Abstract

Image inpainting plays an important role in restoration of cultural relics, pictures beautification. Criminisi algorithm creates good results in large-area inpainting. However, it does still have some deficiencies such as over-extending. In this paper, two improved algorithms based on prior knowledge of the boundary had been proposed by simulating the idea of manual repairing. An algorithm, by simulating the strategy that the next inpainted pixel will be near to the prior one, named nearer neighbor first algorithm, can void the random bounding of the to-be-inpainted pixle. Another algorithm, by simulating the strategy that the inpainting process, named no-inpainted first algorithm, will be in multiple directions, can void the inpainting process in a single direction. The results reveal that the neighborhood-first algorithm performs better than Criminsi algorithm in repairing the missing structure while the unrepaired-first algorithm performs better than Criminsi algorithm in repairing the missing texture.

Highlights

  • Image inpainting is widely used in repairing damaged works of art

  • The traditional algorithms can be classified into two categories:(1) small region inpainting methods based on partial differential equations such as BSCB model and CDD model and(2) large region inpainting methods based on exemplar such as Criminisi algorithm, which use the structure and data information to design a priority function to select a best inpainting pixel[1]

  • The deep-learning methods require high computing overhead, while the traditional algorithm is the image inpainting algorithm based on sample block, which is widely used due to its ability to remove and repair large areas of objects, simple implementation and reasonable results

Read more

Summary

Introduction

Image inpainting is widely used in repairing damaged works of art. With the widespread application of digital images, it is necessary to research image repaired of digital images, which enables it to meet the needs of photo beautification and removal of specific objects in film and television special effects production. The existing algorithms include the deep learning-based methods and the traditional image methods. The deep-learning methods require high computing overhead, while the traditional algorithm is the image inpainting algorithm based on sample block, which is widely used due to its ability to remove and repair large areas of objects, simple implementation and reasonable results. The second category improves the patch matching method to reduce the uncoordinated block boundary [2,5,6,7,10,12].The last category improves the speed by improving the search strategy[3,5,8, 9,11,13] Please follow these instructions as carefully as possible so all articles within a conference have the same style to the title page. This paragraph follows a section title so it should not be indented

Exemplar-based image inpainting algorithm
The improved method of neighborhood-first
Results and comparisons
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call