An Improved Alogrithm of MRF-Based Exemplar Inpainting
Junxi Sun · Journal of Changchun University of Science and Technology · 2011
For eliminating cumulative error in unknown area by the means of exemplar-based image inpainting,the spa-tial geometrical information of isophotes is introduced in the image inpainting algorithm of basing on MRF model ex-emplar.Priority function of algorithm for exemplar-based image inpainting is redefined and updating rule of confidence term is modified.The experimental results are given,which show that the garbage and small structure gaps due to the cumulative error can be avoided effectively and favourable textural and structural information can be preserved in the proposed method.