论文部分内容阅读
目前现有的视频编码标准 (H.2 6 3,MPEG- 4等 )都是采用半像素双线性插值的算法来对图像进行运动预测 ,为了提高视频编码的效率 ,该文在应用维纳滤波概念的基础上提出了一种新的半像素插值算法。该算法通过构造非等重的维纳滤波器来完成对图像的运动预测功能 ,它充分利用了相邻像素点间的相关性 ,同时还保持了图像中变化较大的边缘信息。实验表明 :该算法相对于双线形插值、Cubic插值、B-Spline和非线性插值在性能上有了明显的提高 ,相同比特率时其峰值信噪比可以提高 0 .6～ 1d B。
At present, the existing video coding standards (H.263, MPEG-4, etc.) all use half-pixel bilinear interpolation to predict the motion of the image. In order to improve the efficiency of video coding, Based on the concept of filtering, a new half-pixel interpolation algorithm is proposed. By constructing a non-equal Wiener filter, this algorithm performs the motion prediction function of the image, which takes full advantage of the correlation between adjacent pixels and keeps the edge of the image changing greatly. Experiments show that the proposed algorithm has significantly improved performance compared with bilinear interpolation, Cubic interpolation, B-Spline and nonlinear interpolation, and its peak signal-noise ratio can be increased by 0.6-6dB at the same bit rate.