


• We can have 4-, 8-, or m-paths depending on the type of
adjacency specified.

41
Connectivity
• S: a subset of pixels in an image
• Two pixels p and q are said to be connected in S if there exists
a path between them consisting entirely of pixels in S
• We can have 4-, 8-, or m-connectivity depending on the type
of path specified.

42
Region
• R: a subset of pixels in an image
• R is called a region if every pixel in R is connected to any
other pixel in R
• Boundary (border or contour) of a region: set of pixels in the
region that have one or more neighbors that are not in R

0 1 1 0 0 1 1 0
0 1 1 1 0 1 1 1
0 0 1 0 0 0 1 0
0 0 0 0 1 0 0 0

43
Distance measures
• For pixels p,q, and z with coordinates (x,y), (s,t) and (v,w),
respectively, D is a distance functions if:

44
Distance measures
• D4 distance

• D8 distance

Pixel values D4 distances D8 distances


0 1 1 2 1 2 1 1 1
0 1 0 1 0 1 1 0 1
0 0 1 2 1 2 1 1 1

45
Distance measures
• Dm distance: length of the shortest m-path between two pixels
• D4, D8 distance between p and q are independent of the pixels
along the path
• Dm depends on the value of the pixels between p and q

0 0 1 0 0 1
1 1 0 0 1 0
1 0 0 1 0 0
Dm=3 Dm=2

46
Math tools in image processing
• Linear operators
• Arithmetic operations
• Logic operations
• Spatial/Geometric operations
• Vector and matrix operations
• Image transforms

47
Linear & Non-linear operations
• H: an operator whose inputs and outputs are images
• H is linear if for any two images f and g and any two scalars a
and b
H(af+bg)=aH(f)+bH(g)

48
Arithmetic Operations
• Arithmetic operation between two images is performed
between corresponding pixel pairs
• Example: Averaging of noisy images for noise reduction

49
Logical Operations
• Logical operations are typically applied to binary images.
• We can think of foreground being 1 and background 0.

50
Geometric Transformations
• Geometrical transformations: modify the spatial relationships
between pixels in an image
• Geometrical transformations: called rubber-sheet
transformations
– They may be viewed as printing an image on a sheet of rubber and
then stretching this sheet according to some predefined rules

51
Geometric Transformations
• Geometrical transformation consists of two basic operations:
1. Spatial transformation: defines the rearrangement of pixels on the
image plane
2. Gray level interpolation: deals with the assignment of gray levels to
pixels in the spatially transformed image

52
Spatial Transformations
• Image f with pixels coordinates (v,w) has undergone
geometric transformation to produce an image g with
coordinates (x,y)

• Example: (x,y)=T{(v,w)}=(v/2, w/2)


– Shrinking of the size of f(x,y) by one-half in both directions.

53
Spatial Transformations
• Affine transform

• Can scale, rotate, translate or sheer depending of the values


chosen for elements of T

54
Spatial Transformations

55
Image Registration
• Image registration: to align two or more mages of the same
scene
• The images are available transformation is unknown
• Input image: image the we wish to transform
• Reference image: image against which we want to register
the input image

56
Image Registration
• Solution: spatial relocation is formulated by the use of
tiepoints.
• Tiepoints: a set of pixels whose locations in input and
reference images are known
• Suppose the geometrical transformation process within the
region is modeled by a pair of bilinear equations:
– x=c1v+c2w+c3vw+c4
– y=c5v+c6w+c7vw+c8
– 8 known tiepoints, 8 unknown ci

57
Gray-level Interpolation
• Depending on the values of ci x and/or y can be noninteger
for integer values of (v,w)
– x=c1v+c2w+c3vw+c4
– y=c5v+c6w+c7vw+c8
• We need inferring gray-level values at noninteger locations
(gray-level interpolation)

58
Image Registration

59

You might also like