公开(公告)号：WO1994026068A1

公开(公告)日：1994-11-10

申请号：PCT/EP1994001360

申请日：1994-04-29

Applicant: ALCATEL N.V. ,  CUCCHI, Silvio ,  LAZZER, Maurizio

Inventor： ALCATEL N.V.

IPC: H04N07/13

CPC classification number: H04B14/06 ,  H03M3/00 ,  H04N19/30 ,  H04N19/593 ,  H04N19/60 ,  H04N19/61 ,  H04N19/619 ,  H04N19/63

Abstract: A method for the differential coding and decoding and related circuits are disclosed here. The differential coding is a special type of coding which, instead of coding directly an input signal, carries out the difference between this signal and a predicted signal of it. This allows the reduction of the information to be transmitted. The invented method is capable of further reducing the signal redundancy and therefore the information to be transmitted.

Abstract translation: 这里公开了差分编码和解码及相关电路的方法。 差分编码是一种特殊类型的编码，代替直接编码输入信号，执行该信号与其预测信号之间的差异。 这允许减少要发送的信息。 本发明的方法能够进一步降低信号冗余度，从而进一步降低要发送的信息。

公开(公告)号：WO1994024822A1

公开(公告)日：1994-10-27

申请号：PCT/JP1994000610

申请日：1994-04-11

Applicant: SONY CORPORATION ,  ODA, Tsuyoshi

Inventor： SONY CORPORATION

IPC: H04N07/13

CPC classification number: H04N9/8042 ,  H04N5/781 ,  H04N5/85 ,  H04N5/9264 ,  H04N19/00 ,  H04N19/107 ,  H04N19/115 ,  H04N19/124 ,  H04N19/14 ,  H04N19/146 ,  H04N19/15 ,  H04N19/152 ,  H04N19/172 ,  H04N19/177 ,  H04N19/192 ,  H04N19/423 ,  H04N19/577 ,  H04N19/587 ,  H04N19/59 ,  H04N19/60 ,  H04N19/61

Abstract: Inputted image signals are coded, for example, by performing prediction encoding, DCT conversion, quantization in predetermined quantization step sizes, and variable-length coding so as to generate first coded data. From the amount of the first coded data per unit of time, for example, per frame or GOP and the total amount of data usable, the amount of codes assigned per frame or GOP is determined. On the basis of this amount of codes assigned, the inputted image signals are coded every given period of time to generate second coded data. Thus variable rate coding in which the coding rate changes every period of time is implemented to provide high-quality images which are uniform as a whole without increasing the quantization step sizes of images even though there are continuous frames whose patterns are complicated, unlike a conventional apparatus. The second coded data thus generated are variable rate ones. Therefore, by recording them on an image recording medium, it is possible to utilize a limited capacity of storage effectively, and prolong the recording time of the image recording medium. Then, from this image recording medium, it is possible to reproduce high-quality images which are uniform as a whole.

Abstract translation: 输入图像信号例如通过执行预测编码，DCT转换，预定量化步长的量化和可变长度编码进行编码，以便产生第一编码数据。 从每单位时间的第一编码数据的量，例如每帧或GOP和可用数据的总量，确定每帧或GOP分配的代码量。 基于分配的这个代码量，输入的图像信号每给定一段时间被编码以产生第二编码数据。 因此，实现编码率每个时间周期变化的可变速率编码，以提供整体均匀的高质量图像，而不增加图像的量化步长，即使存在图案复杂的连续帧，这与传统的 仪器。 这样生成的第二编码数据是可变速率数据。 因此，通过将它们记录在图像记录介质上，可以有效地利用有限的存储容量，并延长图像记录介质的记录时间。 然后，从该图像记录介质，可以再现整体上均匀的高质量图像。

公开(公告)号：WO1994000951A1

公开(公告)日：1994-01-06

申请号：PCT/GB1993001352

申请日：1993-06-29

Applicant: BRITISH TELECOMMUNICATIONS PUBLIC LIMITED COMPANY ,  WELSH, William, John

Inventor： BRITISH TELECOMMUNICATIONS PUBLIC LIMITED COMPANY

IPC: H04N07/13

CPC classification number: H04N19/649 ,  H04N19/20 ,  H04N19/30 ,  H04N19/94

Abstract: Video signals are coded (102) at a first relatively low frame rate. One or more portions of the image area, for example the mouth of a person speaker; recognised by feature extractor (106), are coded (110) at a higher frame rate. Preferably the second coding operates by resolving the pixel values of the image portion - considered as a vector - relative to a set of orthogonal reference vectors (stored in 112) to produce coefficients which are then transmitted. Preferably, the feature extractor (106) normalises the image portion as regards scaling, position and rotation and sends data on these parameters as side information.

Abstract translation: 视频信号以第一相对较低的帧速率被编码（102）。 图像区域的一个或多个部分，例如人的扬声器的嘴; 由特征提取器（106）识别，以更高的帧速率被编码（110）。 优选地，第二编码通过相对于一组正交参考向量（存储在112中）来解析被认为是向量的图像部分的像素值来产生随后被发送的系数。 优选地，特征提取器（106）关于缩放，位置和旋转对图像部分进行标准化，并且将这些参数上的数据作为辅助信息发送。

公开(公告)号：WO1993002527A1

公开(公告)日：1993-02-04

申请号：PCT/CH1992000149

申请日：1992-07-16

Applicant: LABORATOIRE DE TRAITEMENT DES SIGNAUX ,  KUNT, Murat ,  POPAT, Ashok, C. ,  NICOULIN, André ,  BASSO, Andrea ,  LI, Wei

Inventor： LABORATOIRE DE TRAITEMENT DES SIGNAUX

IPC: H04N07/13

CPC classification number: H04N19/126 ,  H04N19/124 ,  H04N19/60

Abstract: A method for quantizing and entropy coding coefficients resulting from sub-band decomposition or transform, in order to dispense with the variable occupancy, feedback-controlled buffer memory which is usually necessary in such systems when coded bits must be transmitted at a constant fixed frequency. This method results in reduced complexity and an improved uniformity of quantization accuracy. The method comprises two steps. The first involves a joint retrieval of a random sample of the set of sub-band/transform coefficients and of a corresponding set of probability density functions representative of the known statistical characters of the coefficients. In the second step, the resulting joint random sample is subjected to an iterative procedure to determine the quantization increment which results in the desired overall bit rate. The margin of error for said quantization increment comprises both a statistical component (stemming from the limited sample size) and a deterministic component (reflecting the imperfection in the adequacy of the iterative procedure).

Abstract translation: 一种用于量化和熵编码由子带分解或变换产生的系数的方法，以便在编码比特必须以恒定的固定频率发送时，省略通过必要的可变占用反馈控制的缓冲存储器。 这种方法导致复杂度降低并且提高了量化精度的均匀性。 该方法包括两个步骤。 第一个涉及一组子带/变换系数的随机样本和表示系数的已知统计特征的相应的概率密度函数集的联合检索。 在第二步骤中，所得到的联合随机样本经历迭代过程以确定产生期望的总比特率的量化增量。 所述量化增量的误差幅度包括统计分量（源于有限样本量）和确定性分量（反映迭代过程的充分性的不完备）。

公开(公告)号：WO1993002526A1

公开(公告)日：1993-02-04

申请号：PCT/CH1992000148

申请日：1992-07-16

Applicant: LABORATOIRE DE TRAITEMENT DES SIGNAUX ,  KUNT, Murat ,  DUFAUX, Frédéric ,  MOCCAGATTA, Iole ,  EBRAHIMI, Touradj ,  CAMPBELL, George ,  GEURTZ, Alexander

Inventor： LABORATOIRE DE TRAITEMENT DES SIGNAUX

IPC: H04N07/13

CPC classification number: H04N19/94 ,  H04N19/61 ,  H04N19/63 ,  H04N19/85

Abstract: A method comprising an image decomposition step using multiresolution pyramidal Gabor transformation. The transformed image data arranged in a pyramid of sub-bands are then selected in three groups according to the spatial frequency level of the sub-band data, wherein the sub-band data of each group are coded by a respective one of three coding operations which are performed in parallel and each of which has a specific performance matching the properties of the data in each group. A reverse transformation operation is then performed to recreate the images. The method uses movement-compensation prediction to reduce the time correlation between the two images without movement vector transmission to the receiver being absolutely necessary. The method further uses filters providing decomposition into sub-bands which only comprise value coefficients that are powers of two or the sum of or difference between two powers of two. A previous step allows the data for processing to be converted from interlaced (CCIR 601 format) to progressive modes, based on a movement-compensated interpolation. A final progressive to interlaced mode conversion step may be provided in the decoder.

Abstract translation: 一种包括使用多分辨锥度Gabor变换的图像分解步骤的方法。 然后根据子带数据的空间频率水平，分三组选择排列在子带的金字塔中的变换后的图像数据，其中各组的子带数据由三个编码操作 它们并行执行，并且每个具有与每个组中的数据的属性匹配的特定性能。 然后执行逆变换操作以重新创建图像。 该方法使用运动补偿预测来减少两个图像之间的时间相关性，而不需要对接收机的矢量传输。 该方法还使用提供分解成子带的滤波器，其仅包括作为2的幂的值系数或者两者的和之和或两者之间的差。 先前的步骤允许基于运动补偿插值将处理数据从隔行扫描（CCIR 601格式）转换成逐行模式。 可以在解码器中提供最后的逐行扫描模式转换步骤。

公开(公告)号：WO1992010909A1

公开(公告)日：1992-06-25

申请号：PCT/EP1991002320

申请日：1991-12-05

Applicant: DEUTSCHE ITT INDUSTRIES GMBH ,  BUTERA, Bill

Inventor： DEUTSCHE ITT INDUSTRIES GMBH

IPC: H04N07/13

CPC classification number: H04N7/122 ,  H04N19/51 ,  H04N19/90

Abstract: The video data of a frame sequence are coded and decoded by batches, whereas each nth main frame data field (F1, F4) (= key frame) is transmitted as completely as possible. From two adjacent main frame data fields (F1, F4), n-1 phantom image data fields (P1, P2) are calculated at the emitter and receiver sides by means of n-1 motion fields (M1, M2). In addition, n-1 differential frame data fields (D1, D2) are determined on the emitter side, then transmitted with the main frame data fields (F1, F4). On the receiver side, the original sequence of the data fields (F1, F2, F3, F4,...) is again reconstructed from the transmitted data fields (F1, D1, D2, F4, ...) is again reconstructed from the transmitted data fields (F1, D1, D2, F4, ...) by a renewed calculation of the motion fields (M1, M2).

Abstract translation: 帧序列的视频数据通过批次进行编码和解码，而每第n个主帧数据字段（F1，F4）（=关键帧）尽可能完整地发送。 从两个相邻的主帧数据场（F1，F4），通过n-1个运动场（M1，M2）在发射机和接收机侧计算n-1个幻像图像数据场（P1，P2）。 另外，在发射极侧确定n-1个差分帧数据场（D1，D2），然后与主帧数据字段（F1，F4）一起发送。 在接收机侧，从发送的数据字段（F1，D1，D2，F4，...）再次重建数据字段（F1，F2，F3，F4，...）的原始序列， 通过运动场（M1，M2）的更新计算，发送的数据场（F1，D1，D2，F4，...）。

公开(公告)号：WO1990013971A1

公开(公告)日：1990-11-15

申请号：PCT/EP1990000645

申请日：1990-04-21

Applicant: DEUTSCHE THOMSON-BRANDT GMBH ,  KEESEN, Heinz-Werner

Inventor： DEUTSCHE THOMSON-BRANDT GMBH

IPC: H04N07/13

CPC classification number: G10L19/02 ,  H04N19/527 ,  H04N19/86 ,  H04N19/98

Abstract: The signals are processed in blocks. When signals are coded and decoded in blocks, the limits of the blocks may be visible or audible during reproduction of certain picture or audio signals, causing interference. By post-filtering (471, 472, 473) transformed, quantized and inversely transformed signal scanning values in the coder, or transformed and inversely transformed signal scanning values in the decoder, the visibility or audibility of signal jumps at the limits of blocks can be reduced.

公开(公告)号：WO1989004577A1

公开(公告)日：1989-05-18

申请号：PCT/EP1988001021

申请日：1988-11-11

Applicant: KEHLER, Waldemar

IPC: H04N07/13

CPC classification number: H04N19/124 ,  H04N19/50

Abstract: Generalized range prediction code modulation system for groups of signals with range overlap top-down (Cauchy's convergence principle) and bottom-up for successive range-dependent, sharper, reduced-length coding of the (multidimensional) function-address space, using the ranges of signals, differences, vectors, predicates, unsharpness, relative offsets and code length prediction. One aim of the invention is PCM coding with substantially reduced data rate and simple operations in comparison with conventional processes.

公开(公告)号：WO1988008237A1

公开(公告)日：1988-10-20

申请号：PCT/US1988001157

申请日：1988-04-13

Applicant: GENERAL ELECTRIC COMPANY

Inventor： GENERAL ELECTRIC COMPANY ,  SCHIFF, Leonard, Norman

IPC: H04N07/13

CPC classification number: H04N19/152 ,  H04N19/124 ,  H04N19/50 ,  H04N19/503

Abstract: A digital television transmitter has a buffer (28) that receives a low resolution low pass filtered signal (26). A high resolution high pass filtered signal (30) is applied to the buffer only when the buffer can accept it without overflowing. A digital television receiver receives a digital television picture signal (32') and preferably a transmitter buffer occupancy state signal (34'). A low resolution low pass filtered signal (22') is applied to an adder (68). A high resolution high pass filtered signal (24') is applied to the adder when the buffer state is low and also when it is intermediate and no substantial motion is present. The transmitter and receiver are useful in video teleconferencing.

公开(公告)号：WO1988007311A1

公开(公告)日：1988-09-22

申请号：PCT/US1988000533

申请日：1988-02-26

Applicant: EASTMAN KODAK COMPANY

Inventor： EASTMAN KODAK COMPANY ,  LIN, Shaw-Yueh

IPC: H04N07/13

CPC classification number: H04N5/945 ,  G11B20/1809 ,  H03M13/00 ,  H03M13/35 ,  H03M13/43

Abstract: The sample values of a digitized image (14) are encoded utilizing the amount of redundancy in the digital data to generate check bits which are attached to blocks of the data prior to transmission or recording (Fig. 2A). After transmission or recording, the recovered data is re-encoded using the original encoding rules, and any resultant errors which have changed the redundancy are reflected in the re-encoded check bits (Fig. 2B). The original check bits attached to the data are compared to the re-encoded check bits (26) to determine if the errors are correctable (27), and, if so, what steps must be taken to correct the errors (30) and to restore the original redundancy to the recovered data (31).