公开(公告)号：EP1732329A1

公开(公告)日：2006-12-13

申请号：EP05425407.3

申请日：2005-06-06

Applicant: STMicroelectronics S.r.l.

Inventor： Filippini, Gianluca ,  Biffi, Bruno ,  Rovati, Fabrizio Simone ,  Piccinelli, Emiliano

IPC: H04N7/26 ,  H04N7/50

CPC classification number: H04N19/61 ,  H04N19/51

Abstract: A moving-image signal, such as typically a luminance signal organized in blocks of pixels (10; 20) is coded via a technique that envisages the steps of:
- comparing a block to be coded (10) with a plurality of candidate prediction blocks (20);
- determining, for each candidate prediction block (20), a respective value of an index (SRS) representing the difference between the block to be coded (10) and each candidate prediction block (20); and
- choosing between the candidate prediction blocks (20), as a function of the respective value of the aforesaid index (SRS), a prediction block to be used for coding (30, 40) of the block to be coded.
The signal is sampled pixel by pixel on the block to be coded (10) and on the plurality of candidate prediction blocks (20), thus generating respective surfaces representing the pattern of the signal in the block to be coded (10) and in the candidate prediction blocks (20). Chosen as index is an index of the parallelism between the aforesaid respective surfaces.

Abstract translation: 运动图像信号，例如典型地组织在像素块（10; 20）中的亮度信号的编码通过一种技术来编码，该技术设想以下步骤： - 将待编码的块（10）与多个候选预测块 （20）; - 为每个候选预测块（20）确定表示待编码块（10）与每个候选预测块（20）之间的差的索引（SRS）的相应值; 以及 - 根据上述索引（SRS）的相应值，在候选预测块（20）之间进行选择，用于要编码的块的编码（30,40）的预测块。 在要编码的块（10）上和在多个候选预测块（20）上逐像素地采样信号，从而生成表示要编码的块（10）中的信号的模式的各个表面，并且在 候选预测块（20）。 作为指标选择上述各表面之间的平行度的指标。

公开(公告)号：EP0944245B1

公开(公告)日：2001-07-25

申请号：EP98830163.6

申请日：1998-03-20

Applicant: STMicroelectronics S.r.l.

Inventor： Piccinelli, Emiliano ,  Pau, Danilo ,  Zuccaro, Amedeo

IPC: H04N5/00 ,  H04N7/36

CPC classification number: H04N19/51 ,  G06T7/207 ,  G06T7/223 ,  H04N19/56

公开(公告)号：EP1381242A2

公开(公告)日：2004-01-14

申请号：EP03014057.8

申请日：2003-06-23

Applicant: STMicroelectronics S.r.l.

Inventor： Pau, Danilo ,  Alfonso, Daniele ,  Piccinelli, Emiliano ,  Signorelli, Tiziana

IPC: H04N9/64 ,  G09G5/36

CPC classification number: G09G3/3611 ,  G09G3/3688 ,  G09G5/363 ,  G09G2340/02 ,  G09G2340/06 ,  H04N9/64

Abstract: An RGB digital video signal destined to be displayed on a display such as a liquid crystal display (LCD) is converted from the RGB colour space to the YUV colour space. The signal converted into the YUV colour space is subjected to at least a processing operation selected among a sub-sampling operation (24) and a data compression operation (26). The signal is then stored in a memory and the signal read from said memory (12) is then subjected to at least a return operation (28, 30) complementary to the aforesaid processing operation (24, 26). The signal subjected to the aforesaid return operation is lastly reconverted from the YUV colour space to the RGB colour space, thus being susceptible to be displayed on the display.

Abstract translation: 将要显示在液晶显示器（LCD）的显示器上的RGB数字视频信号从RGB色彩空间转换为YUV色彩空间。 转换成YUV色彩空间的信号至少经受从子采样操作（24）和数据压缩操作（26）中选择的处理操作。 然后将信号存储在存储器中，然后从所述存储器（12）读取的信号至少经受与上述处理操作（24,26）互补的至少一个返回操作（28,30）。 经过上述返回操作的信号最后从YUV颜色空间重新转换为RGB颜色空间，因此易于显示在显示器上。

公开(公告)号：EP1296524A1

公开(公告)日：2003-03-26

申请号：EP01830597.9

申请日：2001-09-20

Applicant: STMicroelectronics S.r.l.

Inventor： Bartolucci, Vilim ,  Pau, Danilo ,  Piccinelli, Emiliano

IPC: H04N7/28 ,  H04N7/50

CPC classification number: G06T9/008 ,  H04N19/124 ,  H04N19/139 ,  H04N19/14 ,  H04N19/15 ,  H04N19/152 ,  H04N19/17 ,  H04N19/174 ,  H04N19/176 ,  H04N19/61 ,  H04N19/94

Abstract: A process for encoding digital video signals (IS) organized in frames (12) comprises the operations of dividing said frames into blocks starting from macroblocks subjected to motion-compensation (13, 14) and applying to said blocks a discrete cosine transform (15) in such a way as to generate respective sets of coefficients. The said sets of coefficients are then assembled (20) by being organized into sets of vectors (X) by means of masking (M 1 , M 2 , M 3 , M 4 ). Once the variance of the vectors has been detected (214), the vectors themselves are quantized (217 to 220) on a number of available bits by means of a pyramid vector quantizer (22), associating to the vectors respective quantization pyramids having given sizes according to the variance detected and to the number of available bits. Finally, the vectors are encoded with respective codewords.

Abstract translation: （IS）组织的一种用于编码数字视频信号处理中的帧（12）包括将所述帧到从经过运动补偿（13，14）的宏块开始，并施加到所述块的离散余弦变换块（15）的操作 在寻求一种方式，以产生respectivement组系数。 系数的所述组，然后通过掩模（M1，M2，M3，M4）的借助于被组织成矢量集（X）的组装（20）。 一旦向量的方差已经检测（214），所述载体本身是由一个金字塔矢量量化器（22）的方式量化（217至220）上的多个可用的位，缔合到载体respectivement量化金字塔具有给定尺寸的 gemäß检测到的方差和可用数位的数目。 最后，将载体与编码码字respectivement。

公开(公告)号：EP1381242A3

公开(公告)日：2008-01-16

申请号：EP03014057.8

申请日：2003-06-23

Applicant: STMicroelectronics S.r.l.

Inventor： Pau, Danilo ,  Alfonso, Daniele ,  Piccinelli, Emiliano ,  Signorelli, Tiziana

IPC: H04N9/64 ,  G09G5/36

CPC classification number: G09G3/3611 ,  G09G3/3688 ,  G09G5/363 ,  G09G2340/02 ,  G09G2340/06 ,  H04N9/64

Abstract: An RGB digital video signal destined to be displayed on a display such as a liquid crystal display (LCD) is converted from the RGB colour space to the YUV colour space. The signal converted into the YUV colour space is subjected to at least a processing operation selected among a sub-sampling operation (24) and a data compression operation (26). The signal is then stored in a memory and the signal read from said memory (12) is then subjected to at least a return operation (28, 30) complementary to the aforesaid processing operation (24, 26). The signal subjected to the aforesaid return operation is lastly reconverted from the YUV colour space to the RGB colour space, thus being susceptible to be displayed on the display.

公开(公告)号：EP1294195A1

公开(公告)日：2003-03-19

申请号：EP01830589.6

申请日：2001-09-14

Applicant: STMicroelectronics S.r.l.

Inventor： Concion, Davide ,  Piccinelli, Emiliano ,  Pau, Danilo

IPC: H04N7/26 ,  G06T9/00

CPC classification number: H04N19/40 ,  H04N19/59

Abstract: The DCT macroblocks (MB) comprised in an MPEG video bitstream are divided into blocks (B1, B2, B3, B4), each of which comprises a plurality of microblocks. In each DCT block (B1, B2, B3, B4), the significant frequencies are identified and preserved, isolating a corresponding microblock (mB1, mB2, mB3, mB4) preferably consisting of the microblock on the top left of each block (B1, B2, B3, B4) and setting to zero the coefficients of the remaining microblocks. On the microblock thus isolated there is performed an inverse discrete cosine transform (16), and the microblock thus obtained is merged (18) with the homologous microblocks obtained from the other blocks comprised in a respective starting macroblock (MB), so as to give rise to a merging block. The merging block thus obtained undergoes a discrete cosine transform (20) so as to obtain a final block, which can be assembled into a macroblock (MBD) with converted format.

Abstract translation: 包含在MPEG视频比特流中的DCT宏块（MB）被划分成块（B1，B2，B3，B4），每个块包括多个微块。 在每个DCT块（B1，B2，B3，B4）中，识别和保存有效频率，隔离优选地由每个块（B1，B2）的左上角的微块组成的相应微块（mB1，mB2，mB3，mB4） B2，B3，B4），并将剩余微块的系数设置为零。 在这样分离的微块上，执行反离散余弦变换（16），并且将由此获得的微块与从包含在相应起始宏块（MB）中的其它块获得的同源微块合并（18），以便给出 上升到合并块。 如此获得的合并块经历离散余弦变换（20），以获得最终的块，其可以被组合成具有转换格式的宏块（MBD）。

公开(公告)号：EP0917363A1

公开(公告)日：1999-05-19

申请号：EP97830605.8

申请日：1997-11-17

Applicant: STMicroelectronics S.r.l.

Inventor： Zuccaro, Amedeo ,  Pau, Danilo ,  Piccinelli, Emiliano

IPC: H04N7/00 ,  H04N7/36

CPC classification number: H04N5/145 ,  G06T7/238 ,  G06T2207/10016 ,  G06T2207/20052 ,  H04N19/51 ,  H04N19/56

Abstract: A motion estimator operating on a recursive mode reduces remarkably the number of operations per pixels required by the particular coding process being implemented, typically the coding according to the MPEG standard. The method is based on the correlation existing among motion vectors associated to macroblocks in a homologous position in temporally adjacent images and also associated to macroblocks belonging to the same picture and spatially adjacent to the current macroblock being processed.
By admitting this double correlation the calculation burden is greatly reduced.

Abstract translation: 以递归模式操作的运动估计器显着减少了正在实现的特定编码过程所需的每像素的操作次数，通常是根据MPEG标准的编码。 该方法基于与时间相邻图像中的同源位置中的宏块相关联的运动矢量之间的相关性，并且还与属于相同图像的宏块相关联，并且与正在处理的当前宏块空间相邻。 通过承认这种双重相关性，大大减少了计算负担。

公开(公告)号：EP1231794A1

公开(公告)日：2002-08-14

申请号：EP01830227.3

申请日：2001-03-30

Applicant: STMicroelectronics S.r.l.

Inventor： Pau, Danilo ,  Alfonso, Daniele ,  Basso, Fabrizio ,  Piccinelli, Emiliano ,  Cremonesi, Alessandro

IPC: H04N7/26

CPC classification number: H04N19/59 ,  H04N19/40 ,  H04N19/48 ,  H04N19/90

Abstract: A process and a system for generating, starting from an MPEG input bitstream (IS), an MPEG output bitstream (OS), the output bitstream (OS) having a resolution (Hor/N x Vert/M) modified with respect to the resolution (Hor x Vert) of the input bitstream (IS). The process comprises the operations of:

distinguishing (100), in the input bitstream (IS), first portions and second portions which respectively substantially do not affect and do affect the variation of resolution; and
subjecting (114 to 122) said second portions of the input bitstream (IS) to a function of modification of the resolution obtained by filtration in the domain of the discrete cosine transform (DCT), and then transferring (134), to said output bitstream (OS), said second portions subjected to filtering in the domain of the discrete cosine transform.

The invention also relates to the corresponding computer program product.

Abstract translation: 用于从MPEG输入比特流（IS），MPEG输出比特流（OS）开始生成具有相对于分辨率修改的分辨率（Hor / N×Vert / M）的输出比特流（OS）的处理和系统 （Hor×Vert）的输入比特流（IS）。 该过程包括以下操作：在输入比特流（IS）中区分（100）分别基本上不影响并且确实影响分辨率变化的第一部分和第二部分; 并且输入比特流（IS）的所述第二部分（114至122）具有通过在离散余弦变换（DCT）的域中通过滤波获得的分辨率的修改的功能，然后将（134）传送到所述输出 比特流（OS），所述第二部分在离散余弦变换的域中进行滤波。 本发明还涉及相应的计算机程序产品。

公开(公告)号：EP1003336A1

公开(公告)日：2000-05-24

申请号：EP98830689.0

申请日：1998-11-17

Applicant: STMicroelectronics S.r.l.

Inventor： Piccinelli, Emiliano ,  Pezzoni, Luca ,  Pau, Danilo

IPC: H04N7/36

CPC classification number: H04N19/18 ,  H04N19/107 ,  H04N19/112 ,  H04N19/124 ,  H04N19/137 ,  H04N19/139 ,  H04N19/14 ,  H04N19/149 ,  H04N19/152 ,  H04N19/159 ,  H04N19/172 ,  H04N19/176 ,  H04N19/61

Abstract: The efficiency of a method of coding video data according to a standard wherein several types of fields (P, B) beside reference or Intra (I) fields are identified and various coding options exist for said non-Intra fields (P, B) including the removal of temporal redundancy from non-Intra fields (P, B) by way of an algorithm of motion estimation and identification of predictor macroblocks providing the best approximation according to a prediction mode selectable among respective predicting modes of the different types of non-Intra fields (P, B), calculating the discrete cosine transform (DCT) for blocks of data of said macroblocks of data according to a frame mode of decomposition, wherein the blocks are composed of lines of data belonging to a even lines semifield and to a odd lines semifield or in a field mode of decomposition, wherein the blocks are composed of lines of data belonging to the same semifield, quantizing and coding the data resulting from said discrete cosine transform (DCT) carried out by blocks of data to be stored, is increased by operating a choice between field or frame modes and motion comparation or no motion comparation with an algorithm of assessment and comparison of significative complexity indexes of macroblocks.

Abstract translation: 视频数据进行译码雅丁到标准worin几个类型的字段（P，B）的旁边参考或内部（I）字段被识别和各种编码选项的方法的效率为所述非帧内的字段（P，B），包括存在 从非内部字段去除时间冗余的（P，B）通过运动估计和预测器提供雅丁的最佳近似的预测模式的不同类型的非内部的respectivement预测模式之中选择的宏块的识别的算法的方式 字段（P，B），计算离散余弦变换（DCT）用于数据的雅丁分解的帧模式中，所述宏块的数据的块，worin的块由属于偶数行半字段和一个数据行的 奇数行半字段或分解的场模式，worin的块由属于同一个半字段，量化和编码数据从所述得到的离散余弦变换的数据线的 （DCT）由要被存储的数据的块进行的，通过操作场或帧模式和运动或无运动比较 - 比较 - 之间的选择与评估和宏块的复杂性有意义的指标比较的算法增加。

公开(公告)号：EP0944245A1

公开(公告)日：1999-09-22

申请号：EP98830163.6

申请日：1998-03-20

Applicant: STMicroelectronics S.r.l.

Inventor： Piccinelli, Emiliano ,  Pau, Danilo ,  Zuccaro, Amedeo

IPC: H04N5/00 ,  H04N7/36

CPC classification number: H04N19/51 ,  G06T7/207 ,  G06T7/223 ,  H04N19/56

Abstract: By relaying on a temporal correlation among successive pictures beside on a spatial correlation of motion vectors of macroblocks of the currently processed picture and by the use of a hierarchical recursive motion estimation algorithm, the hardware complexity of video coders complying with the MPEG-2 standard can be greatly reduced without an appreciable loss of quality of the video images been transferred.
The method of motion estimation and a hardware embodiment of a coder are described and performance is compared with a prior motion estimation system.

Abstract translation: 通过在当前处理的图像的宏块的运动矢量的空间相关性之间的连续图像之间的时间相关性和通过使用分级递归运动估计算法之间进行中继，根据MPEG-2标准的视频编码器的硬件复杂度可以 大大减少，而不会明显损失已转移视频图像的质量。 描述运动估计的方法和编码器的硬件实施例，并将性能与先前的运动估计系统进行比较。