Automatic Image Inpainting In Wavelet Domain Using successive elimination algorithm
summer project pal|
Active In SP
Joined: Jan 2011
19-01-2011, 05:53 PM
Image inpainting refers to ﬁlling in the missing parts or modifying the damaged parts of an image in a visually plausible way.‘Image inpainting’, an artistic term used from ancient times, refers to restoration or retouching works of paintings. This technique can be used for restoring the missing parts of an image or for removing the unwanted objects from an image. In digital images, role of image inpainting techniques grow from mere restoration of images, photographs and ﬁlms to powerful image enhancement and image completion techniques. These techniques have massive range of applications like special effects in images and videos, image compression for transmission and storage, zooming, superresolution and image fusion. Digital image inpainting tries to modify
the digital images in such a way that the modiﬁed region is hardly noticeable. A novel approach for digital image inpainting in wavelet domain is presentedA novel approach for digital image inpainting in wavelet domain is presented here. We make use of successive elimination algorithm, which can automatically ﬁll the damaged regions depending on the surrounding information in the image. We consider the problems of missing or damaged coefﬁcients in both spatial and wavelet domain.
The advantage of this method is that sum of absolute difference is used for similarity measure. Also computational efﬁciency is highly improved. Usually during digital image inpainting, human intervention is needed to indicate the damaged part in an image. In this work, a novel algorithm for automatic detection of ink sprayed image using a wavelet based approach is also incorporated so that image inpainting is fully automatic. The damaged regions in the distorted image are automatically detected and these regions are then effectively restored.
Automatic Image Inpainting In Wavelet Domain Using successive elimination algorithm.pdf (Size: 7.57 MB / Downloads: 220)