Block-based multiscale error concealment using low-rank completion

Abstract

In this paper, we introduce a novel block-based multiscale error concealment method using low-rank completion. The proposed method searches for similar blocks and utilizes low-rank completion to recover the missing pixels. In order to make the full use of the hidden redundant information of images, we seek for more similar blocks by building an image pyramid. The blocks collected from the pyramid are more similar to each other, which leads to a more accurate recovery. Moreover, instead of recovering the missing block at once, we propose a ringlike iterative process to partially minimize the number of unknown pixels and further enhance the recovery result. Experimental results demonstrate the effectiveness of the proposed method.

DOI: 10.1109/APSIPA.2014.7041587

5 Figures and Tables

Cite this paper

@article{Li2014BlockbasedME, title={Block-based multiscale error concealment using low-rank completion}, author={Mading Li and Jiaying Liu and Chong Ruan and Lu Liu and Zongming Guo}, journal={Signal and Information Processing Association Annual Summit and Conference (APSIPA), 2014 Asia-Pacific}, year={2014}, pages={1-4} }