REPRESENTATION

&
DESCRIPTION
Unit-9
1
 2

Basic
 After image has been segmented into
regions, methods required to represent and
describe them

 Representing regions involved two choices

 External characteristics
 Internal characteristics
 3

Basic
 External representation
 when the primary focus is on shape
characteristics

 Internal representation
 when the primary focus is on regional
properties such as color and texture

 Some times both of representations are used

 4

Basic
 Features selected as descriptors should be
insensitive to variations in size, translation
and rotation
 Representation

5
 6

Method
1. Chain codes
2. Polygonal approximations
3. Signatures
4. Boundary segmentation
5. Skeletons
 7

1 Chain code
 Used to represent a boundary by a
connected sequence of straight-line
segments of specified length and direction
 8

1 Chain code
 9

1 Chain code
 Chain code is depends on the starting point
 10

1 Chain code
 Code can be normalized with respect to
starting point

 Treat chain code as circular sequence of

direction numbers
 Redefine starting point so that resulting
sequence of numbers form an integer of
minimum magnitude
 11

1 Chain code
 Normalize also for rotation by using first
difference of chain code

 First difference

 Counting number of direction changes

(counterclockwise) that separate two adjacent
elements of the code
 12

1 Chain code
 e.g.
first difference of 4-direction chain code
10103322 is 3133030

1
10103322
3133030
2
3
 13

1 Chain code
 Disadvantages

 Quite long

 Small disturbance along the boundary due to

noise or imperfect segmentation cause
changes in the code
 14

2 Polygonal approximations
 Boundary can be approximated with
arbitrary accuracy by a polygon

 For closed curve

 Approximation is exact when
number of segments in polygon
is equal to number of points in boundary
 15

2 Polygonal approximations
 Minimum perimeter polygons

Consider object as rubber

Concatenated cell
Suppose rubber allowed to shrink
Encloses boundary
Takes this shape
 16

3 Signatures
 The idea behind a signature is to convert a
two dimensional boundary into a
representative one dimensional function
 17

3 Signatures
 Signatures are invariant to location, but will
depend on rotation and scaling

 Starting at the point farthest from the

reference point or using the major axis of the
region can be used to decrease dependence
on rotation

 Scale invariance can be achieved by either

scaling the signature function to fixed
amplitude or by dividing the function values
by the standard deviation of the function.
 18

3 Signatures
 19

4 Boundary segmentation
 Decompose a boundary into segments

 Use of the convex hull of the region enclosed

by the boundary is a powerful tool for robust
decomposition of the boundary
 20

4 Boundary segmentation
 21

5 Skeletons
 Produce a one pixel wide graph that has the
same basic shape of the region, like a stick
figure of a human. It can be used to analyze
the geometric structure of a region which
has bumps and “arms”
 22

5 Skeletons
 23

5 Skeletons
 Before a thinning algorithm:
 A contour point is any pixel with value 1 and
having at least one 8-neighbor valued 0.
 Let
N ( p1 )  p2  p3  ...  p8  p9
T ( p1 ) : the number of 0 - 1 transitio ns
in the ordered sequence
p2 , p3 ,..., p8 , p9 , p2
 24

5 Skeletons
 Step - 1
Flag a contour point p1 for deletion if the
following conditions are satisfied
 25

5 Skeletons
 Step - 2
Flag a contour point p1 for deletion again.
However, conditions (a) and (b) remain the
same, but conditions (c) and (d) are changed
to
 26

5 Skeletons
 A thinning algorithm:

 Applying step 1 to flag border points for

deletion
 Deleting the flagged points
 Applying step 2 to flag the remaining border
points for deletion
 Deleting the flagged points

 This procedure is applied iteratively until no

further points are deleted
 27

5 Skeletons
 One application of
skeletonization is
for character
recognition.
 A letter or
character is
determined by the
center-line of its
strokes, and is
unrelated to the
width of the stroke
lines
 Boundary description

28
 29

Method
1. Simple Descriptors
2. Shape numbers
3. Fourier descriptors
4. Statistical moments
 30

1 Simple Descriptors
 Length of boundary
 One of the simplest description
 Number of pixels in boundary gives a rough
approximation of its length

 Curvature
 Defined as rate of change of slope
 The difference between the slops of adjacent
boundary segments is used as a descriptor of
curvature at the point of intersection of
segments
 31

2 Shape numbers
 32

2 Shape numbers
 The shape number of a boundary is defined
as the first difference of smallest magnitude

 The order “n” of a shape number is defined

as the number of digits in its representation
 33

2 Shape numbers
 34

3 Fourier Descriptors
 The x-y coordinates of the boundary are
treated as the real and imaginary parts of a
complex number

 Then the list of coordinates is Fourier

transformed using the DFT

 The Fourier coefficients are called the Fourier

descriptors
 35

3 Fourier Descriptors
 The basic shape of the region is determined
by the first several coefficients, which
represent lower frequencies

 Higher frequency terms provide information

on the fine detail of the boundary
 36

3 Fourier Descriptors
 37

4 Statistical moments
 Moments are statistical measures of data
 They come in integer orders
 Order 0 is just the number of points in the
data
 Order 1 is the sum and is used to find the
average
 Order 2 is related to the variance, and order
3 to the skew of the data
 Higher orders can also be used, but don’t
have simple meanings
 38

4 Statistical moments
 Let r be a random variable, and g(ri) be
normalized (as the probability of value ri
occurring), then the moments are
K1 K1
 n (r )  (ri  m) n g (ri ) where m  ri g (ri )
k 0 i 0
 Region descriptors

39
 40

Method
1. Simple Descriptors
2. Topological descriptors
3. Texture
4. Moments of two-dimensional functions
 41

1 Simple Descriptors
 Area of a region
 number of pixels in the region

 Perimeter of a region
 Length of its boundary

 Compactness of a region
 (perimeter)2/area
 42

1 Simple Descriptors
 Mean and median of the gray levels

 Minimum and maximum gray-level values

 Number of pixels with values above and

below the mean
 43

1 Simple Descriptors
 44

2 Topological descriptors

Topological property 1:
the number of holes (H)

Topological property 2:
the number of connected
components (C)
 45

2 Topological descriptors
 Topological property 3:
Euler number
 Number of connected components subtract
the number of holes
 E=C-H

E=0 E= -1
 46

2 Topological descriptors
 47

3 Texture
 48

4 Moments of two-dimensional functions

 49

4 Moments of two-dimensional functions

Table 11.3 Moment invariants for the images in Figs. 11.25(a)-(e).

 50

END
OF
UNIT – 9