A study and comparison of different image segmentation algorithms

20
1 6/24/22 A Study and Comparison of Different Image Segmentation Algorithms

Transcript of A study and comparison of different image segmentation algorithms

Page 1: A study and comparison of different image segmentation algorithms

1

May 2, 2023

A Study and Comparison of Different Image Segmentation Algorithms

Page 2: A study and comparison of different image segmentation algorithms

2

Preamble

Project Title : Detection of counterfeit Indian currency note Seminar Title : A Study and Comparison of Different Image

Segmentation Algorithms In our project we are dividing an image into 3x3 grid and

extract the required features and compare it with the database. The paper has given vivid description about the different segmentation algorithms, which are used in the applications like pattern recognition and image analysis.

Page 3: A study and comparison of different image segmentation algorithms

3

Agenda

Introduction

Image segmentation techniques

Results and discussion

Conclusion

References

Page 4: A study and comparison of different image segmentation algorithms

4

Section-1 : INTRODUCTION

Page 5: A study and comparison of different image segmentation algorithms

5

Continued..

Page 6: A study and comparison of different image segmentation algorithms

6

Section-2:Image Segmentation Techniques

Page 7: A study and comparison of different image segmentation algorithms

7

Thresholding base Image Segmentation

Image segmentation based on a thresholding is the simplest technique

In this technique we set a threshold value (mostly from the histogram of the image)

Pixel lying above or below can be classify as object and background

This technique convert a gray scale image into binary image

This technique will give good result if background and object has large variation in their intensity value

The disadvantage is that it will not be able to identify multiple object.

Page 8: A study and comparison of different image segmentation algorithms

8

Image segmentation algorithms can be classified into two classes:

• Global segmentation algorithms

• Local segmentation algorithms

Algorithms:

Page 9: A study and comparison of different image segmentation algorithms

9

Region-base Image Segmentation

Region based segmentation can be done in two ways: Region Growing Data Clustering

1. Region Growing: Region growing is simplest in region base image segmentation techniques. In this technique, a seed point is chosen at random, then neighboring pixels are check, with some criterion, to determine whether those neighboring pixels are added to the initial seed points or not.

Page 10: A study and comparison of different image segmentation algorithms

10

2. Data clustering: Data clustering method initially assume whole image as a single cluster and then use mathematics and statistics to create number of clusters within the image.Two types of clustering are possible: Hierarchical clustering Partitional clustering

In the hierarchical clustering, we can change the numbers of cluster during the process.

In the partitional clustering, we must decide the numbers of cluster before processing.

Continued…

Page 11: A study and comparison of different image segmentation algorithms

11

Edge based segmentation techniques are first find the edges by using different-different operators.

Since an object can be represented by its edges. So we can segment the image by simply finding edges in the image. A typical approach to segmentation using edges is:

• compute an edge image from original image• process the edge image for broken edges• transform the result to an ordinary segmented image

by filling in the object boundaries watershed segmentation technique is one more technique

which can be used to process the edge image.

Edge base Image Segmentation

Page 12: A study and comparison of different image segmentation algorithms

12

Section-3: Results and Comparison

Original Images

Page 13: A study and comparison of different image segmentation algorithms

13

Result obtained by Otsu's Algorithm

Page 14: A study and comparison of different image segmentation algorithms

14

Result obtained by K-Mean Clustering technique

Page 15: A study and comparison of different image segmentation algorithms

15

Result obtained by quad tree technique

Page 16: A study and comparison of different image segmentation algorithms

16

Result obtained by delta E technique

Page 17: A study and comparison of different image segmentation algorithms

17

Result obtained by FTH technique

Page 18: A study and comparison of different image segmentation algorithms

18

Used and compared the performance 6 image segmentation algorithms.

We have apply these algorithm on a very simple image (bear) to a very complex image (man woman).

From the simulation results, we can conclude that if test image is simple (one object) than Delta-E perform better as compare to other algorithm.

Although Otsu's and Kmean algorithm perform similar to Delta-E but they falsely consider background as a object.

As the Complexity of input test image increases, for single object segmentation, performance degraded. As we can see from tiger food image. for Complex image, if we run the same code for number of objects, than we might get the good performance and this would be a future work.

Conclusion

Page 19: A study and comparison of different image segmentation algorithms

19

Questions??

Stay Hungry Stay Foolish “Steve Jobs”

Page 20: A study and comparison of different image segmentation algorithms

20