This paper provides an H.264 algorithm that makes the detection of all-zero blocks more efficient in the absence of frames. This paper includes the following contents, first introduces the background of video compression technology, describes the development trend of video coding at home and abroad, and explains the process of H.264 coding. Then we analyze the integer transform, content-based adaptive (CAVLC) and adaptive block transform (ABT) technique combined with macroblock size. Then, the paper analyzes the H.264 all-zero block detection in detail, and compares the traditional method and the detection efficiency of the new detection method. Finally, Compared with the traditional coding results to make an analysis.

  At the same time, in 264, the residual signal obtained by motion prediction requires integer transformation and the amount can only be encoded first. In many cases, especially at low bit rate conditions, the overall probability is high with zero quantization coefficient blocks. In this paper, we compare the characteristics of several existing all-zero block detection algorithms. The 4x4 integer transforms its own characteristics, integer transformation, and quantization process inversion, and uses secondary governance and obtains a method for detecting all zero- New algorithm. The experimental results show that this method can not improve the detection rate of zero block and reduce the required computational cost under the premise of improving the quality of video coding.

Keywords: Video coding, all zero block detection, H264, intra prediction

 目  录

摘   要 2

Abstract 3

1  绪  论 9

1.1 课题背景 9

1.2  视频压缩编码的理论基础 10

1.3  视频压缩编码标准 11

1.3.1  H.261视频编码标准 11

1.3.2  H.263视频编码标准 11

        1.3.3  MPEG-1视频编码标准 12

    1.4  H.264视频编码标准 13

    1.5 视频质量标准 13

        1.5.2 客观质量的测量 14

2  H.264算法分析 14

    2.1  H.264对宏块的编码 15

        2.1.1  16 x16帧内预测 16

        2.1.2  4x4帧内预测模式 17

        2.1.3  色差分量的帧内预测 18

2.2  帧间预测 18

        2.2.1  树型构造的运动补偿 18

        2.2.2  1/4像素精度的运动估计 19

        2.2.3  多参考帧模式选择 19

2.3  变换与量化 20

        2.3.1  整数DCT 变换 20

        2.3.2  量化 21

2.4  熵编码 21

        2.4.1  Exp-Golomb码 22