基于自适应鲁棒性光流的差错掩盖

Error Concealment Based on Adaptive Robust Optical Flow

编码数字视频序列经过受噪声影响的信道传输时,通常会出现图像信息丢失。该文提出一种基于自适应鲁棒性光流的差错掩盖方法,作为解码端的工具解决这样的问题。该文利用光流技术能有效获取物体运动估计的特性,对丢失块进行逐像素点的恢复,既避免了图像模糊,又消除了块效应。在光流的估计中,自适应地调整了目标泛函中的数据保持项与空间连贯项之间的关系,并引入Lorentz函数来构造目标泛函,提高了光流的鲁棒性。仿真结果表明,该文提出的方法无论在主观视觉评价,还是在客观的数值标准下,都能比现有的误差掩盖方法恢复出质量更好的图像。

During transmitting encoded digital video sequences over noisy communication channels, the information of the images will be lost. An error concealment method based on robust optical flow is proposed, as a post-processing tool at the decoder side to solve this problem. The property of optical flow technique on efficiently obtaining motion estimation is used to recover the lost block in pixel-wise manner. This processing not only avoids the image blurring, but also reduces the blocking effect. In the estimation of optical flow, the relation between data conservation term and spatial coherence term of the objective function is adaptively adjusted. In order to increase the robustness, Lorentzian function is used to construct the energy function. Simulation results show that the proposed method can recover the higher quality image on both subjective visual evaluation and objective numerical metrics, comparing to the existing error concealment methods.