针对图像修补的不足和视频帧间编码的特点, 提出利用改进的图像修复的帧间差错隐藏联合算法. 根据丢失宏块的周围宏块的运动相关性和正确接收情况对丢失宏块进行运动性和可用性判断, 如果丢失宏块属于静止宏块且可用性不为零, 则采用改进的图像修补隐藏帧间差错, 否则采用外边界匹配准则进行差错隐藏; 并在图像修补中对置信度、等照度和修复方向等方面进行了改进. 该算法既发挥了图像修复对纹理丰富区域的修复优势, 也克服了图像修补对运动区域的隐藏效果不好的缺陷. 实验结果表明, 该算法可以带来主观和客观质量上的提高, 并且明显克服了外边界匹配算法带来的视频抖动.
To overcome the defects of image inpainting and satisfy the characters of video inter-frame coding, a hybrid inter-frame error concealment algorithm using improved image inpainting is proposed in this paper. The movability and availability of the lost MB is judged according to the motion correlation and the receiv-ing status of neighboring MBs. If the lost MB is static and its availability is not zero, the inter concealment based on improved inpainting is used. Otherwise, the outer boundary matching algorithm (OBMA) is used to repair the lost MB. Here, the inter image inpainting is modified with confidence function, isolux function and repairing direction. The algorithm not only takes the advantage of image inpainting for repairing the complex texture area, but also overcomes the defect of image inpainting for repairing the motion area. The experiment results show that the proposed algorithm improves the subjective and objective quality of the video, and reduces the shake induced by the concealment of OMBA.