There are various real world situations where, a portion of the image is lost or damaged which needs an image restoration. A prior knowledge of the image may not be available for restoring the image, which demands for a knowledge derivation from the image itself. Restoring the lost portions of the image based on the knowledge obtained from the image area surrounding the lost area is called as digital image in-painting. The information content in the lost area could contain structural information like edges or textural information like repeating patterns. This knowledge is derived from the boundary area surrounding the lost area.
Keywords— Image restoration, digital image in-painting.
[...] The results would look natural enough that observer without prior knowledge of the original image will not notice the gaps. A Preliminary version of this is “Cloning Brush tool” in Adobe Photoshop where the user has to provide the information of what to fill in. There are two major methods of restoring the missing data. The traditional method concentrates on the structural information called the isophotes which are lines of similar color. A series of differential equations are used iteratively to extend these lines in to the missing area using the information obtained from the boundary pixels. [...]
[...] The structure is measured in terms of magnitude of image gradient. D = α Where is the maximum value of the image gradient in the patch, α is a normalization factor. The value of α is 255 for a typical gray scale image. Sobel's Gradient operator [ is used to calculate the image gradient. The restoration process is given by the following algorithm: 1. Extract the manually selected target region and its initial front (boundary) Repeat until there is no boundary pixel: a. [...]
[...] The inpainting algorithm discussed in this paper accepts a damaged image as input with the damaged areas marked in special color. The algorithm first groups the pixels with the special color as unknown. For ease of comparison, we adopt notation similar to that used in the inpainting literature. The lost area which is the region to be filled is called as the target region indicated by Ω as shown in the fig.1a. The boundary between the known and the unknown area or the contour is denoted by δΩ. [...]
[...] Copy image data from Ψ ˆq to Ψ ˆp for all pixels belonging to the target region. f. Update for the newly filled pixels. target size. Fig 3.3 b and 3.4 b shows the corresponding restored images. The time and accuracy for this image is tabulated in Table Fig 3.1 : Screen shot with Input image Fig 3.2 : Screen shot with Partially restored image III EXPERIMENTATION The algorithm is implemented in Java. For any image the patch size should be slightly greater than the smallest texture element (Texel). [...]
APA Style reference
For your bibliographyOnline reading
with our online readerContent validated
by our reading committee
