Image Completion with Hybrid Interpolation in Tensor Representation

Abstract

The issue of image completion has been developed considerably over the last two decades, and many computational strategies have been proposed to fill-in missing regions in an incomplete image. When the incomplete image contains many small-sized irregular missing areas, a good alternative seems to be the matrix or tensor decomposition algorithms that yield low-rank approximations. However, this approach uses heuristic rank adaptation techniques, especially for images with many details. To tackle the obstacles of low-rank completion methods, we propose to model the incomplete images with overlapping blocks of Tucker decomposition representations where the factor matrices are determined by a hybrid version of the Gaussian radial basis function and polynomial interpolation. The experiments, carried out for various image completion and resolution up-scaling problems, demonstrate that our approach considerably outperforms the baseline and state-of-the-art low-rank completion methods.

Tensor Methods in Computer Vision and Deep Learning

“…By representing visual as tensors, models addressing inverse problems exploit low-rank tensor decompositions (Section IV-B), or rely on robust variants (Sections IV-C IV-E) to estimate a low-rank tensor which forms a basis from which missing data can be imputed or denoised. Along this line of research, several methods for image and video inpainting and/or denoising have been developed by estimating the low-rank tensor using either nuclear norm-based regularizers, or by assuming an explicit low-rank tensor structure described by decompositions such CP, Tucker, Tensor Ring, or TT decomposition [218], [219], [220], [221], [92], [99], [100], [101], [98]. For the same task, methods based on tensor-structured separable dictionary learning (Section IV-E) have also been proposed [4], [115], [11].…”

Special Issue on Intelligent Processing on Image and Optical Information

“…The incomplete images can contain many missing pixels distributed over the entire image. In the paper entitled 'Image Completion with Hybrid Interpolation in Tensor Representation', Rafał Zdunek and Tomasz Sadowski [18] proposes an interpolation algorithm for a wide spectrum of image-completion problems. Their algorithm outperforms other conventional methods with a considerably shorter computational runtime.…”

Patch size adaptive image inpainting

“…D. Thanh et al [10] proposed an adaptive image inpainting using adaptive parameters estimation, parameters can be estimated based on the information of the image, such as discrete gradient. R. Zdunek et al [11] use hybrid interpolation in tensor representation for inpainting image, it only It can only repair a small area. In order to get similar patch for filling image, S. Yang et al [12] proposed multi-patch match with adaptive size, this method improves the accuracy of image matching and saves time.…”