Sie sind auf Seite 1von 9

IMAGE INPAINTING USING COLOR STAMPING, BASED ON

EXEMPLAR TECHNIQUE
A DISSERTATION
submitted in partial fulfillment of the
requirement for the award of degree of
MASTER OF TECHNOLOGY
in
COMPUTER NETWORK AND ENGINEERING
by
POOJA SINGH
(En.No.- GE-12792015)



DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
GRAPHIC ERA UNIVERSITY, DEHRADUN
DEHRADUN 248002 (INDIA)
JULY 2014






CANDIDATES DECLARATION
I hereby certify that the work which is being presented in the Dissertation entitled IMAGE
INPIANTING USING COLOR STAMPING BASED ON EXEMPLAR TECHNIQUE in
partial fulfillment of the requirements for the award of the Degree of Master of Technology in
Computer network and Engineering and submitted in the Department of Computer Science and
Engineering of the Graphic Era University, Dehradun is an authentic record of my own work
carried out during a period from August, 2013 to July, 2014 under the supervision of Manish
Sharma, Assistant Professor, Department of Computer Science and Engineering of the Graphic
Era University, Dehradun.
The matter presented in this dissertation has not been submitted by me for the award of
any other degree of this or any other Institute/University.
(POOJA SINGH)
This is to certify that the above statement made by the candidate is correct to the best of
our knowledge.

Signature Head of Department (Manish Sharma)
Supervisor

The Viva-Voce examination of Pooja Singh has been held on


Signature of Internal Examiner Signature of External Examiner


566/6, Bell Road, Clement Town, Dehradun, Uttarakhand, Web Site : www.geu.ac.in





ABSTRACT
Image inpainting is the process of filling in missing parts of damaged images based on
information gathered from surrounding areas. Inpainting techniques have found widespread use
in applications that include, but are not limited to error recovery, red-eye removal, multimedia
editing. Due to an extensive research in the field of digital inpainting and to rapid advances in
technology, new and improved inpainting techniques are continuously proposed. In this thesis,
we introduce a novel algorithm for digital inpainting that attempts to replicate the basic
techniques used by professional restorators but in a more efficient manner. An efficient
algorithm known as the color stamping algorithm, has been proposed in this thesis. In this
algorithm the color is propagated into the target region pixel by pixel method is more efficient as
it does not require any iteration to complete the filling process. The algorithm automatically does
this in a way that it looks reasonable to the human eye. Details that are hidden/occluded
completely can be recovered by this method. Therefore the objective for color stamping is not to
recover the original image, but to create some new image that has a close resemblance with the
original image. The actual color values are computed using exemplar based technique. In this
project a simultaneous structure and texture propagation has been achieved by a single efficient
algorithm, computational efficiency is achieved by a pixel based stamping method. A number of
examples on real and synthetic images demonstrate the effectiveness and the quality of the
algorithm in removing large objects and occlusion.
Keywords: object removal, image inpainting, Exemplar, texture synthesis.






ACKNOWLEDGEMENTS
First of all, I would like to express my deep and sincere gratitude to my dissertation supervisor
Mr. Manish Sharma, for his kindness, encouragement, continuing guidance, and valuable support
during my research work. His broad knowledge and critical thinking have been of great value for
me.
His valuable suggestions and opportune guidance enabled me to observe the different aspects of
the research in a new glow. From him, I learned valuable lessons in doing research. It was a great
honor for me to work with him.
I am also be thankful for all my faculties for their valuable advice, suggestions and also for their
efforts and time spent in providing me perceptive comments which have greatly improved the
quality of my research work.
I would like to thank my family that has encouraged and supported me all through my M.Tech.
Journey. Finally, I am also very grateful to all the people who helped me directly or indirectly all
through my research.



Date: (Pooja Singh)








Candidates Declaration

Abstract I
Acknowledgements Ii
Table of Contents Iii
List of Abbreviations V
List of Figures Vi
List of Tables vii

1. Introduction
1.1 Overview 1
1.2 Applications of inpainting 2
1.3 Tool 2
1.4 Motivation 2
1.5 Problem statement 2
1.6 Thesis objective 3
1.7 Organization of thesis 3
2. Background and literature survey
2. Literature survey 4
2.1 Background 4
2.2 Basic inpainting techniques 5
2.2.1 Texture synthesis technique 5
2.2.2 PDE based technique 7
2.2.3 Exemplar based technique 8
2.3 Inpainting algorithms 10
3. Methodology
3.1 Implementation methodology 17
3.2 Proposed algorithm 19
4. RESULT
4.1 Experimental results and comparison 22
4.2 image00.jpg results 23
4.3 dbird.jpg result 24
4.4 fish.jpg result 25
4.5 sea.jpg result 26
4.6 images02.jpg result 27
4.7 Graphical representation of proposed and exemplar technique 28

5. Conclusion & scope for future work
5.1 Conclusion 29
5.2. Scope for future work 29
References30















LIST OF ABBREVIATIONS
TSBI Texture synthesis based inpainting
PDE Partial Differential Equation
MSE Mean Square Error
PSNR Peak Signal to noise ratio

GUI Graphical user interface

MATLAB Matrix laboratory






LIST OF FIGURES
Figure no.Captionpage no.
Figure 1.1 Removing large object from an image 9
Figure 2.1 Image restored with our algorithm 12
Figure 2.2 procedures basic texture synthesis 13
Figure 2.3 pixels based and patch based differentiation 14
Figure 2.4 PDE based example 15
Figure 2.5 Structure propagation by exemplar-based texture synthesis 17
Figure 4.1 Original image of (image00.jpg) 23
Figure 4.2 binary mask region23
Figure 4.3 Masked RGB image23
Figure 4.4 Exemplar result 23
Figure 4.5 proposed result23
Figure 4.6 Original image of (dbird.jpg)24
Figure 4.7 Binary mask region24
Figure 4. Masked RGB image24
Figure 4.9 Exemplar result24
Figure 4.10 proposed result 24
Figure 4.11 Original image of (fish.jpg) 25
Figure 4.12 binary mask region 25
Figure 4.13 masked RGB image 25
Figure 4.14 Exemplar result 25
Figure 4.15 proposed result 25
Figure 4.16 original image of(sea) 26
Figure 4.17 binary mask region 26
Figure 4.18 masked (RGB) 26
Figure 4.19 exemplar result 26
Figure 4.20 proposed result 26
Figure 4.21 original image of (images02) 27
Figure 4.22 binary mask region 27
Figure 4.23 masked RGB 27
Figure 4.24 exemplar result 27.
Figure 4.25 proposed result 27
Figure 4.26 graphical representation between both MSE values 28
Figure 4.27 graphical representation between both PSNR values 28
LIST OF TABLES
Table no. CaptionPage no.
Table 1.1Selected inpainting algorithm 11
Table 3.2 Table for comparison MSE and PSNR values 22

Das könnte Ihnen auch gefallen