image segmentation

1. Narmada Reddy
Kari
Module -2
Image Segmentation

2. Image Segmentation
 Segmentation attempts to partition the pixels of an image
into groups that strongly correlate with the objects in an
image
 Typically the first step in any automated computer vision
application
4/28/2021
Narmada Reddy Kari
2

3. Introduction to image segmentation
Usually image segmentation is an initial and vital
step in a series of processes aimed at overall image
understanding
 Applications of image segmentation include
 Identifying objects in a scene for object-based
measurements such as size and shape
 Identifying objects in a moving scene for object-based
video compression (MPEG4)
 Identifying objects which are at different distances from
a sensor using depth measurements from a laser range
finder enabling path planning for a mobile robots
4/28/2021
Narmada Reddy Kari
3

4. Introduction to image segmentation
4/28/2021
Narmada Reddy Kari
4
 Example 1
• Segmentation based on greyscale
• Very simple ‘model’ of greyscale leads to inaccuracies in
object labelling

5. Introduction to image segmentation
 Example 2
• Segmentation based on texture
• Enables object surfaces with varying patterns of grey to
be segmented
4/28/2021
Narmada Reddy Kari
5

6. Introduction to Image segmentation
4/28/2021
Narmada Reddy Kari
6

7. Introduction to image segmentation
 Example 3
• Segmentation based on motion
• The main difficulty of motion segmentation is that an
intermediate step is required to (either implicitly or
explicitly) estimate an optical flow field
• The segmentation must be based on this estimate and not,
in general, the true flow
4/28/2021
Narmada Reddy Kari
7

8. Introduction to image segmentation
4/28/2021
Narmada Reddy Kari
8

9. Introduction to image segmentation
 Example 3
• Segmentation based on depth
• This example shows a range image, obtained with a laser
range finder
• A segmentation based on the range (the object distance
from the sensor) is useful in guiding mobile robots
4/28/2021
Narmada Reddy Kari
9

10. Image Segmentation
• Group similar components (such as, pixels in an image,
image frames in a video) to obtain a compact
representation.
• Applications: Finding tumors, veins, etc. in medical
images, finding targets in satellite/aerial images, finding
people in surveillance images, summarizing video, etc.
• Methods: Thresholding, K-means clustering, etc.
4/28/2021
Narmada Reddy Kari
10

11. Image Segmentation
 Segmentation algorithms for monochrome images
generally are based on one of two basic properties of gray-
scale values:
 Discontinuity
• The approach is to partition an image based on abrupt changes in
gray-scale levels.
• The principal areas of interest within this category are detection of
isolated points, lines, and edges in an image.
 Similarity
• The principal approaches in this category are based on
thresholding, region growing, and region splitting/merging.
4/28/2021
Narmada Reddy Kari
11

12. Image Segmentation
 Segmentation algorithms generally are based on
one of two basis properties of intensity values
• Discontinuity: to partition an image based on
abrupt changes in intensity (such as edges)
• Similarity: to partition an image into regions
that are similar according to a set of
predefined criteria.
4/28/2021
Narmada Reddy Kari
12

13. Image Segmentation
 Image Segmentation
4/28/2021
Narmada Reddy Kari
13

14. Image Segmentation
 Image Segmentation
4/28/2021
Narmada Reddy Kari
14

15. Image Segmentation
 Detection of discontinuities:
• There are three basic types of gray-level
discontinuities:
points , lines , edges
• The common way is to run a mask through the image
4/28/2021
Narmada Reddy Kari
15

16. Point Detection
 Note that the mark is the same as the mask of Laplacian
Operation
 The only differences that are considered of interest are
those large enough (as determined by T) to be
considered isolated points.
 |R| >T
4/28/2021
Narmada Reddy Kari
16

17. Point Detection
4/28/2021
Narmada Reddy Kari
17

18. Line Detection
4/28/2021
Narmada Reddy Kari
18
 Horizontal mask will result with max response when a line
passed through the middle row of the mask with a constant
background.
 the similar idea is used with other masks.
 note: the preferred direction of each mask is weighted with a
larger coefficient (i.e.,2) than other possible directions.

19. Line Detection
 Apply every masks on the image
 let R1, R2, R3, R4 denotes the response of
the horizontal, +45 degree, vertical and -45
degree masks, respectively.
 If, at a certain point in the image
 |Ri| > |Rj|, for all j≠i,
 That point is said to be more likely
associated with a line in the direction of
mask i.
4/28/2021
Narmada Reddy Kari
19

20. Line Detection
 Alternatively, if we are interested in detecting all lines in
an image in the direction defined by a given mask, we
simply run the mask through the image and threshold the
absolute value of the result.
 The points that are left are the strongest responses, which,
for lines one pixel thick, correspond closest to the
direction defined by the mask.
4/28/2021
Narmada Reddy Kari
20

21. Line Detection
4/28/2021
Narmada Reddy Kari
21

22. Edge Detection Approach
• Segmentation by finding pixels on a region boundary.
• Edges found by looking at neighboring pixels.
• Region boundary formed by measuring gray value
differences between neighboring pixels
4/28/2021
Narmada Reddy Kari
22

23. Edge Detection
 An edge is a set of connected pixels that
lie on the boundary between two regions.
 An edge is a “local” concept whereas a
region boundary, owing to the way it is
defined, is a more global idea.
4/28/2021
Narmada Reddy Kari
23

24. Edge Detection
4/28/2021
Narmada Reddy Kari
24

25. Edge Detection
4/28/2021
Narmada Reddy Kari
25

26. Edge Detection
4/28/2021
Narmada Reddy Kari
26

27. Edge Detection
4/28/2021
Narmada Reddy Kari
27
 Detection of discontinuities: Image Derivatives

28. Edge Detection
 First column: images and gray- level
profiles of a ramp edge corrupted by
random Gaussian noise of mean 0 and
= 0.0, 0.1, 1.0 and 10.0, respectively.
 Second column: first-derivative
 images and gray-level profiles.
 Third column : second- derivative
images and gray- level profiles.
4/28/2021
Narmada Reddy Kari
28

29. Edge Detection
4/28/2021
Narmada Reddy Kari
29
 Gradient Operator

30. Edge Detection
4/28/2021
Narmada Reddy Kari
30

31. Edge Detection
4/28/2021
Narmada Reddy Kari
31

32. Edge Detection
4/28/2021
Narmada Reddy Kari
32

33. Edge Detection
4/28/2021
Narmada Reddy Kari
33

34. Edge Detection
4/28/2021
Narmada Reddy Kari
34

35. Edge Detection
4/28/2021
Narmada Reddy Kari
35

36. Edge Detection
4/28/2021
Narmada Reddy Kari
36

37. Edge Detection
4/28/2021
Narmada Reddy Kari
37

38. Edge Detection
4/28/2021
Narmada Reddy Kari
38

39. Edge Linking and Boundary
Detection
4/28/2021
Narmada Reddy Kari
39

40. Edge Linking and Boundary
Detection
4/28/2021
Narmada Reddy Kari
40

41. 4/28/2021
Narmada Reddy Kari
41

42. Edge Linking and Boundary
Detection
4/28/2021
Narmada Reddy Kari
42

43. Global Processing Via Hough transform
4/28/2021
Narmada Reddy Kari
43

44. Global Processing Via Hough transform
4/28/2021
Narmada Reddy Kari
44

45.  Global processing: The Hough Transform
4/28/2021
Narmada Reddy Kari
45
Hough Transform

46. Global processing: The Hough
Transform
4/28/2021
Narmada Reddy Kari
46

47. Global processing: The Hough
Transform
4/28/2021
Narmada Reddy Kari
47

48. Global processing: The Hough
Transform
4/28/2021
Narmada Reddy Kari
48

49. Global processing Via Graph-Theoretic
Techniques
4/28/2021
Narmada Reddy Kari
49

50. Global processing Via Graph-
Theoretic Techniques
4/28/2021
Narmada Reddy Kari
50

51. Global processing Via Graph-
Theoretic Techniques
4/28/2021
Narmada Reddy Kari
51

52. Global processing Via Graph-Theoretic
Techniques
4/28/2021
Narmada Reddy Kari
52

53. Global processing Via Graph-
Theoretic Techniques
4/28/2021
Narmada Reddy Kari
53

54. Image Segmentation
4/28/2021
Narmada Reddy Kari
54

55. Hough Transform
 Global processing: The Hough Transform
4/28/2021
Narmada Reddy Kari
55

56. Global processing: The Hough Transform
4/28/2021
Dr. S.K.Singh
56

57. Image Thresholding
 What is thresholding?
 Simple thresholding
 Adaptive thresholding
4/28/2021
Narmada Reddy Kari
57

58. Thresholding
 Suppose that an image, f(x,y), is composed of light
objects on a dark backround, and the following
figure is the histogram of the image.
 Then, the objects can be extracted by comparing
pixel values with a threshold T.
4/28/2021
Narmada Reddy Kari
58

59. Thresholding – A KeyAspect
• Most algorithms involve establishing a threshold level of
certain parameter.
• Correct thresholding leads to better segmentation.
• Using samples of image intensity available, appropriate
threshold should be set automatically in a robust
algorithm i.e. no hard-wiring of gray values
4/28/2021
Narmada Reddy Kari
59

60. Automatic Thresholding
Use of one or more of the following:-
Intensity characteristics of objects
1. Sizes of objects
2.Fractions of image occupied by objects
3.Number of different types of objects
Size and probability of occurrence – most popular
Intensity distributions estimate by histogram computation.
4/28/2021
Narmada Reddy Kari
60

61. Automatic Thresholding Methods
 Some automatic thresholding schemes:
1. P-tile method
2. Iterative threshold selection
3. Adaptive thresholding
4/28/2021
Narmada Reddy Kari
61

62. 4/28/2021
Narmada Reddy Kari
62

63. P-Tile Thresholding
4/28/2021
Narmada Reddy Kari
63

64. P-Tile Thresholding
 Basic global thresholding:
Based on the histogram of an image Partition the image
histogram using a single global threshold
 Basic global thresholding:
The success of this technique very strongly depends on
how well the histogram can be partitioned
4/28/2021
Narmada Reddy Kari
64

65. Iterative P-Tile Thresholding
 The Basic global thresholding:
1. Select an initial estimate for T (typically the
average grey level in the image)
2. Segment the image using T to produce two
groups of pixels: G1 consisting of pixels
with grey levels >T and G2 consisting
pixels with grey levels ≤ T
3. Compute the average grey levels of pixels
in G1 to give μ1 and G2 to give μ2
4/28/2021
Narmada Reddy Kari
65

66. Thresholding
4/28/2021
Narmada Reddy Kari
66

67. P-Tile Thresholding
4/28/2021
Narmada Reddy Kari
67

68. Limitation of P-Tile thresholding:
4/28/2021
Narmada Reddy Kari
68

69. Limitation of P-Tile thresholding:
4/28/2021
Narmada Reddy Kari
69

70. Limitation of P-Tile thresholding:
4/28/2021
Narmada Reddy Kari
70

71. Adaptive Thresholding
 Thresholding – Basic Adaptive Thresholding
4/28/2021
Narmada Reddy Kari
71

72. Adaptive Thresholding
 Thresholding – Basic Adaptive Thresholding
4/28/2021
Narmada Reddy Kari
72

73. Adaptive Thresholding
 Thresholding – Basic Adaptive Thresholding
4/28/2021
Narmada Reddy Kari
73

74. Adaptive Thresholding
 Thresholding – Basic Adaptive Thresholding
4/28/2021
Narmada Reddy Kari
74

75. Adaptive Thresholding
 Thresholding – Basic Adaptive Thresholding
4/28/2021
Narmada Reddy Kari
75

76. What is a Region?
• Basic definition :- A group of connected pixels with
similar properties.
• Important in interpreting an image because
they may correspond to objects in a scene.
• For that an image must be partitioned into regions
that correspond to objects or parts of an object.
4/28/2021
Narmada Reddy Kari
76

77. Region-Based vs. Edge-Based
4/28/2021
Narmada Reddy Kari
77

78. Region-Oriented Segmentation
 Region Growing
• Region growing is a procedure that groups pixels or subregions into
larger regions.
• The simplest of these approaches is pixel aggregation, which starts
with a set of “seed” points and from these grows regions by
appending to each seed points those neighboring pixels that have
similar properties (such as gray level, texture, color, shape).
• Region growing based techniques are better than the edge-based
techniques in noisy images where edges are difficult to detect.
4/28/2021
Narmada Reddy Kari
78

79. Region-Oriented Segmentation
4/28/2021
Narmada Reddy Kari
79

80. Region-Oriented Segmentation
4/28/2021
Narmada Reddy Kari
80

81. Region-Oriented Segmentation
 Region Splitting
 Region growing starts from a set of seed points.
 An alternative is to start with the whole image as a single
region and subdivide the regions that do not satisfy a
condition of homogeneity.
 Region Merging
 Region merging is the opposite of region splitting.
 Start with small regions (e.g. 2x2 or 4x4 regions) and
merge the regions that have similar characteristics (such
as gray level, variance).
 Typically, splitting and merging approaches are used
iteratively.
4/28/2021
Narmada Reddy Kari
81

82. Region-Oriented Segmentation
4/28/2021
Narmada Reddy Kari
82

83. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
83
 Visualize an image in 3D: spatial coordinates and gray levels.
 In such a topographic interpretation, there are 3 types of points:
Points belonging to a regional minimum
• Points at which a drop of water would fall to a single minimum.
(The catchment basin or watershed of that minimum.)
• Points at which a drop of water would be equally likely to fall to
more than one minimum. (The divide lines or watershed lines.)
Watershed lines

84. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
84
 The objective is to find watershed lines.
 The idea is simple:
• Suppose that a hole is punched in each regional minimum and that the
entire topography is flooded from below by letting water rise through the
holes at a uniform rate.
• When rising water in distinct catchment basins is about the merge, a dam
is built to prevent merging. These dam boundaries correspond to the
watershed lines.

85. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
85

86. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
86
 Start with all pixels with the lowest possible value.
• These form the basis for initial watersheds
 For each intensity level k:
• For each group of pixels of intensity k
• If adjacent to exactly one existing region, add these
pixels to that region
• Else if adjacent to more than one existing regions,
mark as boundary
• Else start a new region

87. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
87
Watershed algorithm might be used on the gradient image instead of the
original image.

88. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
88
Due to noise and other local irregularities of the gradient, over segmentation
might occur.

89. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
89
A solution is to limit the number of regional minima. Use markers to specify
the only allowed regional minima.

90. Watershed Segmentation Algorithm
4/28/2021
Narmada Reddy Kari
90
A solution is to limit the number of regional minima. Use markers to specify
the only allowed regional minima. (For example, gray-level values might be
used as a marker.)

91. THANK YOU