Nonconvex Tensor Relative Total Variation for Image Completion

Image completion, which falls to a special type of inverse problems, is an important but challenging task. The difficulties lie in that (i) the datasets usually appear be multi-dimensional; (ii) unavailable or corrupted data entries are randomly distributed. Recently, low-rank priors have gained importance matrix completion problems and signal separation; however, due complexity multi-dimensional data, using prior by itself often insufficient achieve desirable requires more comprehensive approach. In this paper, different from current available approaches, we develop new approach, called relative total variation (TRTV), under tensor framework, effectively integrate local global image information for processing. Based on our proposed model embedded with TRTV p-shrinkage nuclear norm minimization suitable regularization established. An alternating direction method multiplier (ADMM)-based algorithm framework presented. Extensive experiments terms denoising tasks demonstrate not only effective also superior existing approaches literature.