Essay: Examplar-based inpainting

This system merges two different features of image processing. These are the image inpainting and video watermarking. The video inpainting introduces a novel framework for examplar-based inpainting. This examplar-based inpainting consists in performing first the inpainting on a rough version of the input image. The main advantage of this form of approach is that it is very easy to inpaint low-resolution pictures than high-resolution ones. The computational complexity and visual quality is covered in this manner. However, with this less sensitive to the inpainting’s method parameter setting, we can be inpainted several times using different configurations low-resolution input picture can. After that using inpainted video watermarking is done that is inpainted video is used for applying watermarking on it. The practice of video watermark has a distinguished history. In this project we have to apply QR based video watermarking on the inpainted video for that input is given by the user.
KEYWORD: Examplar-based inpainting, Quick Response (QR) Code, singular value decomposition (SVD)
I. INTRODUCTION:
Today, there are number of researches are performing on the images and for that research field, image has become useful phenomenon. For capturing memories, the images are only used in old days. But now images have changed their face.  There  may  be  two-dimensional,  or  three-dimensional  images.  The  photos  or  videos  may  be  captured  by  using digital or analogous cameras of other capturing devices. Today, images can be very helpful for encryption, processing, authentication, sharing etc. purpose. The main purpose of videos or images is to store the memories of some important moments.  In  image,  due  to  extra  part  or  distortion  sometimes  useful  images  get  discarded  or  deleted.  For  restoring image or painting seems as natural as its original version a super resolution (SR) algorithm is very useful for guessing  and filling in the lost image information. By using image inpainting on video we have to remove the target object form the video file. To recover the selected part form the image we need to select the area to inpaint and then pass particular frame to  the  super  resolution  algorithm.  For  removing  the  objects  which  are  not  required,  the  Exemplar-based inpainting  is  very  useful.  There  is  more  efficient  algorithm  is  a  Super-resolution  algorithm  which  can  produce  very efficient. Output as compare to other inpainting algorithm, Inpainting  used  for  scratch  removal.  The  removal  of  object,  text  and  other  automatic  modification  of  images are  include  in  the  next  applications.  To  remove  objects  from  images  and  fill  the  hole  by  taking  information  from neighborhood is the process of object removal. There are number of  methods to removing the object from the image. Even  we can remove the object by using paint software. Also there are other software’s which can easily remove the object from image.
This technique is observes the image and start replacing the selected part from left to right. Here some automatic image inpainting methods are there. This play an important role by the image inpainting technology in computer graphics and has many applications such as old films renovation, remove object from photos and videos. This method try to replace the  unwanted  part  of  image  by  using  different  restoring  techniques.  So  by  using  image  inpainting  method  one  can replace the original image part which are related to the other parts of images or it may be different than the image part.
To  improve  the  image  quality  from  remove-undesired  object,  there  varies  the  reason  behind  region  completion varies. The object removal starts with find the undesired points or pixels, making the area where the object previously occupies a hole. These hole can be fill up with the help of graphical holes filling technique.  From the  LR  HR  method  we  are using the dictionary  method in  which pixels can be obtained from the dictionary. The  image  pixels  can  be  removed  with  exact  matching  pixels  from  the  dictionary  database.  The  super  resolution method use the Bergman iterations technique to recover the parts from the image. The HR image is estimated with the height and width of image. There is proposed a new regularization method which is based on multi scale morphological filters.
II. LITERATURE SURVEY
In this section we discussed about literature survey on inpainting and watermarking of video.To get idea about different kinds of inpainting and watermarking technique and their work we have made some literature survey as follow:
The diffusion based or the  exemplar based techniques are two techniques of this system. It leads to the development of hierarchical approach of super-resolution based inpainting because of it is having some limitation.
In [1], M. Bertalmio, G. Shapiro, V. C asella’s, and C. Baluster shows for filling the some loosed portion of the image that image inpainting is only used. Bu t for high quality images this meth o d is not suitable. It uses patch based inpainting. The area at which the inpainting algorithm is to b e apply is selected here manually by the user. Here this area is marked as the sigma notation. Masking on image is denoted by sigma. In this by using Eros and lounges algorithm masking is removed. For filling the losses inside the image this method is responsible but this feeling is not reasonable.
In [2], M. Ashikhmin introduce a method to provide the synthesis of natural textures only. A simple texture synthesis algorithm that is well-suited for a specific class of naturally occurring textures. This class includes quasi-repeating patterns consisting of small objects of familiar but irregular size, such as flower fields, pebbles, forest undergrowth, bushes and tree branches.
In[3], C. Ballester, V. Caselles, J. Verdera, M. Bertalmio, and G. Sapiro introduce A variational approach for filling-in regions of missing data in graylevel and colour images. The approach is based on joint interpolation of the image gray-levels and gradient/isophotes directions, smoothly extending in an automatic fashion the isophote lines into the holes of missing data.
In [4], M. Bertalmio, L. Vese, G. Sapiro, and S. Osher present an algorithm for the simultaneous filling-in of texture and structure in regions of missing image information. The basic idea is to first decompose the image into the sum of two functions with different basic characteristics, and then reconstruct each one of these functions separately with structure and texture filling-in algorithms.
In [5], Sylvain Lefebvre, Hugues gives a texture synthesis scheme based on neighbourhood matching, with contributions in two areas: parallelism and control. This scheme defines an infinite, deterministic, a periodic texture, from which windows can be computed in real-time on a GPU. It attain high-quality synthesis using a new analysis structure called the Gaussian stack, together with a coordinate up sampling step and a subpass correction approach.
III. PROPOSED APPROACH FRAMEWORK AND DESIGN
A. Problem Definition
Video inpainting by using image inpainting is one of the challenging process. Previously different methods are available for image inpainting but not for video. There should be some system that helps us to remove all unwanted parts from video. Video should contain only that chicks which we want. Also with the video inapainting their so system that provides us help so that we can also provide so form of security. This can be implemented using watermarking.
B. Objective
• To remove the missing holes from the image.
• To maintain the image quality after inpainting.
• To remove the unwanted part from the video file.
•To apply video watermarking on the inpainted video file.
C. Proposed System
The whole system contain three different modules.
1. User module
2. Exemplar Method
3. Watermarking Module
In the user module the input video is taken from user with inpainting input. After that, video splitting is done on that video to obtain the images from video. These images are maintained into one directory for convention. On this directory the Exemplar method is apply to remove the unwanted part from that video which is user want to remove it from final video.
Fig 1: System Architecture
When whole Exemplar Algorithm done its working, the splitted video is combined to form one single video. This video is then given as input to the Watermaking. In this module the watermarking is applied to the inpained video for that the input for watermaking is again taken from user. User provides the part which is to be watermarked. Using this input the watermarking is done.
D. Mathematical Model
1. User Module:
Set (C) = {c0, c1, c2, c3}
C0= Select the video to inpaint.
C1=select the object to be removed from the image.
C2=Check for object removal.
C3=get Output video.
2. Exemplar Method:
Set (T) = {c1, c2, d0, d1, d2, d3}
d0=Give linear structure higher priority.
d1=Assign each pixel priority value.
d2=Do the structure propagation.
d3 = Inpaint the image.
3. Video Splitting and Combining Module:
Set (S) = {c1, c2, d3, s0, s1, s2}
s0= Get the input video from user.
s1= Split the video into number of images.
s2= Combine the video after inpainting.
4. Watermarking Module
Set (W) = {w1, w2, w3, s2}
W1= select company logo.
W2=select company name.
W3= create watermarked video
Union and Intersection of project:
Set (P) = {c1, c2, d0, d1, d2, d3}
Set (t) = {c1, c2, d3, s0, s1, s2}
Set P intersect Set T={c1, c2,d3}
Fig2:  Set(P intersection t)
E. RESULT
1. Registration Page
This is the registration window, here user can register and get his username and password. If username already present then system just validate and display the message to respective user. In this window, email and mobile number is accepted only when b oth are valid. Register usercan get the OTP as e-mail on registered mai l id. User can get OTP at every login.
Fig 3: Registration Page
2. Login window
This window show login frame, here registered user can login by using username and password generated at registration window. User can receive an email on his registered mail id which contains the OTP f or the login. This OTP is only for authentication purpose only. After entering correct OTP which was sent on mail by the system user can access the next window. On this window one link also available for new user, they can click this link and register themselves
Fig 4: Login page
3. File Menu
After login user can see this window, in file menu of this window multiple menu items. User can select his video for inpaint, also he can op en an i mage for inpaint directly. He can save image,also he can click on save as menu item. Here need to select video for frame extraction. After
clicking on op en video menu item, browse video for inpaint will be opened.
Fig 5: File Menu
IV. CONCLUSION
In  our  study  we  are  studied  different  method  of  image  inpainting  by  using  image  inpainting  and  applying  video watermarking on it. For giving better output using this inpainting method is and by finding exact match of the pixel, it overcomes the limitations of the all existing  work done by previous authors.  For filling  the gaps in the image it  uses the  super  resolution  algorithm.  Here  it  can  result  in  better  and  efficient  output  because  multiple  Image  inpainting techniques combine.A novel inpainting app roach has been presented in this proposed system. The input picture is first down sampled and several inpaintings are performed. The low-resolution inpainted pictures are combined by globally minimizing an energy term. Once the combination is completed, a hierarchical single image super resolution meth o d is applied to recover detail s at the native resolution. Experimental results on a wide variety of images have demonstrated the effectiveness of the proposed method.

;