中文摘要：

Intra帧图像对整个视频图像起着非常重要的作用。基于图像局部边缘类型和走向分析,提出一种能提高Intra帧图像质量的错误隐藏算法。首先根据相邻宏块中边缘类型和走向来判断当前出错宏块的边缘类型与走向分成多个小区域,以分块插值恢复数据。对平缓图像,采取时域替代方法恢复数据,既降低算法复杂度,又能保证良好的图像质量。实验结果表明,本算法能有效提高图像的解码质量。

英文摘要：

The quality of intra pictures is very important to the video sequence, which is the basis of predictive pictures subsequently. A novel and effective error concealment algorithm for intra-coded frames is proposed, which utilizes the edge features and directions analysis technique. In the algorithm, the missing macro-blocks are divided into several sub-blocks according to the edge characteristic of the missing macro-blocks, which is achieved through analyzing the neighboring correctly received Macro-blocks, and in each sub-block, the linearity interpolation is employed to recover the pixel of sub-block, thus completing the missing macro-block recovery. If the sequence scene is stable and changing slowly, temporal error concealment is utilized to reduce the computation complexity, while the quality of the concealed image is also satisfied. The result of experiment shows that this approach achieves performance improvement over existing methods in both PSNR and subjective picture quality.