An exemplar-based method for improvement in image completion using Local Binary Pattern

Image restoration or completion is an issue of image processing and machine vision, in which the corrupted regions are reconstructed from safe regions. Exemplar-based completion approaches that are based on patches propose better results compared to pixel-based methods. In this paper, we introduce a novel exemplar-based algorithm to improve the image reconstruction or completion. The proposed method include two kinds of changes. Firstly, the patches texture features are extracted (computation of the local binary algorithm). Secondly, the search region is limited with the purpose of computational complexity reduction and improvement in algorithm operation. The obtained results depict that this approach for improvement in quality of exemplar-based method is efficient