中文摘要：

针对分布式视频编码（DVC）中的边信息重建问题，提出了一种空域平滑的边信息重建算法。该算法在关键帧之间的前向运动估计时，设置代价函数阈值保证目标运动轨迹的线性连续。对于运动平稳区域，采用双向运动补偿插值生成边信息；对于其它区域，在关键帧中确定多个候选块，并选择与待插值块的4个相邻块具有最小边界匹配误差的候选块作为边信息。最后，利用马尔可夫随机场（MRF）结合最大先验概率估计（MAP）对边界不连续的插值块进行空域平滑。实验表明，该算法在降低编码码率的同时提高了解码图像的峰值信噪比（PSNR）。

英文摘要：

Based on the analysis of side information reconstruction in distributed video coding （DVC）, an improved side information interpolation algorithm is presented in this paper. The algorithm sets the threshold function to ensure the linear continuity of target motion tracks. For the smooth motion field, it adopts bi-direction motion estimation and compensation to generate the side information. For other fields, the block which has the minimum edge matching error with four adjacent blocks of the block waiting interpolation is chosen from the candidate blocks. Then the maximum a posteriori probability （MAP） combined with the Markov random field （MRF） is performed on edge inconsistent interpolation blocks for spatial smoothing. The experiment result shows the proposed algorithm can achieve the better peak signal to noise ratio （PSNR） while decreasing the encoding rate.