公开(公告)号：EP1005233A1

公开(公告)日：2000-05-31

申请号：EP98830599.1

申请日：1998-10-12

Applicant: STMicroelectronics S.r.l.

Inventor： Bagni, Daniele ,  De Bei, Mattia ,  Mian, Gian Antonio ,  Sacchi, Maria Luisa

IPC: H04N7/50

CPC classification number: H04N19/174 ,  H04N19/115 ,  H04N19/126 ,  H04N19/137 ,  H04N19/14 ,  H04N19/147 ,  H04N19/152 ,  H04N19/172 ,  H04N19/176 ,  H04N19/61

Abstract: An algorithm based on a pre-analysis of just few slices (GOS) of the current whole picture and/or on a mix of information on the pre-analysis of a slice of the preceding picture and on the actual encoding data of the preceding whole picture is able to implement an efficient constant bit-rate (BR) control with a reduced requisite of buffer memory capacity and a proportionally less costly hardware.
The pre-analysis may be carried out by precoding the GOS with a constant reference quantizer or by entropy computation and the local control of the bit-rate is implemented by way of an integrative-proportional controller.

Abstract translation: 基于对当前整个图像的仅几个切片（GOS）的预分析和/或关于前一图像的切片的预分析的信息的混合以及前一整体的实际编码数据的算法 图像能够实现有效的恒定比特率（BR）控制，减少了缓冲存储器容量的要求，并且成本较低的硬件。 可以通过用恒定的参考量化器预处理GOS或通过熵计算来进行预分析，并且通过积分比例控制器来实现比特率的本地控制。

公开(公告)号：EP0979011A1

公开(公告)日：2000-02-09

申请号：EP98830484.6

申请日：1998-08-06

Applicant: STMicroelectronics S.r.l.

Inventor： Pau, Danilo ,  Bagni, Daniele ,  Pezzoni, Luca

IPC: H04N7/36

CPC classification number: G06F17/30811 ,  H04N5/147 ,  H04N19/105 ,  H04N19/137 ,  H04N19/14 ,  H04N19/142 ,  H04N19/149 ,  H04N19/15 ,  H04N19/152 ,  H04N19/172 ,  H04N19/176 ,  H04N19/179 ,  H04N19/503 ,  H04N19/51 ,  H04N19/577 ,  H04N19/61 ,  H04N19/87

Abstract: The compression and coding of digital data pertaining to video sequences of pictures including motion estimation for removing temporal redundance greatly benefit from an ability to recognize the occurrence of a change of scene in order to command the prediction computation of the pictures using only a forward motion estimation for pictures preceding the change of scene and only a backward motion estimation for picture subsequent the change of scene of a given sequence. According to the method of the invention, a change of scene is reliably detected by checking two distinct indexes. Spurious detections in the event of noisy pictures, zooming and other situations that may affect either one or the other index are prevented.

Abstract translation: 与包括用于去除时间冗余的运动估计的图像的视频序列相关的数字数据的压缩和编码极大地受益于识别场景变化的发生的能力，以便仅使用前向运动估计命令图像的预测计算 对于场景变化之前的图像，以及在给定序列的场景变化之后的图像的后向运动估计。 根据本发明的方法，通过检查两个不同的索引可靠地检测场景的变化。 如果发生嘈杂的图像，变焦和其他可能影响其中一个或其他索引的情况，则会发生杂散检测。

公开(公告)号：EP1372113B1

公开(公告)日：2005-10-19

申请号：EP02012891.4

申请日：2002-06-11

Applicant: STMicroelectronics S.r.l.

Inventor： Bagni, Daniele ,  Benussi, Flavio ,  Biffi, Bruno

IPC: G06T9/00

CPC classification number: G06T9/005 ,  G06T9/007 ,  H04N19/124 ,  H04N19/149 ,  H04N19/15 ,  H04N19/152 ,  H04N19/159 ,  H04N19/172 ,  H04N19/61

公开(公告)号：EP1372113A1

公开(公告)日：2003-12-17

申请号：EP02012891.4

申请日：2002-06-11

Applicant: STMicroelectronics S.r.l.

Inventor： Bagni, Daniele ,  Benussi, Flavio ,  Biffi, Bruno

IPC: G06T9/00

CPC classification number: G06T9/005 ,  G06T9/007 ,  H04N19/124 ,  H04N19/149 ,  H04N19/15 ,  H04N19/152 ,  H04N19/159 ,  H04N19/172 ,  H04N19/61

Abstract: A method for controlling the bit-rate of a bitstream (OS) of encoded video signals at a variable bitrate (VBR), the bitstream being generated by compressing a video sequence of moving pictures (IS), wherein each picture comprises a plurality of macroblocks of pixels compressed by any of transform coding (DCT), temporal prediction, bi-dimensional motion compensated interpolation or combinations thereof, to produce any of I and/or P and/or B frames, the method involving quantization of said macroblocks effected as a function of a quantization parameter ( mquant ). The method includes the steps of:

defining a target bit-rate ( target_bitrate ) as well as maximum positive ( max_err_pos ) and negative ( max_err_neg ) error values between the target bit-rate and an average value of the current bit-rate of the bitstream (OS),
controlling the current bit-rate in order to constrain it between said maximum positive and negative error values, and
defining an allowed range of variation for updating at least one reference parameter ( Q base ) representative of the average value of the quantization parameter (mquant) over each picture, wherein said allowed range is determined as a function of the target bit-rate ( target_bitrate ) and the maximum positive and negative error values ( max_err_pos, max_err_neg ).

Abstract translation: 一种用于以可变比特率（VBR）控制编码视频信号的比特流（OS）的比特率的方法，通过压缩运动图像（IS）的视频序列来产生比特流，其中每个图像包括多个宏块 通过变换编码（DCT），时间预测，双维运动补偿内插或其组合中的任何一个压缩的像素，以产生I和/或P和/或B帧中的任何一个，所述方法涉及所述宏块的量化， 函数的量化参数（mquant）。 该方法包括以下步骤：在目标比特率和比特流的当前比特率的平均值之间定义目标比特率（目标比特率）以及最大正（max_err_pos）和负（max_err_neg）误差值 （OS），控制当前比特率以便将其限制在所述最大正值和负误差值之间，并且定义允许的变化范围，用于更新表示量化参数的平均值的至少一个参考参数（Qbase） （mquant），其中所述允许范围被确定为目标比特率（目标比特率）和最大正和负误差值（max_err_pos，max_err_neg）的函数。

公开(公告)号：EP1843587A1

公开(公告)日：2007-10-10

申请号：EP06425235.6

申请日：2006-04-05

Applicant: STMicroelectronics S.r.l.

Inventor： Alfonso, Daniele ,  Bagni, Daniele ,  Rovati, Fabrizio

IPC: H04N7/01

CPC classification number: H04N7/014 ,  H04N7/0105

Abstract: In a method for converting an input video sequence (SV), comprising digital images (305, 305') organized in frames and operating at a variable frame-rate (v), into an output video sequence (SC), operating at a preset constant frame-rate (c), it is envisaged to store (215) the input video sequence (SV) temporarily and to control fetching of images (305, 305') of said temporarily stored input video sequence (SV). The invention moreover envisages:
controlling (220) fetching of images (305, 305') from the temporarily stored input video sequence (SV) by adjusting an emptying rate (f) to form an intermediate video sequence (SF); and
carrying out an operation of motion-compensated interpolation (225) on the intermediate video sequence (SF) to form the output video sequence (SC) operating at a pre-set constant frame-rate (c), the emptying rate (f) being adjusted (220) as a function of a number (b; b1) of images of the input video sequence (SV) with variable frame-rate (v) temporarily stored (b; b1).
Preferential application is in the reproduction of video sequences in mobile terminals, with the capacity of audio-video conversation, for example for videoconferences.

Abstract translation: 一种用于转换输入视频序列（SV）的方法，包括以帧组织并以可变帧速率（v）操作的数字图像（305,305'）转换成输出视频序列（SC），在预设 恒定帧率（c），设想临时存储（215）输入视频序列（SV）并控制所述临时存储的输入视频序列（SV）的图像（305,305'）的获取。 本发明还设想：通过调节排空速率（f）来控制（220）从临时存储的输入视频序列（SV）中提取图像（305,305'）以形成中间视频序列（SF）; 并且在中间视频序列（SF）上执行运动补偿内插（225）的操作以形成以预定的恒定帧速率（c）操作的输出视频序列（SC），排空速率（f） 作为临时存储（b; b1）的可变帧速率（v）的输入视频序列（SV）的图像数（b; b1）的函数被调整（220）。 优先应用在移动终端中的视频序列的再现中，具有音频 - 视频对话的能力，例如用于视频会议。

公开(公告)号：EP1107609A1

公开(公告)日：2001-06-13

申请号：EP99830751.6

申请日：1999-12-02

Applicant: STMicroelectronics S.r.l.

Inventor： Bagni, Daniele ,  Battistelli, Luca

IPC: H04N7/36 ,  H04N7/26

CPC classification number: H04N19/137 ,  H04N19/105 ,  H04N19/112

Abstract: A method of processing a bitstream of coded data of video sequences of progressive or interlaced pictures divisible in a top half-frame and in a bottom half-frame that comprises estimating motion vectors of groups of pixels belonging to said top half-frame of the current picture in relation to pixels belonging to the bottom half-frame of the preceding picture (MV TOP X,Y (K)) and motion vectors of group of pixels of said bottom half-frame of the current picture in relation to pixels belonging to the top half-frame of the current picture (MV BOT X,Y (K)), respectively, characterized in that it comprises calculating for each macroblock of said top half-frame and a bottom half-frame a respective top motion coefficient (V TOP X,Y (K)) and a bottom motion coefficient (V BOT X,Y (K)), depending on the estimation of the motion vectors of top half-frame (MV TOP X,Y (K)) and bottom half-frame (MV BOT X,Y (K)); recognizing the current picture as an interlaced picture by a substantial equality of the distributions of values of said motion coefficients (V TOP X,Y (K), V BOT X,Y (K)) or as a progressive picture by a substantial inequality of the distributions of values of said motion coefficients (V TOP X,Y (K), V BOT X,Y (K)).

Abstract translation: 处理逐行或隔行画面的视频序列的编码数据的比特流的被分割成上半帧和下半帧的方法，做了包括估计属于当前的所述顶半帧的像素组的运动矢量 图像相对于属于preceding-图片（MVTOPX，Y（K））和属于上半部分的当前图像的相对于所述底部半帧的像素的组的运动矢量以像素的下半帧的像素 当前画面的-frame（MVBOTX，Y（K）），分别，其特征在于在这样做是包括计算用于所述上半帧的每个宏块和底半帧一个respectivement顶部运动系数（VTOPX，Y（K） ）和底部运动系数（VBOTX，Y（K）），这取决于顶半帧（MVTOPX，Y（K））和下半帧的运动向量的估计（MVBOTX，Y（K））; 由所述运动系数（VTOPX，Y（K），VBOTX，Y（K））或值的分布的相当平等识别当前画面为隔行扫描图像由值的分布的相当不等式一渐进图象 所述系数（VTOPX，Y（K），VBOTX，Y（K））的运动。