Hybrid Patch Propagation and Diffusion Technique for Image Inpainting
This paper deals with anovel hybrid technique based onpatch propagation and diffusion method for image inpainting. The presented technique is used to reconstruct a damaged image. Image inpainting is a method to fill the hole with the best plausible, which is created due to damage. The novel hybridization technique of diffusion-based and exemplar-based is presented to overcome the existing problem of inpainting. The method is tested on the image dataset of TUM-IID. The performance of present method is measured using quality factor (QF) analysis, peak signal-noise ratio (PSNR), and comparing similarity by structure similarity index (SSIM). Result demonstrates that the proposed calculation performed better contrast with the existing exemplar-based technique.