10-Basic Video Compression Techniques

mti

1

Lecture 10 Basic Video Compression Techniques

2010.12.8

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

2

Introduction to Video Compression Video Compression with Motion Compensation Search for Motion Vectors Video Coding Standards

H.26x Video Coding Standards MPEG Standards

AVS Standardshttp://www.77cn.com.cn/

2010 (c) Multimedia Technology Institute of Xidian University

mti

3

A video consists of a time-ordered sequence of

frames, i.e., images. An obvious solution to video compression would be predictive coding based on previous frames. Compression proceeds by subtracting images: subtract in time order and code the residual error. It can be done even better by searching for just the right parts of the image to subtract from the previous frame.2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

4

空间冗余 Spatial Redundancy

变换编码帧内预测编码帧间预测运动补偿

时间冗余 Temporal Redundancy

视觉冗余 Visual Redundancy

非线性量化对象编码等http://www.77cn.com.cn/

2010 (c) Multimedia Technology Institute of Xidian University

mti

5

Consecutive frames in a video are similar - temporal

redundancy exists. Temporal redundancy is exploited so that not every frame of the video needs to be coded independently as a new image. The difference between the current frame and other frame(s) in the sequence will be coded - small values and low entropy, good for compression. Steps of Video compression based on Motion Compensation (MC):1. Motion Estimation (motion vector search). 2. MC-based Prediction. 3. Derivation of the prediction error, the difference.2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

6

运动补偿帧间预测(temporal)+ DCT(spatial)

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

7

Each image is divided into macroblocks

of size N×N. (By default, N= 16 for luminance images. For chrominance images, N= 8 if 4:2:0 chroma subsampling is adopted.) Motion compensation is performed at the macroblock level. (The current image frame is referred to as Target Frame. A match is sought between the macroblock in the Target Frame and the most similar macroblock in previous and/or future frame(s) . The displacement of the reference macroblock to the target macroblock is called a motion vector, MV.2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

8

对于基于运动补偿的视频压缩，每个场景在第一帧后，只需要对运动向量和差值宏块编码，因为这些信息足以用于解码并重新生成完整的图像。运动向量的求解是一个匹配问题，即相关性判定问题。由于运动向量搜索的计算复杂度较高，所以通常限制在

一个较小的相邻区域内进行。水平位移i和垂直位移j规定在[-p,p]范围之内，p是一个较小的正整数。下图中所示的搜索窗口大小为(2p+1)×(2p+1)。宏块的中心(x0, y0)可以放在窗口的任意一个单元格中。搜索的目标是找到一个向量作为运动向量，使得两个宏块的差(可选：平均绝对误差MAD或绝对误差和SAD或均方差MSE)最小。

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

9

求解运动向量最简单的方法就是顺序搜索参考帧中的整个 (2 p 1) (2 p 1)大小的窗口，也称为完全搜索。显然顺序搜索算法的代价是相当高的，每个像素的比较需要三个操作，获取一个宏块的运动向量的复杂度为： (2 p 1) (2 p 1) N 2 3 O( p 2 N 2 )例：720×480,30fps,运算量为29.89×109

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

10

对数搜索不是最优方法，但通常是非常有效的，而且代价较低。 2D对数搜索方法的运动向量搜索过程中需要进行多次迭代，类似折半查找过程。 2D对数搜索方法的计算复杂度降低到 O(log 2 p N 2 )。由于p和N在通常情况下是一个数量级的，所以与顺序搜索方式相比，已经得到明显改善。例：720×480,30fps,运算量为1.25×1092010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

11

对偶搜索是一种有效的搜索方法。这种方法首先在水平方向上计算i-1, i, i+1三个点中误差最小的点，

然后以之为中心继续，直到找到误差最小的点，然后再用同样的方法寻找垂直方向上误差最小的点，从而找到最佳匹配。2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

12

分层(多分辨率)方法中初始的运动向量估计是从显著降低分辨率后的图像中获得的。原始图像为第0层，第1层和第2层是通过将上一层图像分辨率减半获得的。初始的搜索从第 2层开始，由于宏块变小了，p值也随之成比例减小，这一层的计算量也大大缩小。运动向量的估计从初始的粗糙到一层层修正，直到第0层。例：720×480,30fps,运算量为0.51×1092010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

13

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

14

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

15

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

16

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

17

数字视(音)频编解码标准

H.26X系列 (ITU标准)

MPEG系列 (

ISO/IEC标准)AVS (中国标准)

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

mti

18

ITU:

H.261

H.262

H.263 H.263+ H.263++

H.264

ISO/IEC: MPEG1 (Part2)

MPEG2 (Part2)

MPEG4 (Part2) MPEG21

MPEG4(Part10)

MPEG7

GB:

AVShttp://www.77cn.com.cn/

2010 (c) Multimedia Technology Institute of Xidian University

mti

19

H.261也称p×64,这是ITU-T(前身为CCITT)最

早制定的关于视频编码的国际标准。考虑到ISDN的数据传输码率以64kbps为单位，因此把H.261标准码率定义为p×64k bps (p= 1~ 30)。H.261标准

主要用于可视电话和电视会议系统。它支持QCIF(p= 1, 2), CIF (p> 2)两种图像输入格式。

2010 (c) Multimedia Technology Institute of Xidian University

http://www.77cn.com.cn/

本文来源：https://www.bwwdw.com/article/8dyj.html