An Efficient Video Interpolation Framework for Unwanted Object Removal

Abstract: When videos and photographs are captured some unwanted objects come in. On removal of such unwanted objects through image editing tools result in blurred image. It may not be removed properly. There may be case that one wants to replace that undesired object with some new object. In such case both texture and structure of the surrounding region should be preserved. Due motioning of camera some portion contains blur pixels. Some images might appear with missing portions and having cracks. There might be titles, headings written on photographs can be removed using image inpainting. In this paper, a framework is implemented for video inpainting using exemplar based image inpainting. Initially, a video containing unwanted objects is processed and frames are extracted containing unwanted objects and inpainted using exemplar based image inpainting. After inpainting, inpainted frames are saved and all frames are merged to a video. Quality of inpainted video is measured by PSNR.

Keywords: Unwanted object, Image Inpainting, Video Inpainting, Texture, Blurred Image.