A Novel Color Image Inpainting Guided by Structural Similarity Index Measure and Improved Color Angular Radial Transform.

摘要: ABSTRACTImageinpaintingisconsideredasapredictiveprocesstocom-pute the missing image data without introducing undesirableartifacts. Most of existing methods in literature workvery well for small regions but introduce blur large holes.Inthispaper, weproposeanovelunifiedframeworkforaffineand flip invariant inpainting color images. The proposedmethod combines structural similarity index measure, an im-proved version angular radial transform, frequencydomain-based registration and Dr. Kekre’s LUV spacebased blending. It searches best candidate that aresimilar to neighbourhood domain eitherinthesameimageorinthelargedatabaseintermsofitsstruc-ture, texture simultaneously thereby improving theprediction accuracy. Experimental results indicate perceptu-ally satisfactory results.Keywords— affine inpainting, struc-turalsimilarityindexmeasure,colorangularradialtransform,frequency-based registration, space based imageblending1. INTRODUCTIONImage is art recovering original imagefrom images which are generally incomplete due variousfactors, including degradation ageing, damage towear tear, details occlusion andloss transmitted through a noisy communica-tion channel. In such situations, there need predictthe information undesir-able artifacts. To observer, inpainted must lookauthentic bearing any trace being tampered with.A number techniques have been proposed inthe literature. They can be divided into many classes namelymethods on convolution using kernel [1], neighbour-hood diffusion isophotes direction [2], Total Variation(TV) model [3], Curvature Driven Diffusion (CDD)