Sunteți pe pagina 1din 19

CHAPTER 2: DIGITAL IMAGE FUNDAMENTALS

Structure of the human eye


3 membranes:

• cornea & sclera


• choroid
• retina

Cornea: tough, transparent


tissue that covers the Iris diaphragm:
anterior surface of the eye. contracts or
d ≈ 20mm expands to control
Sclera: opaque membrane. the amount of light.

Choroid: contains a
network of blood vessels.

Retina: innermost Two types of receptors: cones and rods.


membrane.
Cones: 6-7 million. Each connected to its own
nerve end. Highly sensitive to color.
Cone vision is called photopic vision.

Rods: 75-150 million. Several are connected to a


Lens: suspended by fibers attached to the single nerve end. Serve to give an overall picture
ciliary body (60-70% water, 6% fat, protein). of the field of view. Not involved in color vision.
Rod vision is called scotopic vision.
Absorbs ~8% of the visible light spectrum.
Blind spot: the area on the retina without
Infrared and ultraviolet light are also absorbed. receptors that respond to light.

Fovea: Circular indentation in the retina (~1.5mm


in diameter)

53
IMAGE FORMATION IN THE EYE

The receptors transform radiant


energy into electrical impulses that
are decoded by the brain. What is the
height of
the image?

Focal length varies


from 14mm to 17mm.

56
SENSOR ARRAYS

The energy from


an illumination
source is reflected
from a scene
element. The lens projects
the viewed scene
onto the lens focal
plane.

The sensor array The digitized


The incoming produces outputs image is obtained.
energy is proportional to
collected and the integral of the
focused onto light received at
an image plane. each sensor.

70
SAMPLING AND QUANTIZATION

‰ The output of most sensors is continuous voltage


waveform.
‰ To create a digital image, continuous sensed data should
be converted into digital form.
‰ Two processes: sampling and quantization.
‰ Consider a function f(x,y) representing an image.
‰ f(x,y) is continuous w.r.t.
 x and y coordinates
 Amplitude
‰ Sampling: digitizing the coordinate values
‰ Quantization: digitizing the amplitude values

73
GENERATING A DIGITAL IMAGE

74
IMAGE ACQUISITION WITH A SENSING ARRAY

75
DIGITAL IMAGE REPRESENTATION

f(0,0)

f(0,N-1)

L: # of discrete gray levels

L = 2k
b=MxNxk

f(M-1,0)

76
STORAGE REQUIREMENTS

77
SPATIAL AND GRAY LEVEL RESOLUTION

Sampling is performed by deleting rows and columns from the original image.

78
RESAMPLING INTO 1024X1024 PIXELS

Resampling is performed by row and column duplication.

79
256/128/64/32 GRAY LEVELS

In these
images, the #
of samples is
constant but
the # of gray
levels was
reduced from False contouring
256 to 32.

80
16/8/4/2 GRAY LEVELS

In these
images, the #
of samples is
constant but
the # of gray
levels was
reduced from
16 to 2.

81
ZOOMING DIGITAL IMAGES

‰ Zooming requires 2 steps:


 The creation of new pixel locations.
 The assignment of gray levels to these new locations.
 Nearest neighbor interpolation
„ A 500x500 image will be enlarged 1.5 times.
„ Lay an imaginary 750x750 grid over the original image.
„ The spacing is less than one pixel because the grid is fitted over a
smaller image.
„ For any point in the overlay, look for the closest pixel in the original
image, and assign its gray level to the new pixel in the grid.
„ When all the new pixels are assigned values, expand the overlay grid
to the original specified size to obtain the zoomed image.
„ Pixel replication is a special case that is applicable when the size of
the image needs to be increased an integer number of times.
„ Nearest neighbor interpolation is fast but it produces a checkerboard
effect!
 Bilinear interpolation - A slightly more sophisticated method
„ (x',y'): coordinates of a point in the zoomed image.
„ v (x',y'): gray level assigned to the point.
„ v (x',y') = ax'+by'+cx'y'+d (the 4 coefficients are determined using the 4
nearest neighbors of (x',y').
83
SHRINKING DIGITAL IMAGES

‰ Similar to image zooming.


‰ Shrinking an image an integer number of times
 Pixel replication is replaced by row&column deletion.
‰ Shrinking an image by a non-integer factor
 Expand the grid to fit over the original image.
 Do gray-level interpolation (nearest neighbor or bilinear).
 Shrink the grid back to its original specified size.
‰ It is possible to use more neighbors for interpolation.
 Points are fitted with a more complex surface.
 Generally gives smoother results.
 An important consideration for a number of applications.
„ Image generation for 3D graphics
„ Medical image processing
 The extra computational burden is seldom justifiable for
general-purpose digital image zooming and shrinking.
„ In general, bilinear interpolation is the method of choice.

84
IMAGE ZOOMING: NEAREST NEIGHBOR
INTERPOLATION VS. BILINEAR INTERPOLATION

85
NEIGHBORS OF A PIXEL

4 diagonal neighbors
of p, ND(p)

8-neighbors of p, N8(p) 4-neighbors of p, N4(p)

pixel p at (x,y)

86
ADJACENCY, CONNECTIVITY

‰ V: set of gray level values


‰ 3 types of adjacency
 4- adjacency: 2 pixels p and q with values from V are 4-
adjacent if q is in the set N4(p)
 8- adjacency: 2 pixels p and q with values from V are 8-
adjacent if q is in the set N8(p)
 m- adjacency: 2 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 set N4(p) ∩ N4(q) has no pixels whose values
are from V
‰ A digital 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)
and (xn,yn)=(s,t), and pixels (xi,yi) and (xi-1,yi-1) are adjacent
for 1 ≤ i ≤ n.

87
REGIONS, BOUNDARIES

‰ 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.
 For any pixel p in S, the set of pixels that are connected to it in
S is called a connected component of S.
 If S has only one connected component, it is called a connected
set.
‰ R: a subset of pixels in an image. R is a region of the
image if R is a connected set.
‰ The boundary of a region R is the set of pixels in the region
that have one or more neighbors that are not in R.

88
DISTANCE MEASURES

‰ p with (x,y) D(p,q) = 0 iff p = q


‰ q with (s,t) D(p,q) = D(q,p)
‰ z with (v,w) D(p,z) ≤ D(p,q) + D(q,z)

‰ Euclidean distance between p and q: De(p,q) = [(x-s)2 + (y-t)2]1/2

‰ D4 distance: D4(p,q) = |x-s| + |y-t|

‰ D8 distance: D8(p,q) = max (|x-s| , |y-t|)

‰ D4 and D8 distances between p and q are independent of any


paths that might exist between the points.
‰ For m-adjacency, Dm distance between two points is defined as
the shortest m-path between the points.

89

S-ar putea să vă placă și