宏块

基于MPEG－1的模拟结果表明了基于宏块分类的运动估计法之有效性。

Simulation result based on MPEG 1 shows the effectiveness of the motion estimation method based on macroblock ′ s classification .

本文主要研究在DCT域内基于宏块公共信息的运动补偿快速算法和利用DCT系数进行码率控制算法。

This dissertation presents the research on the fast motion compensation algorithm with the shared macroblock information in the dct domain and rate control algorithm with dct coefficients .

算法优化了宏块帧间模式选择算法的流程

The algorithm optimizes the flow of inter mode of macro-block decision algorithm .

采用该设计思想实现了一个具有256个PE的电路，用于处理16×16的宏块。

A computing architecture with 256 PEs was implemented for the 16 × 16 block .

设计获得MB（MacroBlock宏块）的方法和过程。

Design the method and process to gain the MB ( Macro Block ) .

给出了帧预测宏块、场预测宏块半象素插值的程序实现，进而对逐行与隔行方式下P帧运动补偿的实现做出较为全面的对比分析。

Furthermore , it makes analysis about the comparison of motion compensation in the mode of interlacing and progressive in the P-VOP of MPEG-4 .

自适应阈值的宏块MAD快速帧内算法

Fast adaptive thresholds intra-frame prediction algorithm based on MAD

引入全零DCT宏块估计算法，省去了部分DCT变换。

Introduces the entire zero DCT macro block estimation algorithm to avoid some unnecessary DCT .

MPEG-2到MPEG-4转换编码中场-帧编码宏块转换及运动矢量优化算法

Field to Frame Conversion and Motion Vector Optimize algorithm of MPEG-2 to MPEG-4 Transcoding

压缩域MPEG-2到MPEG-4视频转码中不匹配宏块的复原算法

Robust Mode Mismatch Macroblock Retrieval Algorithm for Transcoding MPEG-2 to MPEG-4 in Compression Domain

采用基于模糊积分的图像失真度量代替MAD，提出一种基于宏块条的自适应码率控制算法。

A class of image distortion metric based on fuzzy integral was referred to and applied to the rate controlling .

采用自顶向下的设计方法，划分了子模块，对各子模块进行仿真，进行了FPGA综合，对宏块滤波所需时钟周期进行了分析。

We do the simulation for each sub-module . At last we do FPGA synthesis and analyze the clock cycles required for the macro block filter .

该算法只需要通过直接抽取MPEG视频流中的B帧、P帧中的宏块类型信息，并对其进行计算就可以检测出视频中存在的镜头边界，从而对场景变换进行精确定位。

The proposed method can detect scene boundaries through using various macroblock type of B-frame and P-frame in MPEG video flow in order to realize accurate scene change positioning .

宏块变换算法通过矩阵运算，在DCT域内从参考帧中抽取对应运动矢量指向参考区域的频域值，并与宏块的DCT值运算，从而实现帧内编码与帧间编码的宏块转换。

The macroblock coding transform algorithm extracts the reference region in DCT domain by matrix computation , which is pointed by motion vector .

然后，本文从JPEG图像相邻宏块的相关性出发，提出一种新的基于C均值聚类的去块效应算法。

Then , proceed from the correlation between adjacent macro blocks in JPEG images , this thesis proposes a new algorithm based on c-means clustering method to reduce blocking artifacts .

算法利用了SKIP模式在视频序列中的使用率信息和时空域相邻宏块与当前宏块编码模式之间的高度相关性。

The proposed algorithm utilized the adoption information of SKIP mode in sequences and high correlation of coding modes between spatio-temporal adjacent macroblocks and current macroblock .

本文提出了一个基于HVS感知特性的、以宏块为基本单位的RDO算法。

This Thesis proposes a novel macroblock-level RDO algorithm based on perceptual features of HVS .

为了提高加密效率，满足实时性的要求，选择I帧的DC差分系数和P帧、B帧的Intra宏块的DC差分系数进行加密。

Order to improving the encryption efficiency , meeting the real-time and security requirements , the DC differential coefficients of I frames and Intra marco-blocks of P / B frames are chosen to be encrypted .

本文首先分析了课题背景和发展现状，简单比较了几种宏块分割模式算法的特点和解码器DSP优化方法。

In this paper the background and the status of the research are analyzed firstly . Several algorithms of the choice of macro block partition modes and methods of DSP optimization are showed briefly .

该算法将全部编码模式进行分类，对SKIP模式采用提前终止判决，而对其它模式可通过判别宏块的一致性和比较宏块的空时相关性逐类型比较，可有效减少判决的模式数量。

It classifies all modes by both homogeneity and space-time correlation of macroblock . The SKIP mode can be early terminated , and other modes can be compared type by type .

而后，介绍了DCT域转码中的其它关键算法，包括：DCT域的运动补偿算法，宏块模式的选择，码速控制算法。为进一步提高转码效率，本文对转码器的性能进行了优化。

Third , the other key algorithms of transcoding in DCT domain was introduced , including : motion compensation in DCT domain , the choosing of macroblock pattern and rate control .

根据像素值在时间和空间的分布，用具有最大出现概率的像素值来重复更新Sprite中相应宏块的像素值。

Each pixel-value of the macro block on sprite is iteratively updated by the value with maximum appearance probability on temporal and spatial distribution .

选择宏块级亮度直方图差异BH衡量各P帧的编码复杂度，并根据相对复杂度调整各帧的目标比特。

Macro-block level luminance histogram difference ( BH ) is applied to measure the encoding complexity of each P frame , and the target bits of each frame are adjusted according to the relative complexity .

然而，传统的视频编码过程都是以宏块为基本处理单位，对一整帧图像中的宏块依次进行处理，这使得系统Cache的性能十分低下。

However , the classical video coding algorithms use macroblock as the basic processing unit and all the macroblocks in a frame should be done sequentially , and such a scheme make the cache performance very low .

通过减小或少量增加宏块中8×8亮度块的DCT交流系数的能量，调整4个亮度块之间的能量关系嵌入一位水印信息。

One bit watermark was embedded by decreasing or increasing less the energy of AC coefficient in 8 × 8 luminance block , and adjusting energy relation between four luminance blocks in macro-block .

VP8在这里犯的致命的错误是没有把宏块级量化作为其核心功能。

The killer mistake that VP8 has made here is not making macroblock-level quantization a core feature of VP8 .

首先，通过DCT变换选择出帧内预测最佳模式和帧间预测最佳模式，然后通过量化操作，进一步选择出当前宏块的最佳编码模式。

Firstly , through the DCT transform get the optimal mode of intra prediction and the optimal mode of inter prediction . Secondly , by quantization operation further choose the optimal coding mode of current macroblock .

最后再由AVS编码系统进行当前宏块预测模式的选择，从而保证当前宏块在嵌入水印信息的情况下，选择一种最优的预测模式。

At last , the prediction modes of current macroblock is determined using standard AVS coding algorithm which will ensure that the best prediction mode is used for current watermarked macroblock .

运动分类机制通过对中心宏块的SAD进行门限判别来完成运动类型的分类，从而获得准确的运动类型预测，而且避免了大量数据存储和排序计算。

The classification of motion type is judged by the threshold of the SAD in the center macro block , which is an approach that can provide accurate motion type prediction and avoid mass storage as well as complex sorting operations .

首先，用时空加权模型来预测MAD；然后，采用宏块层码率控制策略来分配目标比特；最后，依然采用JVT-G012中的二次R-Q模型计算QP。

Firstly , the MAD is predicted by a new time and space weighted model . Secondly , macroblock layer rate control strategy is used to distribute target bit .