CN1217125A - Vedio data encoder and decoder - Google Patents
Vedio data encoder and decoder Download PDFInfo
- Publication number
- CN1217125A CN1217125A CN 96180260 CN96180260A CN1217125A CN 1217125 A CN1217125 A CN 1217125A CN 96180260 CN96180260 CN 96180260 CN 96180260 A CN96180260 A CN 96180260A CN 1217125 A CN1217125 A CN 1217125A
- Authority
- CN
- China
- Prior art keywords
- frame
- video data
- zone
- prediction
- predication
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
This invention relates to compression of prediction error in motion compensated video coding schemes. Prediction frame and motion vectors are used to extract useful information about the properties of the prediction error signal, e. g. its location, directionality, etc. This information does not need to be transmitted because the decoder of the proposed system includes means to extract this information. Information extracted by the encoder and decoder is used to adapt prediction error coding scheme which gives substantial bitrate reductions.
Description
The present invention relates to a kind of encoder and a kind of decoder that the video data of motion compensation encoding is decoded that is used to carry out the motion compensation encoding of video data.The invention still further relates to a kind of motion compensation encoding method and a kind of method that the video data of motion compensation encoding is decoded of video data.
Motion compensated video encoder and decoder are to utilize the successive frame of traditional video order to demonstrate the fact of high similarity and the fact that the difference between the successive frame is mainly produced by the motion in the scene mutually.Thus, motion compensated video coding can the compressed video data amount, for example can be by IEEE, and volume 73, No.4, P523-548, April 1994, H.Mussmann, P.Pirsch and H.Grallert: can recognize in " development of picture coding ".Replace each frame I (x of compressed video order respectively, y), the MC video encoder is compressed in the current video frame that must be encoded and calculates in encoder, P (the x that is called as predictive frame, y) difference between, and represent current video frame I (x, estimated value y) according to the frame of video of one or more fronts.Video Decoder can utilize (the x at current video frame I, y) and predictive frame P (x, y) difference between, by carry out that the frame similar to encoder predicted and by according to received at predictive frame P (x, y) and actual frame I (x, y) difference between is proofreaied and correct predictive frame, build again current video frame I (x, y).
In field of video encoding, it is common using different coded systems to come the zones of different of image is encoded.This such as ITU H.261 and H.263 and in all modern codec of MPEG-1 and MPEG-2 video encoding standard as seen.For example, Tu Xiang some zone can not must be used any temporary transient prediction (being called as intermediate mass (intrablock)) and is encoded.The zone of the frame of term therefore, " frame " representative carrying out motion compensation encoding and decoding.It should be noted that in the context of motion compensation encoding and decoding, term " frame " is not limited to not exist the particular case of intermediate mass.
Encoder all is that (estimated motion vector calculates by estimation in encoder for x, prediction y) according to the present frame I based on estimated motion vector.Estimated motion vector is used to make up predictive frame P in encoder (x, y), based on it, encoder can calculate predication error frame E, and (x, y), promptly (x is y) with actual frame I (x, y) difference between at predictive frame P.In addition, predication error frame E (x, y) and describe present frame I and be transferred to decoder with respect to the estimated motion vector of the estimation pixel displacement of reference frame R and be used for based on motion vector, to the prediction of present frame P with based on the correction of the predicated error of being transmitted to predictive frame thus make up present frame I
*(x, y).
In order further to reduce the data volume that is transferred to decoder, encoder compresses predicated error.As to predicated error compression and result that the predicated error of compression is decompressed, the compressed prediction error frame E of decompression
*(x, y) may somewhat different than E (x, y).Encoder consider since based on estimation and predictive frame P (x, y) calculating and predication error frame E (x, y) calculating is not based on the primitive frame of front, and is based on decoder to build the previous frame that similar mode is built again again to decoder, the difference a little that is caused.
For estimation, build frame I again
*Or primitive frame I can adopt.
(x, compression y) is normally by being separated into entire frame in less zone and by provide a suitable compression method to realize for each zone for predication error frame En.Term " area " is meant in a frame some pixels of forming the adjacent two-dimensional block in space.Predicated error typical case only concentrates in seldom the zone, yet in remaining area, it is very little or be zero.Therefore, the compressed prediction error data generally include two category informations, i.e. the value of compressed prediction error in the addressing information in the zone of the predication error frame that appointment is encoded and this external these zones.
From Proc.Of 1992 European Conference on Singnalling ProcessingEUSIPCO, P1357-1360, M.Moresco, F.Lavagetto and F.Cocurullo, can recognize a kind of algorithm that is used for the motion adaptive video coding in " the Motion Adaptive vector quantization of video coding ", according to it, the video incoming frame is subdivided into three-dimensional (space-time) piece of fixed size, and is cut apart according to the motor activity in each piece.According to the content of motion, in time these pieces are sampled again by the frame that jumps over some.The sheet of the two-dimensional space of the piece that does not also jump over (slice) is by vector quantization and coding.At the receiver place, the spatial pieces of the piece that each has been transmitted is decoded, the sheet in the middle of it by linear interpolation to build whole again.
The objective of the invention is to improve the Code And Decode of video data, so that can realize higher code efficiency and the bit rate that can further reduce coding video frequency data.
According to the present invention, can carry out the encoder of motion compensation encoding and further realize this purpose according to claim 1 pair video data by one by a decoder of decoding according to the video data of claim 15 pair motion compensation encoding.In addition, this purpose can also be carried out the method for motion compensation encoding according to claim 32 pair video data and realizes by a method of decoding according to the video data of claim 33 pair motion compensation encoding by one.Embodiments of the invention are limited by dependent claims.
According to a first aspect of the invention, the characteristic of the characteristic of predictive frame and/or the motion vector code efficiency that is used to reduce the bit rate of coding video frequency data and is used to improve video data.Predicated error concentrates on the silhouette edge edge of moving object in the video scene.The present invention is based on and observes predictive frame P (i.e. pattern of brightness and/or chroma pixel value as edge and analog, are roughly coincide with the position of these profiles of moving object for x, the y) structure in.In presents, the general structure of this predictive frame of term " discontinuity of image brightness " expression.Encoder knows that successively (x, y), motion vector also is transmitted predictive frame P simultaneously.Therefore, (x y), can judge in encoder which zone is important in predication error frame according to predictive frame P, and will be included in address date in the coding video frequency data and must not transmit by the difference Code And Decode about the zone in these predication error frame.
The present invention is also based on recognizing that predicated error concentrates on the zone of high-speed motion action.Therefore, in encoder can with motion vector can be used to judge and will be encoded by encoder and the important area of the predication error frame of decoded device decoding, and must in coding video frequency data, not comprise address date about the important area of these predication error frame.
Preferred video data coding scheme utilization according to this aspect of the invention can be used for the architectural feature of the predictive frame in the encoder, with the characteristic of in encoder, estimating and be transferred to the motion vector of decoder, with the zone of the predication error frame determining to be encoded by encoder with in decoder, determine to be included in the zone of the predictive frame in the coding video frequency data.
Preferably, for one-frame video data is encoded, predication error frame is split into a plurality of zones.Predictive frame is divided into some zones to be similar to the divided mode of predication error frame, like this for each zone in the predication error frame, all has a corresponding zone in predictive frame.Decoder is cut apart predictive frame in the mode that is similar to encoder, and will restore from the predication error frame of the coded data that is received and be included in each respective regions of the predication error frame in the coded data that is received.These frames can be split into the zone of pre-sizing and shape or be split into the zone of size and shape variable.Frame is divided into the zone can be finished adaptively.The size in each zone and shape can be according to local motion actions and/or according to the partial structurtes feature of predictive frame, as the degree according to the partial discontinuous of predictive frame.
Preferably, after encoder was finished cutting apart of frame in a kind of predetermined or adaptive mode, encoder adopted a kind of predetermined Cahn-Ingold-Prelog sequence rule (sequence rule) to determine the zone of the predication error frame that must be encoded.As an example of this Cahn-Ingold-Prelog sequence rule, to the zone that will be encoded, encoder line by line scan these cut apart by frame gained the zone arrangement and all zones that will be encoded are encoded with their found orders.So, decoder can cutting apart according to predictive frame, according to the information that motion vector extracted of encoding, and recover each respective regions from the predication error frame in the prediction error data that is received according to described Cahn-Ingold-Prelog sequence rule by the zone institute foundation of predictive frame and/or predication error frame.
In addition, according to a second aspect of the invention, the All Ranges of predication error frame or institute's favored area adopt the adaptive coding scheme to encode.Institute's favored area can be the zone that is defined as important predication error frame according to a first aspect of the invention.The adaptive coding scheme can be considered the different qualities kind in each predication error frame zone that will be encoded.The example of adaptive coding scheme comprises the classification vector quantization scheme according to some code books, each code book is a described characteristic kind design, orthogonal transform, match tracing technology (matching pursuit techniques) etc., all these technology all are that the present technique field is known.The classification vector quantization scheme is used for the coding of video data for example in signal processing, Elsevier, volume 26,1992, P265, H.-H.-Chen, " the low rate order image by vector quantization " that Y.-S.Chen and W.-H.Hsu showed or at the international conference ICIP 1994 of IEEE about image processing, volume 2, P452, K.Zhang, " the stable motion estimation and the multi-stage vector quantization that are used for sequential compression " that M.Bober and J.Kittle showed and at IEEE about acoustics, voice, the international conference ICASSP ' 94 of signal processing, volume 5, P417 can learn in " rate distortion analysis based on the motion compensated video encoding and decoding of VQ that block size is variable " that S.Liu showed.According to the present invention, when with the adaptive coding scheme, for example classification vector quantizes, orthogonal transform, when the match tracing scheme is used for the zone of predication error frame encoded, can realize the minimizing of bit rate by gone out classified information (for example encoding scheme selection information) by the feature extraction that can be used for the predictive frame in the encoder, coding video frequency data must not comprise classified information like this.When the adaptive coding in the zone of adopting predictive frame, either-or or except can by with the zone of the corresponding predictive frame in zone of the predication error frame that will be encoded in extract the classified information, the characteristic of motion vector can be used for each coding region of predication error frame is classified, codebook selecting for example, and must in coding video frequency data, not comprise classified information.
According to this aspect of the invention, the classified information of being extracted can indicate one to be used for the specific coding method encoded in each zone of predication error frame, maybe can indicate a different set of coding method.This a kind of favourable enforcement on the one hand according to the present invention, classified information is used for each zone that will be encoded and selects a regional code device in many regional code device, and each all is suitable for the special characteristics of individual features kind.Each regional code device sequentially has one group of available different coding method, and video encoder is selected the best approach and coding method selection information is sent to Video Decoder in this group of methods.If coding method selects information by entropy coding (entropy encode), the coding method that then sends to Video Decoder is selected amount of information can utilize the non-normalized probability distribution that the coding method of each regional code device selects and is kept very low.If for each characteristic kind, promptly for each regional code device, the coding method of one or very few number very may be selected, then when the available a greater number bit of the selection of the coding method with less possibility is indicated to Video Decoder, the selection of the coding method of these or very few number can be indicated to Video Decoder with the lesser amt bit.
In selected video encoder, provide the selection of a specific coding method in all coding methods can be by video encoder according to the disabled information of Video Decoder, for example according to the input of initial video Frame, wait according to each structure of the pixel value in the zone that will be encoded of predication error frame and to realize.So Video Decoder can be used for according to the characteristic of the predictive frame that can be used for Video Decoder with according to selecting the regional code device by the coding method selection information that video encoder transmitted.
This enforcement on the one hand of the present invention is favourable, stay under some situations about the unreliability of the characteristic in each zone that will be encoded of predication error frame in the classification according to the characteristic of the characteristic of predictive frame and/or motion-vector field, it can improve the performance of video encoder and Video Decoder.Utilize this enforcement, even in these cases, the selection information that a high adaptivity in a large number of different coding method and accurate coding method also can very few number sends Video Decoder to.
After probability distribution was selected in the coding method of considering each regional code device, coding selection information was carried out the moisture in the soil coding before being preferably in and sending to Video Decoder.The correction of an equivalence provides a default coding method in many coding methods of each regional code device, and whether the video encoder inspection is satisfied with according to the default encoder method of the regional code device of categorizing selection and is used for the appropriate criteria encoded in each zone of predication error frame.If it is satisfied that the default encoder method of selected regional code device is found running, then not transfer encoding of video encoder method selection information is given Video Decoder, and Video Decoder adopts the default coding/decoding method corresponding to the default coding method that video encoder adopted.Otherwise video encoder is ignored this default coding method and is selected a suitable coding method, and selection information is accordingly sent to Video Decoder.
Subsequently, most preferred embodiment of the present invention is described with reference to the accompanying drawings.
Fig. 1 is the schematic diagram according to first embodiment of motion compensated video encoder of the present invention;
Fig. 2 is the detailed maps according to an embodiment of coded prediction error device of the present invention;
Fig. 3 is the example in a zone of predication error frame;
Fig. 4 is the kind of directivity pattern in predictive frame zone and some examples of subclass;
Fig. 5 is for being an optional regional code device of video encoder according to a second embodiment of the present invention;
Fig. 6 is the example of pixel group;
Fig. 7 is the sub-piece that is used for the calculating pixel range of movement;
Fig. 8 is for being first embodiment according to Video Decoder of the present invention;
Fig. 9 is the detailed maps of predicated error decoder shown in Figure 5;
Figure 10 a illustrates an arrangement in zone;
Figure 10 b illustrates an example of prediction error data;
Figure 11 is for being an optional regional decoding device of a Video Decoder according to a second embodiment of the present invention.
Fig. 1 is the schematic diagram according to an embodiment of motion compensated video encoder of the present invention.In the figure, reference marker 1 expression is used for that (x, y) (x, y) difference between is calculated predication error frame E (x, device y) with predictive frame P by frame of video input I.Reference marker 2 expression coded prediction error pieces, reference marker 3 expression predicated error decoding blocks.Reference marker 5 representative be used for by predictive frame P (x, y) and the predication error frame E of decoding
*(x y) calculates the frame of video I that builds again
*(x, calculation element y).Frame of video I who is used to preserve regeneration of reference marker 6 expressions
*(x y) provides motion compensated predictor 7 to be used for reference frame R (x, frame memory y) that calculates predictive frame with being used to.8 expressions are used for estimated motion vector and the exercise estimator of motion vector to MC prediction piece 7 are provided.4 expressions be used to receive the motion vector that the prediction error data exported by coded prediction error device 2 and exercise estimator 8 exported and be used for multiplexed prediction error data and motion vector to be transferred to the multiplexer of a decoder.
Coded prediction error piece 2 of the present invention not only be used to the predication error frame E that receiving system 1 exported (x, y), also be used to receive predictive frame P (x, y) and the estimated motion vector of exercise estimator 8.In addition, predicated error decoding block 3 not only receives the prediction error data that coded prediction error piece 2 is exported, also receive predictive frame P (x, y) and the estimated motion vector of exercise estimator 8.The video encoder running of Fig. 1 is as follows.For the video data I of each frame input (x, y), encoder all predict a predictive frame P (x, y), and comparison prediction frame P (x, y) and the frame of video I of actual input (x, y).This is relatively finished by calculation element 1.So, predication error frame E (x, the y) prediction error data that is provided for piece 2 and compresses with generation through coded prediction error, and be transferred to decoder.(x, compression y) may cause losing of some information to predication error frame E.Considering that Video Decoder (not showing among Fig. 1) is the prediction error data that is received compression by video encoder, rather than predication error frame E (x, y), then Video Decoder will make the prediction to the actual video frame not be based on predication error frame E (x, y), be based on by to the compressed prediction error data decode and the predication error frame E of regeneration
*(x, y).(video encoder is to obtain regeneration predication error frame E with Video Decoder for x, the y) deviation between for fear of the predictive frame P in encoder
*(x, y) identical mode prediction error data that coded prediction error device 2 is exported is encoded, and prediction is not based on actual video data frame I, and (x y), builds video data frame I and be based on corresponding to video data frame again
*(x y) finishes, and this video data frame can be built by the prediction error data that Video Decoder is exported according to piece 2 again.
Coded prediction error and decoding and to produce a predictive frame according to estimation be that the present technique field is known.Many documents about estimation and motion compensation come forth.The details of these themes is easy to learn, therefore there is no need to go through the problem that these and the present invention do not contact directly.
Coded prediction error device 2 comes by predication error frame being divided into a plurality of zones that (x y) encodes to predication error frame E.In addition, coded prediction error device 2 receives and analyses and prediction frame P (x, y), shown in separately dotted line among Fig. 1.In addition, coded prediction error device 2 receives about the information by the estimated motion vector of exercise estimator 8, shown in separately dotted line among Fig. 1.Thus, coded prediction error device 2 be in one judge predication error frame E (x, which zone y) must be encoded, which zone of predication error frame does not comprise important information and must not be transferred to the position of decoder.Coded prediction error device 2 bases are to the analysis of the characteristic of predictive frame with according to this judgement is finished in the analysis of estimated motion vector, and predictive frame and motion vector are also all available in the decoder of received code video data.(x, y) important information in concentrates on the silhouette edge edge of the object that moves in the video scene at predication error frame E in coded prediction error device 2 utilizations of the present invention.When decoder was not known the accurate profile of moving object, (for example discontinuity of image brightness, edge and analog roughly coincide with the position of these profiles for x, architectural feature y) to can be used for predictive frame P in the encoder.Therefore, the coded prediction error device can (x, architectural feature y) be judged the approximate location of the important area of predication error frame by predictive frame P.Except according to predictive frame P (x, y) architectural feature is judged outside the important area of predication error frame, this embodiment frame by frame of coded prediction error device 2 has been analyzed the characteristic of describing owing to the motion vector of kinetic each pixel displacement in scene, to judge the important area of the predication error frame that will be transferred to decoder.If the judgement of important area is based on the architectural feature of predictive frame, for example edge and discontinuity and based on motion vector, then those zones of predication error frame can judge which has described the motion of the profile of moving object.Alternatively, these judgements of the important area of predication error frame can be based on the architectural feature of predictive frame, or based on motion vector.In these more not preferred examples, still may judge some inessential zones of predication error frame.
In addition, also (x, whether each zone y) adopts an adaptive coding scheme important to judge to coded prediction error device 2 of the present invention to predication error frame E.According to the kind of architectural feature, for example (x, contour direction y) come to be each zone selection encoding scheme separately to predictive frame P.Because predictive frame P is (x, y) all available in video encoder and Video Decoder, so it is favourable selecting the encoding scheme of an adaptation according to these characteristics, even when not comprising other selection information in the coded prediction error information, Video Decoder still can be selected a suitable decoding scheme that mates with the selected encoding scheme separately of video encoder for each zone that will be encoded.
Therefore, coded prediction error device 2 of the present invention is used to judge predication error frame En (x by use, y) with the important area that is encoded and be used to predication error frame E (x, each important area y) select the information of the encoding scheme of an adaptation to utilize predictive frame P in video encoder and Video Decoder
n(x y) and the availability of motion vector, and must not transmit (x, the address information of the position of important area y) or about the information of selected encoding scheme about predication error frame E.Predicated error decoder 3 among Fig. 1 is favourable, but former thereby say optional for principle.If be used for compressed prediction error frame E (x, the encoding scheme of important area y) has caused some information dropouts, the estimation by making motion vector and the prediction of frame of video are not based on actual incoming frame, and are based on the regeneration frame of video I by a decode operation gained identical with the performed decode operation of Video Decoder
*(x, y), the part of the information that these are lost can be transferred to Video Decoder again.This principle is that the present technique field is known, therefore, will omit the further discussion of the importance of predicated error decoder 3 here.The similitude of corresponding blocks of prediction in the predicated error decoder 3 of seeing Fig. 1 and Video Decoder will provide the structural explanation and the running details of predicated error decoder 3 subsequently in the description background according to an embodiment of Video Decoder of the present invention.
Multiplexer 4 is one, and be used for will be by the coded prediction error data of coded prediction error device 2 output with by the multiplexed washability parts of motion vector data of exercise estimator 8 outputs.If the transmission channel separately of prediction error data and motion vector can be used, then can omit multiplexer 4.
Fig. 2 is the detailed maps according to an embodiment of coded prediction error device of the present invention.In Fig. 2,21 one of expression are used for that (x y) is divided into the piece in a plurality of zones with predictive frame P.22 one of expression are used for analyzing at predictive frame P (x, the piece of the space discontinuity in each zone y).Do not limit in the specific embodiment of the scope of the invention at this, this piece comprises (x, frame y) is cut apart the directed high pass filter that filtering operation is carried out in each zone that is produced by P.Another selection that is used to implement piece 22 comprises edge detection algorithm or to the algorithm of image brightness value assignment.Piece 24 be used to according to predictive frame P (x, y) cut apart that (x y) is divided into a plurality of zones with this predication error frame E.
In addition, piece 25 is used for a predetermined Cahn-Ingold-Prelog sequence rule (sequence rule) important area of predication error frame E is exported to piece 26, to compress and to send Video Decoder (this not shown in the figures) to.Especially, in giving specific embodiment, piece 25 will be arranged as an arrangement in zone by the All Ranges that cutting operation produced of piece 21, and this arrangement of lining by line scan is to find important area.Then, each found important area is output to piece 26 immediately.Thus, Video Decoder can be to the zone location of the predication error frame that received, and must not receive the information about the relevant position, and this will make an explanation subsequently.The frame of predictive frame P is cut apart and is that the identical mode of predication error frame E carries out to cut apart with piece 24 by piece 21, and each regional importance of predication error frame E is judged the characteristic that is based on predictive frame in this zone like this.In other words, piece 21 and piece 24 performed frames are cut apart in such a way and are carried out, make the respective regions of predication error frame and predictive frame comprise one group of identical pixel coordinate (x, y).This predictive frame P and predication error frame E are split into has the same block size, for example 4 * 4 pixels or 8 * 8 pixels, the zone.It should be noted that cutting apart can be according to predetermined pattern or can be based on the division (segmentation) of the predetermined image that is divided into irregularly shaped zone.Provide a description the information that current encoded frame is split into arbitrary shaped region at these encoders and give in the scheme of decoder, aforesaid cutting apart can be based on this division.
Each zone that is used for will being encoded to predication error frame E of reference marker 26 expression, promptly piece 25 is judged to be each important zone, compresses and encodes.Adaptive region encoder 26 comprises a classification block 23, and this classification block that does not limit the scope of the invention in this specific embodiment is categorized as a kind with each important area of predictive frame: no continuity; Level continuous; Vertical continuous; Diagonal angle, left side oblique line continuity; Right diagonal angle oblique line continuity; The non-directional continuity.Each the regional directional characteristic that is based on predictive frame P about the classification in each zone of predictive frame E carries out, and these characteristics are analyzed and extracted by the discontinuity analyzer in piece 22.One of reference marker 28 expression is used for each regional categorizing selection according to predictive frame E at many optional regional code devices 271 ..., regional code device among the 27n is encoded in order to each zone to predication error frame E.29 expressions are used to select the output of selected regional code device to be connected to be transferred to the selector of Video Decoder with the output of video encoder with the output that is used for institute's favored area.
Fig. 3 has shown predictive frame P
nA zone, in order to illustrate the running of directed high pass filter among Fig. 2.The pixel of each piece representative in the estimation range of predictive frame P of being got ready among Fig. 3.(x, y) represents pixel coordinates is (x, the brightness value of each pixel y) to P.
Level that is produced and vertical difference F
H(x, y) and F
V(x y) represents each location of pixels (x, y) the space discontinuity on vertical and horizontal direction respectively.
Each can further be further divided into some subclass according to the position of corresponding discontinuity in this zone to directed discontinuity classification b in e.Lifting classification b is example, and the first subclass b1 has covered horizontal discontinuity in this regional the first half, and subclass b2 has covered horizontal discontinuity in this regional mid portion.Subclass b3 has covered the horizontal discontinuity of the next part that is arranged in this zone.
Similar, for classification c, d and e, these classifications divide respectively vertical and diagonal position again according to discontinuity, by Fig. 4 obviously as seen.It should be noted that categorical measure certainly for example changes according to the quantity of optional regional code device 271 to 27n with the correlation method that is used for the compression of predictive frame zone.Equally, many different algorithms also can be used for according to classifying in each zone at this regional filter output value.The example of these algorithms for example can be by M.K.Jain, and Prentice Hall International 1989 can learn in pp.347-357 " basic principle of digital image processing ".
The classification of each regional importance of predictive frame P of being judged according to piece among Fig. 2 25, so, grader 23 is selected a suitable regional code device in optional regional code device 271 to 27n, for example a suitable vector quantizer is used for adaptive coding is carried out in the zone of discussing of predication error frame E.In this special embodiment,, provide the special-purpose vector quantizer of the member's who is adapted to each classification or subclass certain statistical characteristic for each classification or subclass.Thus, vector quantizer also is a prior art for the compressed signal with some information losses.Vector quantizer VQ1 ... each vector quantizer is used in the zone that the best approximation of being found in the code book that belongs to this specific vector quantizer replaces predication error frame En among the VQn.Described code book is that a group of producing in respective classes or subclass of most probable is seldom represented, and it is enough that coefficient of vector quantizer output is given the optimum Match code vector in belonging to the code book of this specific vector quantizer.
According to the present invention, the classification in the zone of predication error frame is based on corresponding predictive frame zone.This predictive frame is also available in Video Decoder, needn't comprise classified information in the coding video frequency data like this.But this classified information can be extracted by the similar sort operation of the sort operation of carrying out and encoder is performed by Video Decoder.
Here it should be noted that other compression methods in predication error frame zone can adopt, and for example orthogonal transform technology (DCT, KLT) or the match tracing technology.
Another embodiment according to motion compensated video encoder of the present invention will be discussed subsequently.First embodiment of present embodiment and above-mentioned video encoder is similar, except following part.
Fig. 5 has shown each a embodiment who is applicable to according in the optional regional code device 271 to 27n among second embodiment of video encoder of the present invention.As shown in Figure 5, according to the present invention, each in the optional regional code device 271 to 27n comprises and a plurality ofly is used to carry out the encoder 531 of a plurality of different coding methods to 53k.In the optional regional code device 271 to 27n each comprises also that in addition one is used for selecting one to encode with a respective regions to predication error frame according to an optimal encoder of proper standard in a plurality of corresponding encoder 531 to 53k.Do not limit the scope that the present invention judges according to the selection control 51 of this specific embodiment, wherein can be used for the respective regions to predication error frame E that the coding method by grader 23 selected optional regional code devices makes can encode by minimum encoding error.For this reason, selection control 51 receives predication error frame E so that can visit by each caused encoding error in the available encoder 531 to 53k of grader 23 selected optional regional code devices.If selection control 51 finds to have in the encoder 531 to 53k the encoder performance more than satisfied, selection control 51 selects to realize with lowest bit quantity the encoder of satisfactory performance in these encoders.Selection control 51 output selection information encode for selector 54 to the respective regions with being encoded of predication error frame to drive selected encoder.In addition, the selection information of selection control 51 outputs is provided for moisture in the soil encoder 52, is used for selection information is carried out the moisture in the soil coding.In addition, select information Control selector 55 to link to each other with compressed prediction error data multiplex with moisture in the soil coding selection information and the output of selected encoder with a input with selected encoder one multiplexer 56 in the encoder 531 to 53k.The output of multiplexer 56 is the output of a corresponding encoded device of optional regional code device 271 to 27n.
In this embodiment according to video encoder of the present invention, grader 23 shown in Figure 2 comes the square area of predictive frame P is classified based on the variation of the image brightness in predictive frame and corresponding each zone, the zone separately predication error frame that will be encoded.In this specific embodiment, the variable quantity of image brightness value is by the 22 following calculating of discontinuity analyzer:
Wherein (x is that (i, the brightness value of the pixel of j) locating, M are at the every row of square area and the pixel quantity of every row, and q is the mean value of the brightness value of pixel in each predictive frame zone for each regional coordinate at predictive frame y) to p.
So discontinuity analyzer 22 calculates group's amount of exercise (cluster activity measure) CAM for a plurality of intended pixel groups in each zone of predictive frame.Each pixel group is corresponding to a specific classification, and this particular category can be distinguished by grader 23.Fig. 6 has shown by way of example can be by 13 different pixels groups of the different qualities classification of grader 23 difference corresponding to 13.The pixel that belongs to corresponding group of the group 1 to 13 shows with shade in Fig. 6.
In each group, group's amount of exercise CAM is calculated as follows in this specific embodiment:
For position in respective regions (i, each pixel of j) locating, by analyze four comprise pixel (i, 2 * 2 sub-block of pixels j) come calculating pixel amount of exercise pam (i, j).Shown among Fig. 7 for pixel (i, the position that j) this is a little.So, the variation of calculating the pixel value in each sub-piece.Varlu is used in these variations, varru, and varll and varrl represent.If a sub-piece pixel is in the outside of this respective regions, promptly (i, when j) being this regional boundary pixel, then to be considered to be 0 in the variation of this sub-piece when the position.Then in this position (i, pixel motion amount j) is:
pam(i,j)=max(varlu,varll,vartu,varrl) (2)
For each group (a), this group amount of exercise cam (a) can be calculated as follows
In the formula, the pixel motion amount of Ca representative all pixels in group a and with the product of the normalization factor w (a) of group a.For group a=1 to a=13 shown in Figure 6, the value of w (a) can be selected as follows
In this specific embodiment, 14 classifications of grader 23 difference, and each zone that will be encoded is according in these 14 classifications are classified according to the position of the main discontinuity of the image brightness in each zone of predictive frame.13 classifications correspond respectively among 13 groups shown in Figure 6.Do not show the 14th classification among Fig. 6, it is to be assigned to the zone that does not comprise any significant discontinuity.
Operation corresponding to the classification in the zone of the predictive frame in a zone of the predication error frame that will be encoded can be carried out as described below.If the variation in the zone of calculating according to formula (1) is less than predetermined threshold, then this zone is considered to not comprise significant discontinuity, and grader 23 is distributed to this zone with classification 14.On the other hand, if if the variation in the zone of calculating according to formula (1) greater than described predetermined threshold, then group a is found, its group amount of exercise can (a) is the highest, indicate main discontinuity in group a, and grader 23 is distributed to this zone with classification a.According to this classification, grader 23 is selected in the optional regional code device 271 to 27n and the corresponding regional code device of this classification.
According to this specific embodiment, each in the optional regional code device 271 to 27n comprises a subset encoder 531 of one group of 13 encoder to 53k, and each is applicable to 13 pixel groups shown in Figure 6 in these 13 encoders.Distinguishingly, each is applicable on its relevant pixel group and carries out two-dimension discrete cosine transform in these 13 encoders.A plurality of encoders 531 to 53k of each optional regional code device 271 to 27n comprise the corresponding encoded device of the classification that is applicable to corresponding optional regional code device, also comprise the encoder that is applicable to the classification relevant with corresponding optional regional code device.For each classification, every other optional regional code device comprises that one group of identical encoder 531 is to 53k.Usually,,, cannot ignore if a different classes of probability is divided greater than the probability that divides into respective classes in a zone for a particular category, then this different classes of be the classification of being correlated with.In this particular example, found that the classification relevant with classification 1 is classification 7,9 and 13.The classification relevant with classification 7 is classification 1,4 and 13.The classification relevant with classification 11 is classification 11 and 13.The classification relevant with classification 13 is other all categories.The related category of other all categories is directly followed symmetry and is considered.
According to this embodiment of video encoder of the present invention utilize one be not with by grader 23 determine that the encoder of the corresponding encoder of classification is very lower than operating better probability corresponding to such other encoder, simultaneously corresponding to best very high this fact of probability of such other encoder running performance, the coding method that is transferred to Video Decoder like this selects amount of information very low, and can only represent with bit number seldom.It is to be carried out with a kind of known method based on the corresponding selection probability of each optional regional code device by moisture in the soil encoder 52 that this selection information is encoded.
Usually, after the uncertainty that in the classification of considering based on the predication error frame in corresponding predictive frame zone, is comprised, select relevant classification.It should be noted that when in the present embodiment, there is the probability that other related categories are provided in each related category during corresponding to the classification of an optional regional code device.For example, a specific classification can have the related category different with other any classifications.
Fig. 8 has shown first embodiment according to Video Decoder of the present invention.This Video Decoder is applicable in the mode identical with video encoder and produces a predictive frame P.In addition, this Video Decoder is from by extracting identical information the predictive frame P of video encoder be used for to predication error frame coding, and the information that will extract be used for to the information decoding that received be used to regenerate predication error frame E
*
Distinguishingly, the decoder of Fig. 8 comprises that a coding video frequency data that is used to receive from video encoder is the demodulation multiplexer (demultiplexer) of prediction error data and motion vector with being used for the data demultiplexing that is received.Piece 32 one of expression are used to receive the prediction error data of demultiplexing and are used for by the coded prediction error information regeneration predication error frame E that is received
*The predicated error decoder.33 the expression be used for by predictive frame P (x, y) and regeneration predication error frame E
*Calculate the frame of video I of regeneration
*Calculation element.The frame of video I of this regeneration
*Be kept in the frame memory 34, and and then export as decode video data by Video Decoder.35 expressions are used for being used as reference frame R and being made up the motion compensated predictor of predictive frame P based on what receive from motion vector video encoder and that be described in the estimation pixel displacement among the present frame I by being stored in frame memory 34.MC fallout predictor 35 use the pixel displacement of reference frame R and these estimations obtain predictive frame P (x, y).
Predicated error decoder 32 is the prediction error data by the video encoder transmission of received code not only, also receive from the predictive frame P of MC fallout predictor 35 and the motion vector that sends by video encoder, and based on predictive frame P with carry out the decoding of predicated error based on motion vector.This can be represented by the dotted line among Fig. 8.
Based on predictive frame P with based on motion vector, predicated error decoder 32 is in a kind of like this position, the silhouette edge edge that has utilized the significant error in predication error frame E roughly to concentrate the predictive frame be in the athletic performance above the fold, and judge that based on the characteristic that this obvious zone of having only predication error frame of video encoder is encoded which zone of predication error frame E transmitted by video encoder.Because the MC prediction 35 predictive frame P that produced corresponding in video encoder by the MC fallout predictor 7 predictive frame P that produces, can carry out predicated error decoding according to this Video Decoder of this embodiment of the present invention, and must not receive information about the position of the important area of predication error frame.Extraly or selectively, predicated error decoder 32 can be carried out adaptive decoding, and must not extract information about the predication error frame of each coding, because decoder can be to be used for predictive frame P the zone that will be encoded of the predication error frame identical mode of classifying is extracted this information from predictive frame P with video encoder.
The detailed maps of the coded prediction error device 32 of Fig. 9 displayed map 8.In Fig. 9, predictive frame P who is used for being provided by MC fallout predictor 35 of 41 expressions is divided into the device in a plurality of zones.Cutting apart of predictive frame to carry out with mode identical in video encoder.In other words, frame segmenting device 41 complete functions corresponding to the frame segmenting device 21 in the coded prediction error device of Fig. 1.
Discontinuity analyzer of reference marker 42 expression, its function be corresponding to the function of the discontinuity analyzer 22 that provides in video encoder, also is suitable for piece 42 about the corresponding explanation of piece 22, therefore just do not repeating here.Which zone that piece 44 expression among Fig. 9 is used for judging predication error frame is by video encoder encodes and be included in the prediction error data of compression.For this reason, decision maker 44 is carried out one and is similar to the judgement of the performed important area of piece among Fig. 2 25 and the operation of decision.If Video Decoder is designed to operate with first embodiment of above-mentioned video encoder, then the predictive frame P that is provided for piece 41
nEach location of pixels in each zone (x, y), piece 44 receives the level and vertical difference F by 42 outputs of discontinuity analyzer
V(x, y) and F
H(x, y), and by will be at all positions in each zone (x, the F that y) locates
V(x, y) and F
H(x, the degree of the discontinuity of being found is judged by absolute value phase Calais y) in each zone.The judgement of being implemented according to the piece among Fig. 2 25, if gained and greater than the corresponding predetermined threshold of predetermined threshold that is adopted with piece 25, piece 44 judges that the respective regions of predictive frame P comprises significant discontinuity.Otherwise piece 44 judges that this respective regions does not comprise significant discontinuity.
In addition, piece 44 is gone back receiver, video encoder institute transmitting moving vector.Based on this, piece 44 is judged the athletic performance amount that each is regional by the amplitude of the motion vector of calculating respective regions and with amplitude peak and with the piece 25 among Fig. 2 for the corresponding threshold of described second threshold value that corresponding operating adopted.The identical standard that piece 25 is adopted based on piece 25 is judged this predictive frame P
nA zone whether important, if promptly in this zone the degree of discontinuity and athletic performance amount all be found to be significantly, then piece 25 judges corresponding zone to be important.
Piece 44 is applied to a predetermined Cahn-Ingold-Prelog sequence rule each several part of the prediction error data that will be received and distributes to corresponding zone, and this rule is used for the predefined procedure rule of the important area of prediction of output error frame corresponding to piece in video encoder 25.Distinguishingly be, in this particular example, piece 44 is arranged as a combination in zone with the All Ranges of the cutting operation gained that the mode similar to 25 arrange regionals of piece in video encoder carried out piece 425, and this zone combination of lining by line scan is to seek important zone.Part corresponding to the prediction error data that each received in a zone of predication error frame is assigned to important area, and the order in the described area arrangement is corresponding to the order that receives appropriate section.An example of this operation is shown in Figure 10 a and 10b.The combination 105 in Figure 10 a piece 21 and zone that the performed cutting operations of piece 41 produce in Video Decoder in video encoder.This combination comprises four row 0-3 and four lines 0-3.Dash area is represented important area in this combination 105, and non-shaded portion is inessential zone.
Figure 10 has shown a prediction error data stream that receives from video encoder.This data flow is made of part 101 to 104, and each part comprises the prediction error data of the respective regions of predication error frame.In the present example, the operation of lining by line scan that piece 44 carries out in combination 105 has caused the order subsequently of important combination: (0,1); (1,2); (2,3); (3,2).This order is corresponding to the order of the each several part in prediction error data.Based on these corresponding order, piece 44 can be distributed to the part 101 to 104 of each reception of prediction error data its respective regions.T express time axle.
Serve as reasons each part of compressed prediction error data of piece 44 outputs of selector of being controlled by grader 43 of reference marker 45 expression, the characteristic classification that is used for judging according to grader is selected a suitable decoder.Thus, adaptive region decoder 48 can be according to predictive frame P
nEach regional classification carry out an operation that reverses mutually with the performed encoding operation of piece in video encoder 26.Each optional regional decoding device 471 to 47n of adaptive region decoder 48 is corresponding to one in the optional regional code device 271 to 27n in the video encoder among Fig. 9.If employing vector quantizer, it comprises one, and be adapted to will be by the code book of the area classification of each regional decoding device 471 to 47n decodings, and each regional decoding device 471 to 47n is adapted to receive an index (index) of giving its code book at its input, the used code book of this code book and corresponding regional code device is identical, and by reading vector corresponding to the code book of institute's reception hint.Certainly, except vector quantization or as selectable, this coded prediction error device and decoder can merge other adaptive prediction error coding and decoding scheme, and for example orthogonal transform (DCT, KLT) or the match tracing technology.
Figure 11 has shown the optional regional decoding device of of Video Decoder according to a second embodiment of the present invention.Video Decoder according to second embodiment is adapted to use with above-mentioned video encoder according to second embodiment.According to this Video Decoder of second embodiment with similar according to the Video Decoder of first embodiment, except operating discontinuity analyzer 22 and the grader 23 that is similar to respectively according to the video encoder of second embodiment according to discontinuity analyzer 42 among Fig. 9 of second embodiment and grader 43, like this, the description of these structures of determining and operation does not just remake repeat specification once more.In addition, the optional regional decoding device 471 to 47n according in the Video Decoder of second embodiment shown in Figure 11 is different from the optional regional decoding device among first embodiment.
As shown in figure 11, comprising one according to each the optional regional decoding device 471 to 47n in the Video Decoder of second embodiment, to be used for the compressed prediction error data from the video encoder of second embodiment that receive are carried out demultiplexing be that coded prediction error information and entropy coding method are selected information.Reference marker 62 expressions receive the entropy decoder that entropy coding method is selected information.Reference marker 631 to 63k represent wherein each all with encoder 531 to 53k in the optional regional code device of optional regional decoding device corresponding to each in corresponding a plurality of decoders.Reference marker 64 and 65 expressions are by the selector of the output control of entropy decoder 62, be used for providing the selected decoder of coded prediction error information, link to each other with the output of corresponding optional regional decoding device with the output that is used for selected decoder to a plurality of decoders 631 to 63k.
Can be advantageously used in various application fields according to the encoder that is used for video data of the present invention with according to of the present invention being used for to the method for video data encoding and decoding.One of them application is a system that is used for transmitting video data, this system comprises one according to encoder of the present invention, be used for the device of transfer encoding video data and one according to the decoder that is used to receive the video data of institute's transmitting video data and output decoder of the present invention.
Another favourable application is one and is used to preserve video data, this system comprises the encoder that is used for according to the present invention video data encoding, be used to preserve the device of coding video frequency data and be used for extracting coding video frequency data and being used to decode and exporting the described video data that extracts from described storage device.
Advantageously, the present invention also can merge use with the different coding mode of the zones of different of the frame that will be encoded.Other zones of this frame are through motion compensation encodings if intraframe codings are passed through in some zones of a frame, then the present invention can be used for these motion compensation encoding zones, and the All Ranges that term " motion compensation encoding of a frame of video " should not be construed as limiting to this frame all is the situation of MC coding.
Those skilled in the art should be clear, and the present invention can be the basis of the one or more coded systems in Video Codec, and wherein it can use with coded system of the prior art.When coding region and encoding scheme determine do not produce desirablely as a result the time, encoder can determine to use the mode of a prior art.For example, encoder can be overthrown the selection of coding region and the selection of encoding scheme, and this decision is transferred to this decoder.This may need to use the unwanted function of other mode, but it can judge whether the coding decision that takes place subsequently is obviously better.
Claims (35)
1. encoder that is used for video data is carried out motion compensation encoding comprises:
-be used to estimate to describe the device (8) of motion vector of the scene displacement of frame of video pixel;
-be used for predicting the device (7) of a video data frame (P) based on first video data frame (R) and described motion vector;
-be used for the device (1) that video data frame (P) based on described prediction and second video data frame (I) calculate predication error frame (E);
-be used for described predication error frame (E) coding, and obtain the device (2) of prediction error data;
-be used for described motion vector and described prediction error data are transferred to the device (4) of decoder;
-described being used for comprising described predication error frame (E) apparatus for encoding (2):
-be used for described predication error frame is divided into the device (24) in a plurality of zones; With
-being used for carrying out the decision maker of at least one following operation: judgement will be encoded and judge for each regional at least a coding method in the zone of described predication error frame (E) and be used to described predication error frame (E) is encoded; With
-be used for the regional code device (26) of described predication error frame (E) being encoded according at least one decision of described decision maker;
-wherein said decision maker is applicable to the characteristic of the described prediction video data frame of analysis (p) and carries out at least one decision based on described analysis.
2. according to the encoder of claim 1, it is characterized in that,
-described decision maker comprises and being used for for each regional analysis of described predication error frame device (22) in the discontinuity of the respective regions image brightness of prediction video data frame (P); With
-described decision maker (2) is applicable to based on the analysis of described discontinuity carries out described at least one decision.
3. according to the encoder of claim 2, it is characterized in that,
Described discontinuity analytical equipment (22) comprising:
-be used at the limited difference (F of each zone calculating of predicting video data frame (P) in vertical direction
H) linear filter and calculate in the horizontal direction limited difference (F
V) linear filter; With
-be used for device absolute value and that compare with first predetermined threshold in the output of two filters in each zone;
-described decision maker is applicable to that a zone judging predication error frame (E) is not encoded, if that the respective regions of described prediction video data frame (P) is calculated absolute value and do not exceed described first predetermined threshold.
4. according to the encoder of any one claim of front, it is characterized in that,
-described regional code device (26) comprise be used for to each be determined the zone that will be encoded respectively from a plurality of zones of different encoders (271 ..., 27n) the middle choice device (28,29) of selecting a regional code device.
5. according to the encoder of claim 4, it is characterized in that,
-described decision maker comprises the brightness discontinuity that is used for according to the respective regions of described prediction video data frame (P), a classification in one group of characteristic classification is distributed to the sorter (23) in each zone that will be encoded of described predication error frame (E);
-described choice device (28,29) is applicable to the classification of distributing according to each zone, to each zone that will be encoded select a regional code device (271 ..., 27n).
6. according to the encoder of claim 5, it is characterized in that,
Described sorter (23) is applicable to based on being used at the limited difference (F that calculates in vertical direction with each zone of the regional corresponding prediction video data frame (P) that will be encoded of described predication error frame (E)
H) and calculating limited difference (F in the horizontal direction
V) the output valve that provided of linear filter (22) classify.
7. according to the encoder of claim 5, it is characterized in that,
Described sorter (23) is applicable to based on the respective pixel brightness discontinuity among each the regional different pixels group who concentrates on this prediction video data frame (P) classifies, and each zone of this this prediction video data frame (P) is corresponding to a zone that will be encoded of described predication error frame (E).
8. according to the encoder of claim 6 or 7, it is characterized in that,
-described category does not comprise at least one in the following classification: no discontinuity; The horizontal direction discontinuity; The vertical direction discontinuity; Diagonal angle, left side oblique line directions discontinuity; Right diagonal angle oblique line directions discontinuity; The non-directional discontinuity.
9. according to the encoder of any claim in front, it is characterized in that,
-described regional code device (26) is applicable to by selected orthogonal transform or vector quantization or match tracing scheme and is encoded in each zone that will be encoded of described predication error frame (E).
10. according to the encoder of any claim in front, it is characterized in that,
-each regional code device (271 ... 27n) be applicable to and comprise a plurality of coding methods (531 to 53k);
-offer each regional code device (271 ... 27n) be used for according to the coding method choice device (51,54,56) of the standard relevant from a coding method of corresponding a plurality of coding methods (531 to 53k) selection with the performance of coding method;
11. the encoder according to claim 11 is characterized in that,
Described a plurality of coding method (531 to 53k) for the zones of different encoder (271 ... be different 27n);
12. according to any one encoder in the claim 1 to 11, it is characterized in that,
-be used for predicated error decoder (3) that described predication error frame (E) is decoded; With
-be used for calculating the device (5) of first video data frame based on described prediction video data frame (P) and described decoding predication error frame (E*).
13. the encoder according to any claim in front is characterized in that,
-described decision maker (25) is included as the device of the direction of each the regional calculating amplitude of described prediction video data frame (P) and/or respective motion vectors;
-described decision maker (25) is applicable to based on the direction of amplitude of being calculated and/or described motion vector carries out described judgement.
14. the encoder according to claim 13 is characterized in that,
-described decision maker is applicable to that a zone judging predication error frame (E) is not encoded, if the motion vector amplitude of the described calculating corresponding with this zone does not surpass second predetermined threshold.
15. be used for decoder that the motion compensation encoding video data is decoded, comprise:
-be used to preserve the device (34) of a video data frame (R);
-be used for based on described video data frame (R) and predict the device (35) of a video data frame (P) based on the motion vector data that is received;
-be used for the decoding of the prediction error data that received and obtain a predication error frame (E
*) device (32); With
-be used for predication error frame (E based on described video data frame (P) and described decoding
*) calculate and export the video data frame (I of a renewal
*), and with the video data frame (I of this renewal
*) be kept at the device (33) in the described storage device;
-be used for the described device (32) that the prediction error data of described reception is decoded being comprised:
-be used for carrying out the decision maker of following at least one operation: judge encoding and being included in the zone of the prediction error data that received and each has been encoded and is included at least one decoding scheme of regional determination in the prediction error data that is received of described predication error frame (E); With
-be used for the prediction error data of described reception being decoded to obtain the regional decoding device (48) of predication error frame according at least a decision of described decision maker;
-wherein said decision maker is applicable to the characteristic of the described prediction video data frame of analysis (p) and carries out at least one decision based on described analysis.
16. the decoder according to claim 15 is characterized in that,
Described decision maker comprises
-be used for described prediction video data frame (P) is divided into the device (41) of a plurality of presumptive areas
-be used for device to the discontinuity of each regional analysis image brightness of described prediction video data frame.
-described decision maker is applicable to based on respective regions in described prediction video data frame (P) described predication error frame (E
*) described at least one the decision of execution.
17. the decoder according to claim 16 is characterized in that,
Described discontinuity analytical equipment comprises:
-be used at the limited difference (F of each zone calculating of predicting video data frame (P) in vertical direction
H) linear filter (42) and calculate in the horizontal direction limited difference (F
V) linear filter (42); With
-be used for device absolute value and that compare with the 3rd predetermined threshold in the output of two filters in each zone;
-described decision maker (44) is applicable to not encode in a zone judging predication error frame and is not comprised in the prediction error data of described reception, if that the respective regions of described prediction video data frame is calculated absolute value and do not exceed described the 3rd predetermined threshold.
18. according to any one decoder in the claim 15 to 17, it is characterized in that,
-described regional decoding device (48) comprises and being used for described predication error frame (E
*) each encoded and be included in the prediction error data that is received the zone respectively from a plurality of zones of different decoders (471 ..., 47n) the middle choice device (45,46) of selecting a regional decoding device.
19. the decoder according to claim 18 is characterized in that,
-described decision maker comprises the brightness discontinuity that is used for according to the respective regions of described prediction video data frame (P), a classification in one group of characteristic classification is distributed to the sorter (43) in each zone that will be encoded of described predication error frame (E*);
-described choice device (45,46) is applicable to the classification of distributing according to each zone, and a regional code device is selected in each zone that will be encoded of described predication error frame (E*).
20. the decoder according to claim 19 is characterized in that,
Described sorter (43) be applicable to based on be used for described predication error frame (E
*) encode and be included in the limited difference (F that calculates in vertical direction in each zone of the regional corresponding prediction video data frame (P) in the prediction error data of described reception
H) and calculating limited difference (F in the horizontal direction
V) the output valve that provided of linear filter (42) classify.
21. the decoder according to claim 19 is characterized in that,
Described sorter (43) is applicable to based on the respective pixel brightness discontinuity among each the regional different pixels group who concentrates on this prediction video data frame (P) classifies, each zone and described predication error frame (E of this prediction video data frame (P)
*) encode and be included in regional corresponding in the prediction error data of described reception.
22. the decoder according to claim 20 or 21 is characterized in that,
-described category does not comprise at least one in the following classification: no discontinuity; The horizontal direction discontinuity; The vertical direction discontinuity; Diagonal angle, left side oblique line directions discontinuity; Right diagonal angle oblique line directions discontinuity; The non-directional discontinuity.
23. any one encoder according to claim 15 to 22 is characterized in that,
-described regional decoding device (48) is applicable to by selected orthogonal transform or vector quantization or match tracing scheme to come described predication error frame (E
*) the error prediction data that received with each decoded zone decode.
24. any one decoder according to claim 18 to 22 is characterized in that,
-each regional decoding device (471 ... 47n) be applicable to and comprise a plurality of coding/decoding methods (631 to 63k);
-be used for selecting information according to the coding method that is received, from corresponding a plurality of coding/decoding methods (631 to 63k), select the coding/decoding method choice device (62,64,65) of a coding/decoding method;
25. the decoder according to claim 24 is characterized in that,
Described a plurality of coding/decoding method (631 to 63k) for the zones of different decoder (471 ... be different 47n);
26. according to any one decoder in the claim 15 to 25, it is characterized in that,
-described decision maker is included as the device (44) of the direction of each the regional calculating amplitude of described prediction video data frame (P) and/or corresponding motion vector;
-described decision maker is applicable to based on the direction of amplitude of being calculated and/or described motion vector and carries out described judgement.
27. the encoder according to claim 26 is characterized in that,
-described decision maker is applicable to judges predication error frame (E
*) a zone do not encode and be not comprised in the prediction error data that is received, if the motion vector amplitude of the described calculating corresponding with this zone surpasses the 4th predetermined threshold.
28. according to any one decoder in the claim 15 to 27, it is characterized in that,
-described the device (32) that is used for that the prediction error data of described reception is decoded is applicable to according to predetermined allocation of space rule ordering ground institute's receiving unit of described prediction error data is distributed to by described decision maker (41,42,44) each zone of being judged;
-each part is corresponding to a coding region of described predication error frame.
29. the decoder according to claim 28 is characterized in that,
-described judgement allocation rule has defined distributes to described each zone with the described order receiving unit of described prediction error data with they are presented when the All Ranges of the described predication error frame of scanning order and/or with the order that they are judged by described decision maker.
30. be used for the system of transmitting video data, comprise according to any one encoder in the claim 1 to 14 and comprising according to any one decoder in the claim 15 to 29.
31. be used to preserve the system of video data, comprise according to any one encoder in the claim 1 to 14, be used to preserve coding video data device and according to any one decoder in the claim 15 to 29.
32. a method that is used for video data is carried out motion compensation encoding comprises:
The motion vector of the scene displacement of frame of video pixel is described in-estimation;
-predict a video data frame (P) based on first video data frame (R) and described motion vector;
-calculate predication error frame (E) based on the video data frame (P) and second video data frame (I) of described prediction;
-to described predication error frame (E) coding, and obtain prediction error data;
-described motion vector and described prediction error data are transferred to decoder;
-described being used for comprises the step of described predication error frame (E) coding:
-described predication error frame is divided into a plurality of zones; With
-carry out at least one following operation: the zone of judgement in described predication error frame (E) will be encoded and judge for each regional at least a coding method and be used to described predication error frame (E) is encoded; With
-according at least one decision of described determination step described predication error frame (E) is encoded;
-wherein said at least one decision is based on the characteristic of analyzing described prediction video data frame (p).
33. be used for method that the motion compensation encoding video data is decoded, comprise:
-preservation video data frame (R);
-predict a video data frame (P) based on described video data frame (R) with based on the motion vector data that is received;
-to the decoding of the prediction error data that received with obtain a predication error frame (E
*); With
-based on the predication error frame (E of described video data frame (P) and described decoding
*) calculate and export the video data frame (I of a renewal
*), and with the video data frame (I of this renewal
*) be kept in the described storage device;
-step of the prediction error data of described reception decoding is comprised:
-carry out following at least one operation: judge encoding and being included in the zone in the prediction error data that is received and each has been encoded and be included at least one decoding scheme of regional determination in the prediction error data that is received of described predication error frame (E); With
-according at least a decision of described decision maker the prediction error data of described reception is decoded to obtain predication error frame;
-wherein said judgement is based on the characteristic of analyzing described prediction video data frame (p).
34. the method for transmitting video data, comprise according to the method for claim 32 to described coding video data, via the video data of transmission medium transfer encoding with according to the method for claim 33 video data that is transmitted is decoded.
35. preserve the method for video data, comprise according to the method for claim 32 to described coding video data, the video data of coding is kept in the storage device, extracts the video data preserved and the video data that is extracted is decoded according to the method for claim 33 by described storage device.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB96180260XA CN1134993C (en) | 1996-04-18 | 1996-04-18 | Vedio data encoder and decoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB96180260XA CN1134993C (en) | 1996-04-18 | 1996-04-18 | Vedio data encoder and decoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1217125A true CN1217125A (en) | 1999-05-19 |
| CN1134993C CN1134993C (en) | 2004-01-14 |
Family
ID=5127772
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB96180260XA Expired - Lifetime CN1134993C (en) | 1996-04-18 | 1996-04-18 | Vedio data encoder and decoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1134993C (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100380984C (en) * | 2002-06-11 | 2008-04-09 | 诺基亚有限公司 | Spatial prediction based intra coding |
| CN101065953B (en) * | 2004-09-15 | 2012-04-18 | 雷斯昂公司 | Flir-to-missile boresight correlation and non-uniformity compensation of the missile seeker |
-
1996
- 1996-04-18 CN CNB96180260XA patent/CN1134993C/en not_active Expired - Lifetime
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100380984C (en) * | 2002-06-11 | 2008-04-09 | 诺基亚有限公司 | Spatial prediction based intra coding |
| CN101065953B (en) * | 2004-09-15 | 2012-04-18 | 雷斯昂公司 | Flir-to-missile boresight correlation and non-uniformity compensation of the missile seeker |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1134993C (en) | 2004-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR100430328B1 (en) | Video data encoders and decoders | |
| KR20220127323A (en) | Tree soup node size per slice | |
| CN1245029C (en) | Motion vector extrapolation for transcoding video sequences | |
| CN1767655A (en) | Multi view point video image parallax difference estimating method | |
| CN110139109B (en) | Image coding method and corresponding terminal | |
| CN1215439C (en) | Apparatus and method for performing scalable hierarchical motion estimation | |
| CN1130919C (en) | Apparatus for encoding contour of regions contained in video signal | |
| US6895051B2 (en) | Video data encoder and decoder | |
| US20060188014A1 (en) | Video coding and adaptation by semantics-driven resolution control for transport and storage | |
| CN1906948A (en) | Image coding apparatus and method for predicting motion using rotation matching | |
| KR20110028563A (en) | Video compression method | |
| CN1310913A (en) | Image compression | |
| CN1166211C (en) | Method and apparatus for encoding motion vector based on number of vilid reference motion vectors | |
| CN1292594C (en) | Coding and decoding method and apparatus using plural scanning patterns | |
| CN1956544A (en) | Methods and systems for video data processing employing continuous/interlaced region predictions | |
| CN1208313A (en) | Method and apparatus for encoding motion vector | |
| CN1774933A (en) | Method and apparatus for determining search range for adaptive motion vector for use in video encoder | |
| CN1232125C (en) | Method for motion estimation (me) through discrete cosine transform (dct) and an apparatus therefor | |
| CN1926879A (en) | A video signal encoder, a video signal processor, a video signal distribution system and methods of operation therefor | |
| CN1662066A (en) | Method for selecting predicting mode within frame | |
| CN1604650A (en) | Method for hierarchical motion estimation | |
| CN1159922C (en) | Method for encoding binary shape signal | |
| EP3469792A1 (en) | Devices and methods for video coding using segmentation based partitioning of video coding blocks | |
| CN1151678C (en) | Method and apparatus for encoding contour image of object in video signal | |
| CN1134993C (en) | Vedio data encoder and decoder |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: NOKIA SIEMENS COMMUNICATIONS INC. Free format text: FORMER OWNER: NOKIA OY Effective date: 20080418 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| C56 | Change in the name or address of the patentee |
Owner name: NOKIA OY Free format text: FORMER NAME OR ADDRESS: NOKIA MOBILE PHONES LTD. |
|
| CP03 | Change of name, title or address |
Address after: Espoo, Finland Patentee after: NOKIA Corp. Address before: Finland Thessaloniki Patentee before: Nokia Mobile Phones Ltd. |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20080418 Address after: Espoo, Finland Patentee after: NOKIA SIEMENS NETWORKS OY Address before: Espoo, Finland Patentee before: NOKIA Corp. |
|
| C56 | Change in the name or address of the patentee | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Espoo, Finland Patentee after: NOKIA SOLUTIONS AND NETWORKS OY Address before: Espoo, Finland Patentee before: Nokia Siemens Networks OY |
|
| CX01 | Expiry of patent term |
Granted publication date: 20040114 |
|
| EXPY | Termination of patent right or utility model |