Sequential Error Concealment for Video/Images by Sparse Linear Prediction

• Published in: IEEE transactions on multimedia Vol. 15; no. 4; pp. 957 - 969
• Main Authors: Koloda, J., Ostergaard, J., Jensen, S. H., Sanchez, V., Peinado, A. M.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.06.2013; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Approximation; Artificial intelligence; Block-coded images/video; Computation; Computer science; control theory; systems; convex optimization; Correlation; Cryptography; Detection, estimation, filtering, equalization, prediction; error concealment; Errors; Exact sciences and technology; Image edge detection; Image reconstruction; Information, signal and communications theory; Interpolation; Linear prediction; Mathematical models; missing data imputation; Pattern recognition. Digital image processing. Computational geometry; Prediction algorithms; Predictive models; Regression; Signal and communications theory; Signal, noise; sparse representation; Telecommunications and information theory; Video coding; Block code; Block-coded images/video; Video signal; Lossy channel; Convex programming; missing data imputation; Missing data; Linear prediction; Coding; Sequential method; Sparse representation; convex optimization; error concealment; Incomplete information; Transmission loss; Pixel; Signal to noise ratio
• ISSN: 1520-9210; 1941-0077
• DOI: 10.1109/TMM.2013.2238524

In this paper, we propose a novel sequential error concealment algorithm for video and images based on sparse linear prediction. Block-based coding schemes in packet loss environments are considered. Images are modelled by means of linear prediction, and missing macroblocks are sequentially reconstructed using the available groups of pixels. The optimal predictor coefficients are computed by applying a missing data regression imputation procedure with a sparsity constraint. Moreover, an efficient procedure for the computation of these coefficients based on an exponential approximation is also proposed. Both techniques provide high-quality reconstructions and outperform the state-of-the-art algorithms both in terms of PSNR and MS-SSIM.