Abstract
Filling holes in an image is achieved in a manner similar to peeling the onion. The order of filling affects the image inpainting results, especially concerning the content of complex images. When high-resolution images are used to extract edge information, they are susceptible to high-frequency information, such as complex textures and noise. Furthermore, edge information is extracted in different resolutions, while the main contour information of the image can be obtained more easily. In this paper, multi-resolution information is used to prioritize which target patches in an image to fill, which helps to elucidate the optimal sequence for image repair. Multi-resolution images provide more information than single-resolution images, and similar patches are computed on multi-resolution images to obtain multiple candidate patches. Similar patch calculations use a variety of information on colors, gradients, and boundaries to more accurately search for similar patches. We chose the most reasonable candidate patch by means of the structural similarity index measure (SSIM). When pasting the patch to fill the target region, we used graph cut technology to eliminate blockiness. Compared with the state-of-the-art repair algorithm, the experimental results prove that the proposed repair algorithm can repair the image very well.
Highlights
Image inpainting aims to fill corrupted regions such as scratches, damaged regions, dates, and so on
We obtained peak signalto-noise ratio (PSNR), structural similarity index measure (SSIM), and feature similarity index measurement (FSIM) values by testing 30 images. as shown in Fig. 5(a)-5(c), and the average values of PSNR, SSIM, and FSIM are shown in Fig. 5(a)-5(c), the PSNR, SSIM and FSIM values of the inpainting results using multi-resolution information to compute priority are greater than those of the single resolution results, it demonstrate that the multi-data term can improve the inpainting effect
To obtain the best valued of α, we tested the accumulative values of the PSNR, SSIM, and FSIM using 30 images
Summary
Image inpainting aims to fill corrupted regions such as scratches, damaged regions, dates, and so on. Bertalmio et al firstly proposed inpainting in Reference [1] His approach was a diffusion-based method, in which the filling process involves diffusing information by isophotes from the known regions to the target regions. H. Liu et al.: Exemplar-Based Image Inpainting With Multi-Resolution Information and the Graph Cut Technique and searches for the matching patch in the whole image. The exemplar-based technique [6] has an advantage for inpainting large region with missing texture information It has several drawbacks: (1) the filling order affects the repair results due to the high-frequency components of full-level images; (2) the best sample patch randomly selected is not necessarily the most suitable patch, which is located by Euclidean distance on the color component; and (3) copying and pasting patches may generate the appearance of visual artifacts such as block effect (blockiness).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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
