A High-Quality Video Reconstruction Optimization System Based on Compressed Sensing

The video compression sensing method based on multi hypothesis has attracted extensive attention in the research of video codec with limited resources. However, the formation of high-quality prediction blocks in the multi hypothesis prediction stage is a big challenge. To solve this problem, this paper constructs a novel compressed sensing-based high-quality video reconstruction optimization system and a range of optimization of prediction blocks models. We combine the high-quality optimization reconstruction of foreground block with the residual reconstruction of background block to improve the overall reconstruction effect of video sequence. The new system mainly includes two parts: the selection of blocks and the optimization of prediction blocks. First, in order to accurately find blocks with large changes and improve the efficiency of the algorithm, we divide blocks into foreground blocks and background blocks by using the difference between the current block and adjacent frame blocks. It can quickly determine the blocks to be optimized. Second, to improve the accuracy of the prediction block, this paper designs multi objective OPBS-NSGA-II model and single objective OPBS-PSO model realizes optimization of prediction blocks. It effectively suppresses the influence of the fluctuation of prediction block on reconstruction and improves the reconstruction performance. In addition, we select the objective function that has a greater impact on the final reconstruction performance through experiments. Experimental results show that the proposed compressed sensing-based high-quality video reconstruction optimization system significantly improves the reconstruction performance in both objective and supervisor quality.