You are on page 1of 7

ABSTRACT

Image fusion is an approach to acquire a composite image with all objects in


focus. The cue point of image fusion is to progress and construct activity level
measurement to evaluate the quality of source images. The Dense Scale Invariant
Feature Transform (SIFT) method is proposed to obtain a multi-focus image
fusion. The main creativity of this effort is that it manifest the greatest
prospective of image local features such as the Dense SIFT used for image fusion.
It also consists of matching densely sampled SIFT features between the two
images, while conserving spatial discontinuities. The use of dense SIFT features
allows rugged matching across different spot or object appearances. The
discontinuity-conserving spatial model allows matching of objects located at
different parts of the locations. Especially, the local feature descriptor can not
only be commissioned as the activity level measurement, but also be used to
match the mis-registered pixels between numerous source images to improve the
clarity of the fused image. In our algorithm, via the sliding window technique, the
dense SIFT descriptor is first used to estimate the activity level of source image
patches to obtain an initial decision map, and then the decision map is enhanced
with feature matching and distinct focus measure comparison. Fusion is done by
comparing the decision maps.

iii
ACKNOWLEDGEMENT

First of all I thank the Lord Almighty for his abundant grace and
countless blessings in making this work a great success.

We extend our gratitude to our Secretary Thiru. R. Mahesh Kumar,


M.B.A., for making us march towards the glory of success.

We also express our sincere thanks to our beloved Principal


Dr. AnantAchary, M.Tech., Ph.D., and our Dean Dr. K. Sarukesi, Ph.D.,
for all the facilities offered.

We are very much grateful to pay our sincere respect and thanks to
our Head of the Department, Dr.P.Subathra, M.E.,Ph.D., who is the source
of inspiration, guidance and help.

Our sincere thanks also go to our project coordinators


Mrs.M.Chengathir Selvi, M.E., Ph.D and Mr.D.Vendan, M.Tech., Ph.D
for offering us the valuable information in the respective field.

We are highly indebted to our project guide Mrs.E.Vakaimalar,


M.E.,Ph.D., for her guidance and constant supervision as well as for
providing necessary information regarding the project and also for her
support in completing the project.

We would like to express our gratitude towards our parents and staff
members of Kamaraj College of Engineering and Technology for their kind
cooperation and encouragement which helped us in completion of this
project.

iv
TABLE OF CONTENTS

CHAPTERNO. TITLE PAGE NO.


ABSTRACT iii
LIST OF FIGURES vii
LIST OF TABLES viii
LIST OF ABBREVIATIONS ix
1 INTRODUCTION 1
1.1 Image Fusion 1
1.1.1 Methods 2
1.1.1.1 Pixel Level Fusion 2
1.1.1.2 Feature Level Fusion 3
1.1.1.3 Decision Level Fusion 3
1.1.2 Domains 3
1.1.2.1 Spatial Domain Fusion 3
1.1.2.2 Frequency Domain Fusion 5
1.1.3 Types of source image 7
1.1.4 Benefits 8
1.1.5 Applications 9
1.2 Dense SIFT 10
1.3 Performance Metrics 11
2 LITERATURE SURVEY 14
3 SYSTEM DESIGN 20
3.1 Existing System 20
3.2 Proposed System 20
3.3 System Architecture 21
4 SYSTEM REQUTREMENT 23
4.1 Hardware Requirements 23
4.2 Software Requirements 23
4.3 Software Specification 23

v
5 SYSTEM IMPLEMENTATION 27
5.1 Input Module 27
5.2 Histogram and Pixel Wise SIFT Image 27
5.3 Dense Scale Invariant Feature
27
Transform
5.4 Decision Map 28
5.5 Fused Image 29
6 RESULTS AND DISCUSSION 30
6.1 Experimental Results 30
6.2 Performance Metrics 34
7 CONCLUSION AND FUTURE WORK 35
7.1 Conclusion 35
7.2 Future work 35
APPENDICES 36
Sample Code 36
REFERENCES 57
PUBLICATIONS 59

vi
LIST OF TABLES

TABLE NO. TABLE NAME PAGE NO.


Table 6.1 Performance Metrics 34

vii
LIST OF FIGURES

FIGURE NO. FIGURE NAME PAGE NO.


Figure 1.1 Levels of Abstractions 2
Figure 1.2 Types of Source Image 7
Figure 3.1 System Architecture 21
Figure 5.1 Dense SIFT Image 28
Figure 5.2 Fused Image 29
Figure 6.1 Source Image 1 30
Figure 6.2 Source Image 2 30
Figure 6.3 Histogram 31
Figure 6.4 Dense SIFT Image 31
Figure 6.5 Fused Image 32
Figure 6.6 Samples 33

viii
LIST OF ABBREVIATIONS

DSIFT Dense Scale Invariant Feature Transform


PSNR Peak to Signal Noise Ratio
JPEG Joint Photographic Expert Group
PNG Portable Network Graphics
SWT Sliding Window Technique
HOG Histogram of Oriented Gradients

ix

You might also like