Essay:

Essay details:

  • Subject area(s): Engineering
  • Price: Free download
  • Published on: 7th September 2019
  • File format: Text
  • Number of pages: 2

Text preview of this essay:

This page is a preview - download the full version of this essay above.

Abstract— Image inpainting refers to an approach used for filling in the missing or corrupted regions of an image. In simple words, Image inpainting or completion is a procedure to restore a damaged image. Inpainting approaches plays a vital role in numerous applications like object elimination, scratch removal, Image restoration etc.Recently various approaches have been proposed. In the past, this problem has been courteously approached by two classes of algorithms: (i) texture synthesis algorithms for creating large image regions from sample textures, and (ii) inpainting techniques for filling in small image gaps. In this paper a new algorithm is proposed for removing large objects from images. Recently various strategies have been proposed different exemplar based image inpainting algorithms to recover the structure of damaged images. We introduce an unusual and efficient exemplar based Image Inpainting Algorithm with investigating natural image patches. This paper will discuss inpainting techniques with respect to restoration of image. The goal of any digital image inpainting algorithm is to rebuild the missing or impaired regions in a visually plausible way.

Keywords— Object Removal, Image Inpainting, Simultaneous Texture and Structure Propagation, exemplar-based inpainting.

Fig1: Object Removal

                  Mrunal Mate, Bhanudas Thakare

Dept. of Computer Engineering,

Sanjivani COE, Kopargaon

Affiliated to Savitribai Phule Pune University

Email:- [email protected]

1. Introduction

This paper reveals an unusual algorithm for removing unwanted, may be large or small objects from digital photographs and replacing them with visually feasible backgrounds. The filling-in of missing region in image is known as Image Inpainting.

Inpainting is the creative activity of modifying an image or video in a form that is not easily observable by an ordinary observer. Image Inpainting has become thoroughly made researchers more curious to do an research in Image processing. The modification of images in a way that is no detectable for an observer who does not know the original image is a practice as artistic creation. The motive was simple, to bring medieval pictures, and then to make this pictures look good by repairing it completely. We can scan these photographs or take snaps and then system will do the rest things. This practice is called ‘retouching’ or ‘inpainting’. We can obtain scratch free (from damaged photographs), high resolution (from faded photographs), image coding and transmission (e.g. Recovery of missing blocks), text removal (names, dates and time etc.)  and likewise. Also image inpainting has been widely researched in the applications of digital effect (i.e. object removal, image supervising, image rescaling).

The schemes that are proposed for image inpainting can be classified into two categories:

• Texture oriented

• Structure oriented

In the previous work, several investigators have used texture synthesis as a way to fill large image regions with pure textures-repetitive two dimensional textural patterns. Example of texture synthesis is the exemplar based technique. This approach effectively produces new texture by sampling and considering colour values from the source. Though the methods of texture synthesis are effective, they have problems in filling holes in photographs of real world images having of composite textures. On the other hand, the structure oriented strategy generates information of the missing region by considering linear structure (called isophotes in the inpainting literature) into the target region with contrast to diffusion. They are motivated by the partial differential equations. Their drawback is that the diffusion process creates some blurriness, which becomes hard when filling larger regions. The technique introduced in this paper uses the strengths of both approaches into a single efficient algorithm. The main attention is on linear structures in structure oriented technique. But, linear

structures in the target region only focus the fill order of is an exemplar based texture synthesis algorithm. The result is an algorithm that has the efficiency and qualitative performances of exemplar based texture synthesis, but which also make use of the image constraints forced by surrounding linear structures. There are many objects and uses of this method. In photography and filmography is used for film renewal, to reverse the drop (e.g., crashes in photographs or scrapes and dust spots in film). It is also used for red-eye reduction, the printed date from photographs and removing objects to original effect. This system can be used to include the lost blocks in the coding and transmission of images, for example, in a flowing video. It can also be used to remove symbols in videos. Also alphanumeric restoration of older paintings for preservation  purposes, reestablishing aged or damaged photographs and films, text deletion and object removal in images for different effects, alphanumeric zooming and edge-based image coding.

2. LITERATURE SURVEY

A. Super Resolution Based Inpainting

This paper provides a unique structure for exemplar-based inpainting. It involves in performing initial the inpainting on a part of the input image. A hierarchical super-resolution formula is then used to obtain details on the missing areas. The advantage of this procedure is that it is easier to inpaint low-resolution photos than high-resolution. The effect is each in the form of computational quality and visual quality. The low-resolution input image is unpainted many times with totally different configurations, to be less sensitive to the parameter setting of the inpainting technique. Results are efficiently combined and details are recovered by a single-image super-resolution formula.

B. Region filling and Object Removal by Exemplar–Based Inpainting

A new formula is proposed for removing large objects from digital images. The main task is to fill the holes in a

picture. In the past, this drawback has been solved by 2 categories of algorithms

1) Texture synthesis algorithms

for producing large image regions from sample textures and

2) Inpainting  techniques

for filling in small image gaps.

This paper represents a unique and efficient formula that results the advantages of those two approaches. A number of algorithms specifically solve this issue for the task of image restoration, where scratches and overlaid text are removed .These image inpainting techniques fill holes in pictures by controlling linear structures (called isophotes within the inpainting literature) into the target region through diffusion.

C. Super-Resolution from a Single Image

Methods for super-resolution will be predominantly divided into two families of methods:

1)  The classical multi-image super resolution

2)  Example-Based super-resolution.

This paper introduced a Unified framework for combining these two families of ways. It is important to learn how this combined approach will be applied to get super resolution from one small image. This study considers various previous and current approaches/strategies in image inpainting. The major roll covered by the study include different approaches for inpainting.

D. Diffusion based Inpainting

Diffusion based Inpainting was the first digital Inpainting approach. Here missing region is filled by diffusing the image information from the known region into the missing region at the pixel level. Basically these algorithms are based on theory of vibrational method and Partial Differential equation (PDE).

E. Texture Synthesis Based Inpainting

Texture synthesis based algorithms are one of the earliest approaches of image Inpainting. And these algorithms are used to fill the missing areas using similar neighbourhoods of the damaged pixels present in an image. The texture synthesis algorithms synthesize the new image pixels. And then try to maintain the local structure of the image. All the earlier Inpainting techniques used these methods to fill the missing region by sampling and copying pixels from the neighbouring area.

F. Exemplar based Inpainting

The exemplar based approach is an important class of inpainting algorithms. And they have shown to be very effective. Normally it consists of two basic steps

1. Priority assignment

2. Selection of the best matching patch.

The exemplar based method samples the best similar patches from the known region, whose similarity is measured by considering certain parameters, and pastes into the target patches in the missing region present in an image. Exemplar- based Inpainting produces the region (target region), that is unknown by the most similar patch in the source region. In accordance with the filling order, the method fills various structures in the missing regions using related information of neighbouring regions. The above mentioned method is an efficient approach for reconstructing large desired regions.

2.  METHODOLOGY

Existing methods can be classified into two main categories. The first category concerns diffusion-based approaches which propagate linear structures or level lines via diffusion based on partial differential equations and variation methods. Unfortunately, the diffusion-based methods tend to introduce some blur when the hole to be filled-in is large. The second family of approaches concerns exemplar-based methods which sample and copy best matches texture patches from the known image neighborhood. These methods have been inspired from texture synthesis techniques and are known to work well in cases of regular or repeatable textures. The first attempt to use exemplar-based techniques for object removal has been reported in. Authors in improve the search for similar patches by introducing an a priori rough estimate of the in painted values using a multi-scale approach which then results in an iterative approximation of the missing regions from coarse to fine levels. In proposed system two main components are the in-painting and the super-resolution algorithms.

More specifically, the following steps are performed:

1. A low-resolution image is first built from the original picture;

 2. An in-painting algorithm is applied to fill-in the holes of the low-resolution picture;

 3. The quality of the in-painted regions is improved by using a single-image SR method.

Fig2:Inpainting Framework

3. SINGLE IMAGE SUPER RESOLUTION ALGORITHM

Super-resolution algorithm Once the inpainting of the low-resolution picture is coBmpleted, a single-image super-resolution approach is used to reconstruct the high resolution of the image. The idea is to use the low-resolution inpainted areas in order to guide the texture synthesis at the higher resolution. As in [11], the problem is to find a patch of higher-resolution from a database of examples. The main steps, illustrated in figure 5 are described below: 1. Dictionary building: it consists of the correspondences between low and high resolution image patches. The unique constraint is that the high-resolution patches have to be valid, i.e. entirely composed of known pixels. In the proposed approach, high-resolution and valid patches are evenly extracted from the known part of the image. The size of the dictionary is a user-parameter which might influence the overall speed/quality trade-off. An array is used to store the spatial coordinates of HR patches (DHR). Those of LR patches are simply deduced by using the decimation factor; 2. Filling order of the HR picture: the computation of the filling order is similar to the one described in Section 3. It is computed on the HR picture with the sparsity-based method. The filling process starts with the patch ψ HR p 10 Olivier Le Meur and Christine Guillemot having the highest priority. This improves the quality of the inpainted picture compared to a raster-scan filling order; 3. For the LR patch corresponding to the HR patch having the highest priority, its K-NN in the inpainted images of lower resolution are sought. The number of neighbours is computed as described in the previous section. The similarity metric is also the same as previous; 4. Weights wp,pj are calculated by using a non-local means method as if we would like to perform a linear combination of these neighbours. However, the similarity distance used to compute the weights is composed of two terms: the first one is classical since this is the distance between the current LR patch and its LR neighbours, noted d(ψ LR p , ψLR p,pj ). The second term is the distance between the known parts of the HR patch ψ HR p and the HR patches corresponding to the LR neighbours of ψ LR p . Say differently, the similarity distance is the distance between two vectors composed of both pixels of LR and HR patches. The use of pixel values of HR patches allows to constraint the nearest neighbour search of LR patches. 5. A HR candidate is finally deduced by using a linear combination of HR patches with the weights previously computed: ψ HR p = X pj'∈'DHR wp,pj × ψp,pj (4) with the usual conditions 0 ≤ wp,pj ≤ 1, and P k wp,pk = 1. 6. Stitching: the HR patch is then pasted into the missing areas. However, as an overlap with the already synthesized areas is possible, a seam cutting the overlapped regions is determined to further enhance the patch blending. The minimum error boundary cut [21] is used to find a seam for which the two patches match best. The similarity measure is the Euclidean distance between all pixel values in the overlapping region. More complex metrics have been tested but they do not substantially improve the final quality. At most four overlapping cases (Left, Right, Top and Bottom) can be encountered. There are sequentially treated in the aforementioned order. The stitching algorithm is only used when all pixel values in the overlapping region are known or already synthesized. Otherwise, the stitching is disabled. After the filling of the current patch, priority value is recomputed and the aforementioned steps are iterated while there exist unknown areas.

Fig: SR-Based Super Resolution Algorithm

3. RESULTS AND ANALYSIS

Super Resolution algorithm comes into picture when the region to be filled is large. Though many algorithms have been researched on image inpainting, this is the best to be used for filling of holes are large. The two types of methods that is, exemplar based inpainting and diffusion based inpainting are combined such that to get results in desired manner. Obviously it is a two-step algorithm. Unlike other, inpainting is applied to the rough part of the image rather than working on the filling of missing areas in original resolution image. According to this algorithm inpainting is not performed directly on full resolution picture but on the image which is smaller than the original, therefore it significantly reduces the computational time needed. Performing inpainting on rough part gives advantages that there is no more existence of noise and also it is very easy. The algorithm does not demand for computational resources always. Thus the algorithm shall not only inpaint an image but also it’ll increase its resolution. Another very important advantage is that the proposed system with its interaction with the system will not need any expertise to interact with it. On national importance it’ll be proved to be a best solution to preserve very old paintings in museums.

Fig: Inpainting Result

4. CONCLUSION

Last decades or two, many inpainting concepts were implemented but this Super Resolution based inpainting is new anatomy proposed to solve problems of image inpainting. Non-parametric patch sampling along with super resolution technique will be the better and give better results than other image inpainting methods.

...(download the rest of the essay above)

About this essay:

This essay was submitted to us by a student in order to help you with your studies.

If you use part of this page in your own work, you need to provide a citation, as follows:

Essay Sauce, . Available from:< https://www.essaysauce.com/essays/engineering/2016-3-14-1457934825.php > [Accessed 23.10.19].