2. Outlines
1 What is segmentation ?
2 Diļ¬erent approach for image segmentation
Discontinuity based
Region based
3 Diļ¬erent edge detector operator
4 Linking of edge points
Local processing
Global processing
5 References
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 2 / 13
3. Image segmentation
ā It is a process for dividing an image into its constituent part.
Q At which level this division should be stopped.
Ans Level of division is application dependent entity.
Detection of movement measurement of vehicle on a road.
Types of image segmentation
1 Discontinuity based approach:
This approach is applicable, where there arise a abrupt changes in the
intensity level in an image.
Isolated points
Lines present in an image
Edges
2 Similarity based approach:
Grouping of those pixels, which are similar in some sense.
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 3 / 13
4. Continuedā
Thresholding operation
Region growing based approach
Region splitting and merging
Discontinuity based approach
Using suitable mask, we may be able for detecting
Isolated points
Lines present in an image
Edges
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 4 / 13
5. Continuedā
R =
1
i=ā1
1
j=ā1
Wi,j f (x + i, y + j)
1 Point detection
For point detection
|R| > T
where T is the given threshold.
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 5 / 13
6. Line detection
2 Line detection
Note: If |Ri | > |Rj | ā i = j then associated mask is more aligned towards
the direction of ith mask.
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 6 / 13
7. Continuedā
3 Edge detection
Note: 2nd order derivative is very sensitive to the noiseā not suitable for
edge detection
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 7 / 13
8. Continuedā
Let f (x, y) is a image signal, where
ā
f
=
Gx
Gy
=
āf
āx
āf
āy
or
f = mag(ā
f
) = [Gx2
+ Gy2
]1/2
ā |Gx| + |Gy|
Direction of ā
f
Ī±(x, y) = tanā1 Gy
Gx
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 8 / 13
10. Results obtained due to Sobel edge operator
Results obtained due to Sobel edge operator
Note: Previt and Sobel operators are the 1st order derivative operators.
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 10 / 13
11. Second derivative operator (Laplacian)
2
f =
ā2
f
āx2
+
ā2
f
āy2
Laplacian of Gaussian operator (LoG):
h(x, y) = eā x2+y2
2Ļ2
Let x2
+ y2
= r2
then
2
h =
r2
ā Ļ2
Ļ4
exp ā
r2
2Ļ2
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 11 / 13
13. References
M. Sonka, V. Hlavac, and R. Boyle, Image processing, analysis, and machine vision.
Cengage Learning, 2014.
D. A. Forsyth and J. Ponce, āA modern approach,ā Computer vision: a modern
approach, vol. 17, pp. 21ā48, 2003.
L. Shapiro and G. Stockman, āComputer vision prentice hall,ā Inc., New Jersey,
2001.
R. C. Gonzalez, R. E. Woods, and S. L. Eddins, Digital image processing using
MATLAB. Pearson Education India, 2004.
Subject: Image Procesing & Computer Vision Dr. Varun Kumar (IIIT Surat)Lecture 26 13 / 13