MODULE 5

CG & FIP
 Dr. T. THIMMAIAH INSTITUTE OF TECHNOLOGY
(Estd. 1986) Oorgaum, Kolar Gold Fields, Karnataka – 563120
(Affiliated to VTU, Belgaum, Approved by AICTE - New Delhi)
NAAC Accredited 'A' Grade, NBA Accredited for CSE, ECE & Mining Engg Programs
DEPA

RTMENT OF COMPUTER SCEINCE AND ENGINEERING

Table of Contents

MODULE: 5
Image Segmentation

Topics Page No.

5.1 Image Segmentation 5-1

5.1.1 Introduction 5-1

5.1.2. Classification of Image Segmentation Algorithm 5-1

5.1.3 Detection of Discontinuities 5-2

5.2 Edge Detection 5-3

5.2.1 Stages in Edge Detection 5-3

5.2.2 Types of Edge Detectors 5-4

5.2.3 First Order Edge Detection 5-5

5.2.4 Second Order Edge Detection 5-7

Question Bank 5-8

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

MODULE 5: Image Segmentation

5.1 Image Segmentation
5.1.1 Introduction
1. What is Image segmentation? Explain the characteristics of the segmentation
process.

Definition: Segmentation is the process of partitioning a digital image into

multiple regions and extracting meaningful regions known as regions of interest (ROI)
for further image.
The characteristics of the segmentation process are the following:
1. If the sub regions are combined, the original region can be obtained. Mathematical can
be stated that ∪Ri = R for i = 1, 2,..., n.
2. The sub regions Ri should be connected. In other words, the region cannot be open-
ended during the tracing process.
3. The regions R1, R2,..., R, do not share any common property. Mathematically, it can
be stated as Ri ∩ Rj = ∅ for all i and j where i≠ j.
4. Each region satisfies a predicate or a set of predicates such as intensity or other image
statistics. Mathematically P (Ri)=True, (P) can be color, texture or any other image
statistics.

5.1.2 Classification of Image Segmentation

Q2. Explain different ways of classifying the image segmentation algorithm?

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 1

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

A) Based on user interaction:

i) Manual method, the object of interest is observed by an expert who traces its
ROI boundaries as well, with the help of software. Hence, the decisions
related to segmentation are made y the human observers.
ii) Automatic segmentation algorithms are a preferred choice as they segment
the structures of the objects without any human intervention. They are
preferred if the tasks need to be carried out for a large number of images.
iii) Semi-automatic algorithms are a combination of automatic and manual
algorithms. In semi-automatic algorithms, human intervention is required in
the initial stages. Normally, the human observer is supposed to provide the
initial seed points indicating the ROI. Then the extraction process is carried
out automatically
B) Based on Pixel relationship
 Pixel similarity relationships with neighboring pixels.
 The similarity relationships can be based on color, texture, brightness, or any
other image statistic.
i) Contextual (region-based or global) algorithm
 Contextual algorithms group pixels together based on common
properties by exploiting the relationships that exist among the
pixels. These are also known as region-based global algorithms.
 In region-based algorithms, the pixels are grouped based on some
so of similarity that exists between them
ii) Non-contextual (pixel-based or local) algorithms.
 Non-contextual algorithms or pixel based or local algorithms ignore
the relationship that exists between the pixels or features.
 To identify the discontinuities that are present in the image such as
isolated lines and edges. These are then simply grouped into a
region based on some global-level property.

5.1.3 Detection of Discontinuities.

3. Explain types of grey level discontinuities in a digital image?

i) Point detection
 An isolated point is a point whose grey level is significantly different
from its background in a homogeneous area.
 The mask is superimposed onto an image and the convolution process is
applied.
 The response of the mask is given as

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 2

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

 Where the f values are the grey level values of the pixels associated with
the image.
 A threshold value T is used to identify the points. A point is said to be
detected at the location on which the mask is centered if R≥ T, where T is
a non-negative integer.

ii) Line Detection

 Four types of masks are used to get the responses, that is, R1, R2, R3,
and R for the directions vertical, horizontal, +45°, and -45°, respectively.
 These masks are applied to the image.
 The response of the mask is given as

 R1 is the response for moving the mask from the left to the right of the
image.
 R2 is the response for moving the mask from the top to the bottom of the
image.
 R3 is the response of the mask along the +45° line
 R is the response of the mask with respect to a line of 45°.

5.2 Edge Detection

5.2.1 Stages in Edge Detection
4. Explain stages in Edge detection.

 Detect sharp changes in image brightness which can capture the important events
and properties.
 The edge detection process are
i) Filtering
 It is better to filter the input image to get maximum performance
explicitly or implicitly.
 It involves smoothing, where the noise is suppressed without
affecting the true edges.
 Uses a filter to enhance the quality of the edges in the image.
Normally, Gaussian filters are used as they are proven to be very
effective for real-time images.

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 3

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

ii) Differentiation
 This phase distinguishes the edge pixels from other pixels.
 The idea of edge detection is to find the difference between two
neighborhood pixels.
 If the pixels have the same value, the difference is 0. This means
that there is no transition between the pixels.
 The non- zero difference indicates the presence of an edge point.
iii) Localization.
 Detected edge are localized
 The process involves determining the exact location of the edge.
 Edge thinning and linking steps to ensure that the edge is sharp
and connected.

5.2.2. Types of Edge Detectors

5. Explain different types of Edge detectors.

i) Derivative filters use differentiation technique to detect the edge.

ii) Template matching filters use template that resemble the target shapes and match
with image.
iii) Gaussian derivatives are very effective for real time images and are used along with
derivatives filters
iv) Pattern fit the surface is considered as a topographic surface with the pixel value
representing altitude

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 4

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

5.2.3 First Order Edge Detection Operators

6. Apply First order edge detection operator on the pixel in the following image.

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 5

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 6

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

5.2.4 Second Order Derivatives Filters

7. Explain Secord order derivative in image segmentation.

 Edge pixel is present at the location where second derivatives is zero

 It is like zero crossing, which can be observed as a sign change
i) Laplacian operator:
 Generate Mask
 Apply the Mask
 Detect the zero crossing(pixels in a
neighboring differ from each other pixel in
sign)
 Laplacian Masks are

0 1 0 1 0 1
1 -4 1 0 -4 0
0 1 0 1 0 1

ii) Laplacian of Gaussian( Marr-Hildrith) operator:

 Generate the mask and apply LoG to the image
 Detect the zero crossing.
iii) Difference of Gausioan (DOG) filter:
 Generate the mask and apply DoG to the image.
 Detect the zero crossing and apply the threshold
to suppress the weak zero crossing
 Display and exit
iv) Canny edge detection:
Step1:
 Convolve the image with Gaussian Filter
 Compute the gradient of resultant smooth image
 Store the edge magnitude and egde orientation in two
separator arrays.

Step 2:

 Thinning of edge by a process known as Non

maxima suppression.

Step 3:

 Apply Hysteresis Thresholding (objective is that

only a large amount of change in gradient
magnitude)

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 7

21CS63 COMPUTER GRAPHICS AND FUNDAMENTALS OF IMAGE PROCESSING 23-24

Question Bank

1) What is formal definition of segmentation?

2) What are the different way in which segmentation algorithm can be
classified
3) List the edge operator.
4) What is meant by first and second order derivatives of an image?
5) What is the problem of second order of second order derivatives?
6) Consider a one dimensional image f(x)= [10 10 10 10 10 40 40 40 40 20 20].
What are the first order and second derivatives .Locate the position of edge?

Prof Sophia S, Asst. Prof., Department of CSE, Dr. TTIT, KGF 5- 8