ABSTRACT
The problem of automatically removing objects in video sequences is called video inpainting, as an extension of the image inpainting terminology that refers to the same problem for still digital images. This terminology was introduced in the year 2000 in the original “Image inpainting” article, by Bertalmı´o et al. [101], that adopted the denomination used in art restoration to denote the modification of a painting in a way that is non-detectable for an observer who does not know the original artwork. Other popular terms are image completion, image fill-in, image repair and image disocclusion, the latter being introduced by Masnou and Morel [265] in their pioneering approach, which treated the image hole or gap as an occluding object that had to be removed. Image denoising algorithms do not apply to image inpainting, because all the original information present in the image gap is discarded. For a recent overview on inpainting see [92].
