The process of repairing damaged area or reconstructing specific area or removing unwanted objects from video is known as video inpainting. Most of the automatic techniques are available to deal with this problem but most of them are unable to repair large holes. To get sharp inpainted area, in this paper we have proposed an efficient algorithm using exemplar-based technique. Here, considered static camera which gives video having stationery background with moving foreground. To detect the region of moving objects we apply edge detection technique. Once object region detected priority assignment to the patches is applied. A natural image has structures and textures. Structure sparsity was measure to find similarities of the patches. The patch having higher sparseness is then selected and its priority is set which is the highest priority among the patches. This patch is then used for further inpainting.

Inpainting texture synthesis patch.