Answers are grounded in this paper and cite their sources.
In this paper, a whole frame loss error concealment algorithm for three-dimensional video coding is proposed. The main concept of the proposed algorithm is to extrapolate the motion vectors for concealing a current error block by jointly considering the available motion vector and the depth information. In addition, the depth information is adopted to help the derivation of reference pixels for concealing errors in the case that suitable motion vectors cannot be obtained by the motion vector extrapolation process alone. Simulation results demonstrate that our proposed algorithm can achieve up to 0.52 dB PSNR, as well as subjective quality improvement, compared to previous work.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: