Title :
Object Removal By Cross Isophotes Exemplar-based Inpainting
Author :
Wu, Jiying ; Ruan, Qiuqi
Author_Institution :
Inst. of Inf. Sci., Beijing Jiaotong Univ.
Abstract :
Exemplar-based inpainting algorithm is a new model used for large region removal. It combines the advantages of both inpainting and texture synthesis. The crucial part of this model is data term which is used to preserve linear structure. In this paper, basing on anisotropic diffusion, deeper mathematical analysis of along isophotes exemplar-based inpainting (AIEI) is given, and it is also proved the data term is the strength of isophotes. A novel cross isophotes exemplar-based inpainting (CIEI) model is proposed, which determines the data term cross the isophotes. Data term in CIEI considers the extent of edge, so CIEI has better linear structure preserving property, and it can fill the target region of natural scene more plausibly. Both theoretical analysis and experiments have verified the validity of the new CIEI model proposed in this paper
Keywords :
image restoration; image texture; along isophotes exemplar-based inpainting; anisotropic diffusion; cross isophotes exemplar-based inpainting; large region removal; linear structure; mathematical analysis; object removal; texture synthesis; Anisotropic magnetoresistance; Data analysis; Filling; Hyperspectral imaging; Image processing; Information science; Layout; Mathematical analysis; Partial differential equations; TV;
Conference_Titel :
Pattern Recognition, 2006. ICPR 2006. 18th International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
0-7695-2521-0
DOI :
10.1109/ICPR.2006.884