Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
3 views

Unit-6 Image Segmentation

Image segmentation is the process of partitioning an image into homogeneous groups of pixels for analysis and interpretation. Techniques include thresholding, which segments based on pixel intensity, and region-based methods, which group connected pixels with similar properties. Key methods discussed are global thresholding, region growing, and region splitting, each with its advantages and challenges.

Uploaded by

Jasneet Kaur Chhabra
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
3 views

Unit-6 Image Segmentation

Image segmentation is the process of partitioning an image into homogeneous groups of pixels for analysis and interpretation. Techniques include thresholding, which segments based on pixel intensity, and region-based methods, which group connected pixels with similar properties. Key methods discussed are global thresholding, region growing, and region splitting, each with its advantages and challenges.

Uploaded by

Jasneet Kaur Chhabra
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 8

Unit-5 Image Segmentation

Introduction
The main objective of image segmentation is to extract various features of the image which can be merged or split
in order to build objects of interest on which analysis and interpretation can be performed. Image segmentation
represents the first step in image analysis and pattern recognition.
Image segmentation refers to the process of partitioning an image into groups of pixels which are homogeneous
with respect to some criterion. Different groups must not intersect with each other, and adjacent groups must be
heterogeneous. Segmentation algorithms are area oriented instead of pixel-oriented. The result of segmentation is
the splitting up of the image into connected areas. Thus segmentation is concerned with dividing an image into
meaningful regions.

IMAGE SEGMENTATION BASED ON THRESHOLDING


Thresholding techniques produce segments having pixels with similar intensities. Thresholding is a useful technique
for establishing boundaries in images that contain solid objects resting on a contrasting background. There exist a
large number of gray-level based segmentation methods using either global or local image information. The
thresholding technique requires that an object has homogenous intensity and a background with a different intensity
level. Such an image can be segmented into two regions by simple thresholding.

Global Thresholding
Global thresholding is the simplest and most widely used of all possible segmentation methods. Global thresholding
will suffer when pixels from different segments overlap in their use of intensities.
In the global thresholding method, the same threshold value is used for the whole image. Global thresholding is
used in many applications because of the simplicity and low computation cost of segmentation. Let, f(x,y) and
g(x,y) be the input image and segmented image, respectively, then segmentation of the image is achieved as follows:

Accuracy of the segmentation depends on the appropriate selection of the threshold value T. Hence, deciding the
value of T for an image is a challenge. Several methods for deciding the appropriate value of threshold T have been
developed by researchers in recent years, such that the segmentation accuracy could be optimized.
The advantages of thresholding-based segmentation methods are simple to implement and overall the segmentation
process is quick and efficient for images with text.
Thresholding-based segmentation methods perform poorly when the images have lots of variation in the intensities
of pixels.
Global Thresholding using Otsu method
Region Based Segmentation
A region in an image is defined as a group of connected pixels having similar properties. In general, objects also
can be interpreted as a group of connected pixels of similar properties. Hence, regions can be used as a key for
image segmentation. In region-based segmentation, pixels corresponding to an object are grouped together.
Region-based segmentation techniques incorporate two basic similarity measures, these are value similarity (pixels
similar to their neighbors are considered in the region) and spatial proximity (spatially close pixels are considered
in the region). Hence, a region or object in any image consists of spatially close pixels with similar gray values.
Two main techniques of region-based segmentation are region growing and region splitting and merging.
Region Growing
The region growing-based segmentation method is a bottom-up approach that starts from a set of initial pixels
(called seeds). These seeds then grow on the basis of some similarity measures (gray level, color, texture). These
seeds can be selected either manually (based on some prior knowledge) or automatically (depending on particular
application).
A pixel is added to a region if it satisfies three conditions:
1. It has not been assigned to any other region.
2. It is a neighbor of that region.
3. The new region that was formed after the addition of the pixel is still uniform.
There are two basic problems with region growing.
Firstly, the accuracy of segmentation is dependent on suitable seed values; however, it is not a simple task to find
good starting points.
Secondly, in some images color may be the appropriate similarity criterion, while in others the image gray levels
are a better choice.

Region splitting
Region splitting essentially employs a similar philosophy, but is the reverse approach to region growing. In this
case we begin the segmentation procedure by treating the whole image as a single region which is then successively
broken down into smaller and smaller regions until any further subdivision would result in the differences between
adjacent regions falling below some chosen threshold.
This method is applicable to images whose number of rows and number of columns are integer power of 2.
Numerical on all the above topics solved in theory sessions.

Sources: DIP by S Jayaraman et. al. TMH Publication.


Fundamentals of DIP by S. Annadurai et. al. Pearson Publication.
Understanding DIP by Vipin Tyagi, CRC Publication.
IP and PR by Frank Y Shih, IEEE-Wiley Press.
DIP by Rafel C. Gonzalez et. al. Pearson Publication.
Fundamentals of DIP- A practical approach with examples in MATLAB by Chris Solomon et. al. Wiley-Blackwell Publication.
DIP by S Sridhar OXFORD University Press Publication.

You might also like