Yasmine ElGlaly

Supervised by: Prof. Essam Hamed Atta, Prof. Taimor M. Nazmy, Prof. Beih Sayed El Desouky

Ain shams University, Faculty of Computer & Information Sciences, Computer Science Department

 

Image inpainting is the process of filling in missing parts of damaged images based on information gathered from surrounding areas. In addition to problems of image restoration, inpainting can also be used in wireless transmission and image compression applications. In this thesis, we have developed an automatic digital inpainting system that enables the user to choose between two complementary approaches. The first is based on the solution of partial differential equation of isophote intensity to fill-in missing portions in the region under consideration, while the second is based on texture inpainting. The filling-in process is automatically done in regions containing completely different structures, textures, and surrounding backgrounds.

 

We have also presented an improved inpainting method based on the exemplar-based image inpainting technique. The developed method enhances the inpainting robustness and effectiveness by including image gradient and second derivative information during the inpainting process.

 

Finally, we validated our developed method and compare the results with previous methods. Our results show that the developed algorithm can reproduce texture and at the same time keep the structure of the surrounding area of the inpainted region. The method proved to be effective in removing large objects from an image, ensuring accurate propagation of linear structures, and eliminating the drawback of "garbage growing" which is a common problem in other methods.

For more details, read this.