CN106851273A

CN106851273A - A kind of motion-vector coding method and device

Info

Publication number: CN106851273A
Application number: CN201710073895.5A
Authority: CN
Inventors: 朱洪波
Original assignee: Beijing QIYI Century Science and Technology Co Ltd
Current assignee: Beijing QIYI Century Science and Technology Co Ltd
Priority date: 2017-02-10
Filing date: 2017-02-10
Publication date: 2017-06-13
Anticipated expiration: 2037-02-10
Also published as: CN106851273B

Abstract

The present invention proposes a kind of motion-vector coding method, including：The predicted motion vector predicted composition motion vector set for obtaining setting quantity is searched from the motion vector of the reference prediction encoding block of image block to be encoded；Judge in predicted motion vector set, if exist and image block motion vector identical predicted motion vector to be encoded；If there is with image block motion vector identical predicted motion vector to be encoded, then encode and indexed with the sequence number of image block motion vector identical predicted motion vector to be encoded；If there is no with image block motion vector identical predicted motion vector to be encoded, then the predicted motion vector in predicted motion vector set, treating coded picture block motion vector carries out differential coding treatment.The predicted motion vector that the cataloged procedure of motion-vector coding method proposed by the present invention is applied to is more, considers that different situations carry out corresponding coded treatment comprehensively, and code efficiency is higher.

Description

A kind of motion-vector coding method and device

Technical field

The present invention relates to encoding video pictures field, more particularly to a kind of motion-vector coding method and device.

Background technology

Block base motion vector coding techniques, be block base hybrid discrete cosine conversion video encoding standard core technology it One.

Existing piece of base motion-vector coding method, such as it is traditional H.263, H.264 in coding method, first from spatial domain or Time domain finds the intermediate value of the encoded encoding block motion vector of the multiple adjacent with encoding block, used as encoding block motion vector Prediction；Then the prediction of encoding block motion vector is subtracted with encoding block motion vector, the difference of encoding block motion vector is obtained, it is right The difference of encoding block motion vector is encoded.In above-mentioned existing coding method, only application is adjacent with encoding block The intermediate value of multiple encoded encoding block motion vectors, as the prediction of encoding block motion vector, to encoding block motion vector Prediction it is less, and for any one encoding block motion vector, carried out using the method for encoding motion vector difference Coding, coding method is simple, and code efficiency is relatively low.

The content of the invention

Defect and deficiency based on above-mentioned prior art, the present invention propose a kind of motion-vector coding method and device, Using multiple predicted motion vectors when being encoded to motion vector, consider that different situations carry out corresponding coded treatment comprehensively, Code efficiency is higher.

A kind of motion-vector coding method, including：

From the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting quantity, described to treat The predicted motion vector of coded picture block motion vector, predicted composition motion vector set；

By the image block motion vector to be encoded respectively with the predicted motion vector set in predicted motion vector Contrasted, judged in the predicted motion vector set, if existed identical with the image block motion vector to be encoded Predicted motion vector；

If in the predicted motion vector set, existing and being predicted with the image block motion vector identical to be encoded Motion vector, then encode and indexed with the sequence number of the image block motion vector identical predicted motion vector to be encoded, as institute State the coding of image block motion vector to be encoded；

If in the predicted motion vector set, in the absence of pre- with the image block motion vector identical to be encoded Motion vector is surveyed, then the predicted motion vector in the predicted motion vector set, the image block to be encoded is moved Vector carries out differential coding treatment, obtains the coding of the image block motion vector to be encoded.

Preferably, described from the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting number Amount, the predicted motion vector of the image block motion vector to be encoded, predicted composition motion vector set, including：

According to efficient video coding HEVC coding methods, from the motion of the spatial domain adjacent encoder block of image block to be encoded to Amount, and in time domain reference image in the motion vector of the encoding block of same position, lookup obtains setting quantity, and correspondence is waited to compile The predicted motion vector of code image block motion vector, predicted composition motion vector set；

In rejecting the predicted motion vector set, in setting the predicted motion vector of quantity, the predicted motion for repeating Vector.

Preferably, from the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting quantity , the predicted motion vector of the image block motion vector to be encoded after predicted composition motion vector set, is treated by described Coded picture block motion vector is contrasted with the predicted motion vector in the predicted motion vector set respectively, is judged in institute In stating predicted motion vector set, if exist with the image block motion vector identical predicted motion vector to be encoded it Before, the method also includes：

It is according to the reference key of the image block motion vector to be encoded and pre- in the predicted motion vector set The reference key of motion vector is surveyed, treatment is zoomed in and out to the predicted motion vector in the predicted motion vector set respectively, Make the predicted motion vector in the predicted motion vector set and the skew represented by the image block motion vector to be encoded Amount is identical.

Preferably, the sequence number rope encoded with the image block motion vector identical predicted motion vector to be encoded Draw, including：

Generation is characterized in the predicted motion vector set, is existed and the image block motion vector identical to be encoded The identification information of reference motion vectors；

Encode the identification information, and coding and the image block motion vector identical predicted motion vector to be encoded Sequence number index.

Preferably, the predicted motion vector in the predicted motion vector set, to the image to be encoded Block motion vector carries out differential coding treatment, obtains the coding of the image block motion vector to be encoded, including：

The X-component of the predicted motion vector in the predicted motion vector set, is calculated the figure to be encoded As the X-component difference of block motion vector；

The Y-component of the predicted motion vector in the predicted motion vector set, is calculated the figure to be encoded As the Y-component difference of block motion vector；

To the X-component difference, the sequence number index of the correspondence X-component difference, the Y-component difference and the correspondence Y The sequence number index of component difference, carries out coded treatment, obtains the coding of the image block motion vector to be encoded.

Preferably, the X-component of the predicted motion vector in the predicted motion vector set, is calculated institute The X-component difference of image block motion vector to be encoded is stated, including：

The X-component of the predicted motion vector obtained in the predicted motion vector set is extracted respectively；

Generation First ray, comprising first predicted motion in the predicted motion vector set in the First ray The X-component of vector；

The X-component of the remaining predicted motion vector in the predicted motion vector set is traveled through, a prediction is often traversed During the X-component of motion vector, the X of the predicted motion vector in the X-component of the predicted motion vector and the First ray is calculated The distance of component；If the X-component of the predicted motion vector in the X-component of the predicted motion vector and the First ray away from From both greater than 1, then the X-component of the predicted motion vector is added into the First ray；

From the First ray, lookup obtains the absolute difference with the X-component of the image block motion vector to be encoded The X-component of minimum predicted motion vector, record is minimum with the absolute difference of the X-component of the image block motion vector to be encoded Predicted motion vector X-component in the First ray sequence number index；

According to the X of the minimum predicted motion vector of the absolute difference with the X-component of the image block motion vector to be encoded Component, and the image block motion vector to be encoded X-component, be calculated the X of the image block motion vector to be encoded Component difference.

Preferably, the Y-component of the predicted motion vector in the predicted motion vector set, is calculated institute The Y-component difference of image block motion vector to be encoded is stated, including：

The Y-component of the predicted motion vector obtained in the predicted motion vector set is extracted respectively；

By the Y-component of the predicted motion vector in the predicted motion vector set, the second sequence and the 3rd sequence are divided into Row；Wherein, the X-component of the corresponding predicted motion vector of Y-component of the predicted motion vector in second sequence, with described In one sequence, the X of the predicted motion vector minimum with the absolute difference of the X-component of the image block motion vector to be encoded divides Amount is identical；The X-component of the corresponding predicted motion vector of Y-component of the predicted motion vector in the 3rd sequence, with described In one sequence, the X of the predicted motion vector minimum with the absolute difference of the X-component of the image block motion vector to be encoded divides Amount is different；

The 4th sequence is generated, comprising first Y of predicted motion vector in second sequence in the 4th sequence Component；

The Y-component of remaining predicted motion vector in second sequence is traveled through, a predicted motion vector is often traversed Y-component when, calculate predicted motion vector in the Y-component of the predicted motion vector and the 4th sequence Y-component away from From；If the distance of the Y-component of the predicted motion vector in the Y-component of the predicted motion vector and the 4th sequence is both greater than 1, then the Y-component of the predicted motion vector is added into the 4th sequence；

The 5th sequence is generated, comprising first Y of predicted motion vector in the 3rd sequence in the 5th sequence Component；

The Y-component of remaining predicted motion vector in the 3rd sequence is traveled through, a predicted motion vector is often traversed Y-component when, calculate predicted motion vector in the Y-component of the predicted motion vector and the 5th sequence Y-component away from From；If the distance of the Y-component of the predicted motion vector in the Y-component of the predicted motion vector and the 5th sequence is both greater than 1, then the Y-component of the predicted motion vector is added into the 5th sequence；

From the 4th sequence and the 5th sequence, lookup obtains the Y with the image block motion vector to be encoded The Y-component of the minimum predicted motion vector of the absolute difference of component, records the Y-component with the image block motion vector to be encoded The minimum predicted motion vector of absolute difference Y-component where sequence, and with the image block motion vector to be encoded Y-component the minimum predicted motion vector of absolute difference sequence of the Y-component where it in sequence index；

According to the Y of the minimum predicted motion vector of the absolute difference with the Y-component of the image block motion vector to be encoded Component, and the image block motion vector to be encoded Y-component, be calculated the Y of the image block motion vector to be encoded Component difference.

Preferably, it is described to the X-component difference, the sequence number index of the correspondence X-component difference, the Y-component difference And the sequence number index of the correspondence Y-component difference, coded treatment is carried out, obtain the volume of the image block motion vector to be encoded Code, including：

It is absolute value sum, the absolute value of the X-component difference to the X-component difference and the Y-component difference, described The minimum prediction of the absolute difference of the X-component of X-component difference, the Y-component difference and the image block motion vector to be encoded Sequence number of the X-component of motion vector in the First ray indexes, characterizes and divide with the Y of the image block motion vector to be encoded The information of the sequence where the Y-component of the minimum predicted motion vector of the absolute difference of amount and the image block motion to be encoded Sequence index difference in sequence of the Y-component of the minimum predicted motion vector of the absolute difference of the Y-component of vector where it Carry out coded treatment.

A kind of motion vector code device, including：

Prediction processing unit, for from the motion vector of the reference prediction encoding block of image block to be encoded, lookup to be obtained Setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition motion vector set；

Judging unit, for by the image block motion vector to be encoded respectively with the predicted motion vector set in Predicted motion vector is contrasted, and is judged in the predicted motion vector set, if existed and the image block to be encoded Motion vector identical predicted motion vector；

First coding unit, judges in the predicted motion vector set for the judging unit, exist with it is described During image block motion vector identical predicted motion vector to be encoded, encode and the image block motion vector identical to be encoded The sequence number index of predicted motion vector, as the coding of the image block motion vector to be encoded；

Second coding unit, judges in the predicted motion vector set for the judging unit, in the absence of with institute When stating image block motion vector identical predicted motion vector to be encoded, the prediction fortune in the predicted motion vector set Moving vector, differential coding treatment is carried out to the image block motion vector to be encoded, obtain the image block to be encoded move to The coding of amount.

Preferably, the prediction processing unit is looked into from the motion vector of the reference prediction encoding block of image block to be encoded Find setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition motion vector set When, specifically for：

According to efficient video coding HEVC coding methods, from the motion of the spatial domain adjacent encoder block of image block to be encoded to Amount, and in time domain reference image in the motion vector of the encoding block of same position, lookup obtains setting quantity, and correspondence is waited to compile The predicted motion vector of code image block motion vector, predicted composition motion vector set；Reject the predicted motion vector set In, in the predicted motion vector of setting quantity, the predicted motion vector for repeating.

Preferably, described device also includes：

Scaling processing unit, for the reference key according to the image block motion vector to be encoded, and the prediction The reference key of the predicted motion vector in motion vector set, respectively to the predicted motion in the predicted motion vector set Vector zooms in and out treatment, the predicted motion vector in the predicted motion vector set is moved with the image block to be encoded Side-play amount represented by vector is identical.

Preferably, first coding unit coding and the image block motion vector identical predicted motion to be encoded to When the sequence number of amount is indexed, specifically for：

Generation is characterized in the predicted motion vector set, is existed and the image block motion vector identical to be encoded The identification information of reference motion vectors；Encode the identification information, and coding and the image block motion vector phase to be encoded The sequence number index of same predicted motion vector.

Preferably, second coding unit, including：

First component calculation unit, for the X-component of the predicted motion vector in the predicted motion vector set, It is calculated the X-component difference of the image block motion vector to be encoded；

Second component computing unit, for the Y-component of the predicted motion vector in the predicted motion vector set, It is calculated the Y-component difference of the image block motion vector to be encoded；

Differential coding unit, for dividing the X-component difference, the sequence number index of the correspondence X-component difference, the Y The sequence number index of amount difference and the correspondence Y-component difference, carries out coded treatment, obtains the image block motion vector to be encoded Coding.

Preferably, predicted motion vector of first component calculation unit in the predicted motion vector set X-component, when being calculated the X-component difference of the image block motion vector to be encoded, specifically for：

Preferably, predicted motion vector of the second component computing unit in the predicted motion vector set Y-component, when being calculated the Y-component difference of the image block motion vector to be encoded, specifically for：

Preferably, the differential coding unit is to the X-component difference, the sequence number index of the correspondence X-component difference, institute The sequence number index of Y-component difference and the correspondence Y-component difference is stated, coded treatment is carried out, the image block fortune to be encoded is obtained During the coding of moving vector, specifically for：

Motion-vector coding method proposed by the present invention, the present encoding block motion vector of acquisition setting quantity is pre- first Survey motion vector, and to set quantity predicted motion vector zoom in and out treatment；Then, contrast is judged in predicted motion vector In whether there is with present encoding block motion vector identical predicted motion vector, if it has, then direct coding predicted motion The sequence number index of vector；If it is not, according to the predicted motion vector of setting quantity, being carried out to present encoding block motion vector Differential coding treatment.In above-mentioned coding process, multiple predicted motion vectors are applied to, and present encoding block is moved When vector is encoded, first in view of whether having in multiple predicted motion vectors and present encoding block motion vector identical Predicted motion vector, if it has, then directly using the coding of the predicted motion vector as present encoding block motion vector coding, Only encode the sequence number index of the predicted motion vector；If it is not, using above-mentioned multiple predicted motion vectors, to current Encoding block motion vector is encoded.The predicted motion vector that its cataloged procedure is applied to is more, and different situations are considered comprehensively Corresponding coded treatment is carried out, code efficiency is higher.

Brief description of the drawings

In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this Inventive embodiment, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis The accompanying drawing of offer obtains other accompanying drawings.

Fig. 1 is a kind of schematic flow sheet of motion-vector coding method provided in an embodiment of the present invention；

Fig. 2 is a kind of schematic diagram of predicted motion vector for obtaining motion vector provided in an embodiment of the present invention；

Fig. 3 is the schematic flow sheet of another motion-vector coding method provided in an embodiment of the present invention；

Fig. 4 is the schematic flow sheet of another motion-vector coding method provided in an embodiment of the present invention；

Fig. 5 is the schematic flow sheet of another motion-vector coding method provided in an embodiment of the present invention；

Fig. 6 is a kind of structural representation of motion vector code device provided in an embodiment of the present invention；

Fig. 7 is the structural representation of another motion vector code device provided in an embodiment of the present invention；

Fig. 8 is the structural representation of another motion vector code device provided in an embodiment of the present invention.

Specific embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.

Shown in Figure 1 the embodiment of the invention discloses a kind of motion-vector coding method, the method includes：

S101, from the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting quantity, institute State the predicted motion vector of image block motion vector to be encoded, predicted composition motion vector set；

Specifically, when the motion vector for treating coded picture block is encoded, image block fortune to be encoded is generally found first The predicted motion vector of row vector, the motion vector for treating coded picture block using predicted motion vector is encoded.In this hair In bright embodiment, from the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting quantity, waits to compile The predicted motion vector of code image block motion vector.Using multiple predicted motion vectors participate in treating the motion of coded picture block to The coding of amount, the code efficiency of coded picture block motion vector is treated to improve.Further, search and obtain setting quantity After predicted motion vector, predicted composition motion vector set.

S102, by the image block motion vector to be encoded respectively with the predicted motion vector set in predicted motion Vector is contrasted, and is judged in the predicted motion vector set, if existed and the image block motion vector to be encoded Identical predicted motion vector；

Specifically, by the motion vector of image block to be encoded, and the motion vector of image block to be encoded reference key, It is right to be carried out with the reference key of each predicted motion vector in predicted motion vector set, and each predicted motion vector Than.If certain predicted motion vector is identical with the motion vector of image block to be encoded, and the predicted motion vector reference Index is also identical with the reference key of image block motion vector to be encoded, then think the predicted motion vector with image to be encoded The motion vector of block is identical.

If in the predicted motion vector set, existing and being predicted with the image block motion vector identical to be encoded Motion vector, then perform the sequence of step S103, coding and the image block motion vector identical predicted motion vector to be encoded Number index, as the coding of the image block motion vector to be encoded；

If specifically, in predicted motion vector set, existed pre- with the motion vector identical of image block to be encoded Motion vector is surveyed, in view of predicted motion vector is the motion vector for having completed coding, its coding is accurate.Therefore, at this In inventive embodiments, when it is determined that there is the motion vector identical prediction fortune with image block to be encoded in predicted motion vector set During moving vector, in order to save operand, directly by the coding with image block motion vector identical predicted motion vector to be encoded As the coding of the motion vector of image block to be encoded, the sequence number index of the predicted motion vector is only encoded.Treating volume When the motion vector of code image block is decoded, the coding that only decoding is indexed to the sequence number of the predicted motion vector, further, The coding for finding the predicted motion vector is indexed by sequence number, is the coding that have found image block motion vector to be encoded, solution Code the predicted motion vector coding, also just complete the decoding for treating coded picture block motion vector.

If in the predicted motion vector set, in the absence of pre- with the image block motion vector identical to be encoded Motion vector is surveyed, then performs step S104, the predicted motion vector in the predicted motion vector set, wait to compile to described Code image block motion vector carries out differential coding treatment, obtains the coding of the image block motion vector to be encoded.

If specifically, in predicted motion vector set, in the absence of pre- with image block motion vector identical to be encoded Survey motion vector, then according to the prediction for treating coded picture block motion vector, treat coded picture block motion vector poor Coded treatment, obtains the coding of image block motion vector to be encoded.It should be noted that in embodiments of the present invention, being beneficial to The motion vector that multiple predicted motion vectors of the motion vector of image block to be encoded treat coded picture block carries out differential coding Treatment, relative to the processing method of prior art, its encoding precision is higher.

Optionally, in another embodiment of the present invention, the reference prediction encoding block from image block to be encoded In motion vector, lookup obtains setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition Motion vector set, including：

Specifically, the spatial domain adjacent encoder block of image block to be encoded, refer to it is spatially adjacent with image block to be encoded, Belong to the encoded image block of same two field picture；The coding of same position in the time domain reference image of image block to be encoded Block, refers to encoded, with image block to be encoded not in same two field picture, but in the image of identical image position Block.The embodiment of the present invention uses HEVC coding methods, from the motion vector of the spatial domain adjacent encoder block of image block to be encoded, and In time domain reference image in the motion vector of the encoding block of same position, lookup obtains setting the predicted motion vector of quantity.

For example, as illustrated in fig. 2, it is assumed that X is image block to be encoded, Y is the same location drawing picture in the reference picture of X in time domain Block.Using the method for the acquisition spatial and temporal prediction motion vector of HEVC, first from the motion of the upper left adjacent image block of X to Predicted motion vector is searched in amount in order, such as predicted motion vector is obtained according to the order of A0, B0, B2, A1, B1, if certain The motion vector of individual position does not exist, then skip it.Then, if being obtained from the same position encoding block of time domain reference image Predicted motion vector is done, as shown at right two positions of stain, C0 and C1.

Specifically, after the predicted motion vector for obtaining setting quantity is searched, contrast confirms the prediction fortune in setting quantity In moving vector, if having the predicted motion vector of repetition, that is, it is confirmed whether have motion vector identical, and reference key is also identical Predicted motion vector.If it has, then rejecting the predicted motion vector of repetition, make the predicted motion in predicted motion vector set Vector is differed.

Optionally, in another embodiment of the present invention, it is shown in Figure 3, performing step S301, from figure to be encoded As in the motion vector of the reference prediction encoding block of block, lookup obtains setting quantity, the image block motion vector to be encoded Predicted motion vector, after predicted composition motion vector set, perform step S303, will the image block motion to be encoded Vector is contrasted with the predicted motion vector in the predicted motion vector set respectively, is judged in the predicted motion vector In set, if exist with before the image block motion vector identical predicted motion vector to be encoded, the method also includes：

S302, the reference key according to the image block motion vector to be encoded, and the predicted motion vector set In predicted motion vector reference key, the predicted motion vector in the predicted motion vector set is zoomed in and out respectively Treatment, make predicted motion vector in the predicted motion vector set with represented by the image block motion vector to be encoded Side-play amount is identical.

Specifically, the reference key of motion vector, represents the side-play amount of the motion vector.If two ginsengs of motion vector Examine index different, it is impossible to think that the two motion vectors are identical.Therefore, in order to guarantee to be looked in predicted motion vector set To with image motion vector identical predicted motion vector to be encoded, the embodiment of the present invention to predicted motion vector set in it is pre- Survey motion vector zoom in and out treatment, make the side-play amount of predicted motion vector, i.e. reference key, with image block to be encoded move to The reference key of amount is identical.Accordingly, while the reference key of predicted motion vector is zoomed in and out, by predicted motion to The coordinate of amount scales identical multiple.It should be noted that before treatment is zoomed in and out to predicted motion vector, the present invention is implemented Example treatment is ranked up to the predicted motion vector in predicted motion vector set first, be specifically first according to index from it is small to Big order, selects the reference key identical predicted motion vector of reference key and image block motion vector to be encoded, then Reference key order from small to large is pressed again, selects other predicted motion vectors.By after above-mentioned treatment, to predicted motion to When amount zooms in and out treatment, actually it is mainly different from the reference key of image block motion vector to be encoded to reference key Predicted motion vector.

For example, it is assumed that the coordinate of a certain predicted motion vector is (4,8), reference key is 2；Image block to be encoded move to The coordinate of amount is (X, Y), and reference key is 1.When so being zoomed in and out to the predicted motion vector, its reference key is reduced 1 Times, its reference key can be made identical with the reference key of image block motion vector to be encoded.Accordingly, by the predicted motion While the reference key of vector reduces 1 times, the coordinate of the predicted motion vector is also reduced 1 times, be changed into (2,4).Therefore, it is right After the predicted motion vector zooms in and out treatment, its coordinate is (2,4), and reference key is 1.

The step in the embodiment of the method shown in step S301, S303~S305 difference corresponding diagrams 1 in the present embodiment S101, S102~S104, its particular content refer to the content of the embodiment of the method shown in corresponding diagram 1, and here is omitted.

Optionally, in another embodiment of the present invention, it is shown in Figure 4, it is described to encode and the image to be encoded The sequence number index of block motion vector identical predicted motion vector, including：

S403, generation are characterized in the predicted motion vector set, are existed and the image block motion vector to be encoded The identification information of identical reference motion vectors；

Specifically, during coded picture block motion vector coding is treated, when by generation predicted motion vector collection Close, and image block motion vector to be encoded and the predicted motion vector in predicted motion vector set are contrasted, find Exist in predicted motion vector set during with image block motion vector identical predicted motion vector to be encoded, generate a mark Will information, and the value of the flag information is set to 0.That is, when the value of the flag information is 0, representing in prediction In motion vector set, exist and image block motion vector identical predicted motion vector to be encoded.

S404, the coding identification information, and coding and the image block motion vector identical prediction fortune to be encoded The sequence number index of moving vector.

Specifically, the flag information and its value that are generated in step S403 are encoded, also, will be with image to be encoded Sequence number index of the motion vector identical predicted motion vector of block in predicted motion vector set is encoded.

For example, it is assumed that total n of predicted motion vector in predicted motion vector set, numbering be mvp0, mvp1 ... mvpn-1.If one predicted motion vector of the motion vector MV of image block to be encoded and any of the above is identical, if this is transported Sequence number index of the moving vector in predicted motion vector set is j, i.e. mvpj, then make ismvdabs0=0, otherwise make Ismvdabs0=1.Then ismvdabs0=0, and coding j are encoded, and terminates to treat the motion vector MV of coded picture block Cataloged procedure.

Step S401, S402, S405 in the present embodiment, the step in embodiment of the method shown in difference corresponding diagram 1 S101, S102, S104, its particular content refer to the content of the embodiment of the method shown in corresponding diagram 1, and here is omitted.

Optionally, in another embodiment of the present invention, it is shown in Figure 5, it is described according to the predicted motion vector Predicted motion vector in set, differential coding treatment is carried out to the image block motion vector to be encoded, obtains described waiting to compile The coding of code image block motion vector, including：

The X-component of S504, predicted motion vector in the predicted motion vector set, is calculated and described waits to compile The X-component difference of code image block motion vector；

Specifically, from predicted motion vector set, extraction obtains the X-component of all predicted motion vectors, used as waiting to compile The prediction of the X-component of code image block motion vector.Then, the prediction of the X-component according to image block motion vector to be encoded, from treating The absolute difference with the X-component of image block motion vector to be encoded is selected in the prediction of the X-component of coded picture block motion vector The prediction of the X-component of minimum image block motion vector to be encoded, the method for copying the H.264 difference of middle calculation of motion vectors, The difference of the prediction of the X-component of image block motion vector to be encoded and the X-component of the image block motion vector to be encoded is calculated, As the difference of the X-component of image block motion vector to be encoded.

The Y-component of S505, predicted motion vector in the predicted motion vector set, is calculated and described waits to compile The Y-component difference of code image block motion vector；

Specifically, from predicted motion vector set, extraction obtains the Y-component of all predicted motion vectors, used as waiting to compile The prediction of the Y-component of code image block motion vector.Then, the prediction of the Y-component according to image block motion vector to be encoded, from treating The absolute difference with the Y-component of image block motion vector to be encoded is selected in the prediction of the Y-component of coded picture block motion vector The prediction of the Y-component of minimum image block motion vector to be encoded, the method for copying the H.265 difference of middle calculation of motion vectors, The difference of the prediction of the Y-component of image block motion vector to be encoded and the Y-component of the image block motion vector to be encoded is calculated, As the difference of the Y-component of image block motion vector to be encoded.

S506, to the X-component difference, the sequence number index of the correspondence X-component difference, the Y-component difference and correspondence The sequence number index of the Y-component difference, carries out coded treatment, obtains the coding of the image block motion vector to be encoded.

Specifically, obtaining the X-component difference of image block motion vector to be encoded, and image block motion vector to be encoded Y-component difference after, according to conventional coding method, to the X-component difference, the sequence number index of the correspondence X-component difference, The sequence number index of the Y-component difference and the correspondence Y-component difference, carries out coded treatment, a series of codings for obtaining, as The coding of image block motion vector to be encoded.

The step S101 in embodiment of the method shown in step S501 in the present embodiment~S503 difference corresponding diagrams 1~ S103, its particular content refers to the content of the embodiment of the method shown in corresponding diagram 1, and here is omitted.

Optionally, in another embodiment of the present invention, the prediction in the predicted motion vector set The X-component of motion vector, is calculated the X-component difference of the image block motion vector to be encoded, including：

For example, after the predicted motion vector for searching the repetition for obtaining is rejected, in the predicted motion vector set for obtaining Predicted motion vector be mvp0, mvp1 ... mvpn-1, in this step process, extraction obtains the X point of each predicted motion vector Amount, i.e. mvpx0, mvpx1 ... mvpxn-1.

For example, generation First ray, wherein comprising mvpx0.

Specifically, with reference to the example above, travel through mvpx1 to mvpxn-1 from small to large one by one, if currently mvpxi and the The distance of existing mvpx is less than or equal to 1 in one sequence, then reject this mvpxi, and this mvpxi otherwise is added into the first sequence Row.

Specifically, with reference to the example above, after having traveled through mvpx1 to mvpxn-1, if the unit in First ray have m, i.e., Mvpxx0 to mvpxxm-1.If the x-component of the motion vector of the image block with coding is mvx, one is selected from First ray The minimum mvpxx with its absolute difference, if its sequence number index is mvpxxt for t, i.e. this mvpxx.

Specifically, with reference to the example above, the X-component difference mvxdiff=mvx- of image block motion vector to be encoded mvpxxt。

Optionally, in another embodiment of the present invention, the prediction in the predicted motion vector set The Y-component of motion vector, is calculated the Y-component difference of the image block motion vector to be encoded, including：

For example, after the predicted motion vector for searching the repetition for obtaining is rejected, in the predicted motion vector set for obtaining Predicted motion vector be mvp0, mvp1 ... mvpn-1, in this step process, extraction obtains the Y point of each predicted motion vector Amount, i.e. mvpy0, mvpy1 ... mvpyn-1.

Specifically, with reference to the example above, according to the absolute difference minimum with the X-component of image block motion vector to be encoded The X-component mvpxxt of predicted motion vector, by mvpy0, mvpy1 ... mvpyn-1 is divided into the second sequence and the 3rd sequence.Second The X-component mvpx of the corresponding predicted motion vector of Y-component of the predicted motion vector in sequence all equal to mvpxxt, the 3rd sequence The X-component mvpx of the corresponding predicted motion vector of Y-component of the predicted motion vector in row is all not equal to mvpxxt.

Specifically, with reference to the example above, traveling through Y points of remaining predicted motion vector in the second sequence from small to large one by one Amount, if the Y-component of current predictive motion vector is less than with the distance of the Y-component of existing predicted motion vector in the 4th sequence Equal to 1, then reject the Y-component of this predicted motion vector, the Y-component of this predicted motion vector otherwise is added into the 4th sequence Row.

Specifically, with reference to the example above, the Y of remaining predicted motion vector divides in the 3rd sequence of traversal from small to large one by one Amount, if the Y-component of current predictive motion vector is less than with the distance of the Y-component of existing predicted motion vector in the 5th sequence Equal to 1, then reject the Y-component of this predicted motion vector, the Y-component of this predicted motion vector otherwise is added into the 5th sequence Row.

Specifically, with reference to the example above, if the y-component of the motion vector of image block to be encoded is mvy, finding the 4th sequence With that mvpy minimum with mvy absolute differences in the 5th sequence.It is no if it is in the 4th sequence, then set isusemvxeq=1 Then set isusemvxeq=0.

Specifically, with reference to the example above, if sequence number indexes of this mvpy in corresponding that sequence is i, i.e., this Mvpy is mvpyi, then the Y-component difference mvydiff=mvy-mvpyi of image block motion vector to be encoded.

Optionally, in another embodiment of the present invention, it is described to the X-component difference, the correspondence X-component difference Sequence number index, the Y-component difference and the correspondence Y-component difference sequence number index, carry out coded treatment, obtain described treating The coding of coded picture block motion vector, including：

Specifically, with reference to the example above, abs (mvxdiff)+abs (mvydiff) is encoded first, abs is then encoded (mvxdiff).If mvxdiff is not equal to zero, then the symbol of coding mvxdiff, if then mvydiff is not equal to 0, that Encode the symbol of mvydiff.Then the index t of mvpx is encoded, isusemvxeq is next encoded, finally encodes mvpy's Index i.

By above-mentioned treatment, that is, complete the coded treatment for treating coded picture block motion vector.Accordingly, to using When the image block motion vector of above-mentioned coding method coding is decoded, with reference to the citing in the various embodiments described above, its process is general State as follows：First as cataloged procedure, mvp sequences are first constructed, ismvdabs0 is then decoded, if 0, then decoding mvp Index j after terminate this process.Otherwise successively decoding abs (mvxdiff)+abs (mvydiff), abs (mvxdiff), The symbol of mvxdiff, the symbol of mvydiff.Mvxdiff and mvydiff is obtained after decoding.Then constructed as cataloged procedure The forecasting sequence of the x-component of image block motion vector, then decodes the index t of mvpx, and mvpxxt is obtained from sequence.Root again Two forecasting sequences of y are constructed as in cataloged procedure according to mvpxxt, isusemvxeq is then decoded, corresponding sequence is selected Row, decode the index i of mvpy, and mvpyi is obtained from the corresponding forecasting sequence of y.Finally obtain x points of image block motion vector Amount mvx=mvxdiff+mvpxxt, y-component mvy=mvydiff+mvpyi.

Shown in Figure 6 the embodiment of the invention also discloses a kind of motion vector code device, the device includes：

Prediction processing unit 601, for from the motion vector of the reference prediction encoding block of image block to be encoded, searching To setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition motion vector set；

Judging unit 602, for by the image block motion vector to be encoded respectively with the predicted motion vector set In predicted motion vector contrasted, judge in the predicted motion vector set, if exist and the figure to be encoded As block motion vector identical predicted motion vector；

First coding unit 603, judges in the predicted motion vector set exist for the judging unit 602 During with the image block motion vector identical predicted motion vector to be encoded, encode and the image block motion vector to be encoded The sequence number index of identical predicted motion vector, as the coding of the image block motion vector to be encoded；

Second coding unit 604, judges in the predicted motion vector set for the judging unit 602, does not deposit When with the image block motion vector identical predicted motion vector to be encoded, according in the predicted motion vector set Predicted motion vector, differential coding treatment is carried out to the image block motion vector to be encoded, obtains the image block to be encoded The coding of motion vector.

Specifically, in the present embodiment unit specific works content, refer to the content of corresponding embodiment of the method, Here is omitted.

Motion vector code device proposed by the present invention, it is single by prediction treatment first when being encoded to image motion vector Unit 601 obtain setting quantity present encoding block motion vectors predicted motion vectors, and to set quantity predicted motion to Amount zooms in and out treatment；Then, the contrast of judging unit 602 judges to whether there is in predicted motion vector and is transported with present encoding block Moving vector identical predicted motion vector, if it has, the then sequence number of the direct coding predicted motion vector of the first coding unit 603 Index；If it is not, the second coding unit 604 according to setting quantity predicted motion vector, present encoding block is moved to Amount carries out differential coding treatment.In above-mentioned coding process, multiple predicted motion vectors are applied to, and to present encoding When block motion vector is encoded, first in view of whether having in multiple predicted motion vectors and present encoding block motion vector Identical predicted motion vector, if it has, then directly using the coding of the predicted motion vector as present encoding block motion vector Coding, only encode the predicted motion vector sequence number index；If it is not, using above-mentioned multiple predicted motions to Amount, encodes to present encoding block motion vector.The predicted motion vector that its cataloged procedure is applied to is more, considers comprehensively not Same situation carries out corresponding coded treatment, and code efficiency is higher.

Optionally, in another embodiment of the present invention, ginseng of the prediction processing unit 601 from image block to be encoded Examine in the motion vector of predictive coding block, lookup obtains setting quantity, the prediction fortune of the image block motion vector to be encoded Moving vector, during predicted composition motion vector set, specifically for：

Specifically, in the present embodiment prediction processing unit 601 specific works content, refer to corresponding embodiment of the method Content, here is omitted.

Optionally, in another embodiment of the present invention, shown in Figure 7, described device also includes：

Scaling processing unit 605, it is for the reference key according to the image block motion vector to be encoded and described pre- The reference key of the predicted motion vector surveyed in motion vector set, transports to the prediction in the predicted motion vector set respectively Moving vector zooms in and out treatment, the predicted motion vector in the predicted motion vector set is transported with the image block to be encoded Side-play amount represented by moving vector is identical.

Specifically, scaling the specific works content of processing unit 605 in the present embodiment, corresponding embodiment of the method is referred to Content, here is omitted.

Optionally, in another embodiment of the present invention, first coding unit 603 is encoded and the figure to be encoded When being indexed as the sequence number of block motion vector identical predicted motion vector, specifically for：

Specifically, in the present embodiment the first coding unit 603 specific works content, refer to corresponding embodiment of the method Content, here is omitted.

Optionally, in another embodiment of the present invention, it is shown in Figure 8, the second coding unit 604, including：

First component calculation unit 6041, for the X of the predicted motion vector in the predicted motion vector set Component, is calculated the X-component difference of the image block motion vector to be encoded；

Second component computing unit 6042, for the Y of the predicted motion vector in the predicted motion vector set Component, is calculated the Y-component difference of the image block motion vector to be encoded；

Differential coding unit 6043, for the X-component difference, the sequence number index of the correspondence X-component difference, described The sequence number index of Y-component difference and the correspondence Y-component difference, carries out coded treatment, obtains the image block motion to be encoded The coding of vector.

Optionally, in another embodiment of the present invention, first component calculation unit 6041 is according to the prediction The X-component of the predicted motion vector in motion vector set, the X-component for being calculated the image block motion vector to be encoded is poor Timesharing, specifically for：

Specifically, in the present embodiment the first component calculation unit 6041 specific works content, refer to corresponding method The content of embodiment, here is omitted.

Optionally, in another embodiment of the present invention, the second component computing unit 6042 is according to the prediction The Y-component of the predicted motion vector in motion vector set, the Y-component for being calculated the image block motion vector to be encoded is poor Timesharing, specifically for：

Specifically, in the present embodiment second component computing unit 6042 specific works content, refer to corresponding method The content of embodiment, here is omitted.

Optionally, in another embodiment of the present invention, the X-component difference of the differential coding unit 6043 pairs, The sequence number index of the sequence number index, the Y-component difference and the correspondence Y-component difference of the correspondence X-component difference, is compiled Code treatment, when obtaining the coding of the image block motion vector to be encoded, specifically for：

Specifically, in the present embodiment differential coding unit 6043 specific works content, refer to corresponding method and implement The content of example, here is omitted.

The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or uses the present invention. Various modifications to these embodiments will be apparent for those skilled in the art, as defined herein General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, the present invention The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one The scope most wide for causing.

Claims

1. a kind of motion-vector coding method, it is characterised in that including：

From the motion vector of the reference prediction encoding block of image block to be encoded, lookup obtains setting quantity, described to be encoded The predicted motion vector of image block motion vector, predicted composition motion vector set；

The image block motion vector to be encoded is carried out with the predicted motion vector in the predicted motion vector set respectively Contrast, judges in the predicted motion vector set, if exist pre- with the image block motion vector identical to be encoded Survey motion vector；

If in the predicted motion vector set, existed and the image block motion vector identical predicted motion to be encoded Vector, then encode and indexed with the sequence number of the image block motion vector identical predicted motion vector to be encoded, is treated as described The coding of coded picture block motion vector；

If in the predicted motion vector set, transported in the absence of with the image block motion vector identical prediction to be encoded Moving vector, the then predicted motion vector in the predicted motion vector set, to the image block motion vector to be encoded Differential coding treatment is carried out, the coding of the image block motion vector to be encoded is obtained.

2. method according to claim 1, it is characterised in that the reference prediction encoding block from image block to be encoded In motion vector, lookup obtains setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition Motion vector set, including：

According to efficient video coding HEVC coding methods, from the motion vector of the spatial domain adjacent encoder block of image block to be encoded, with And in time domain reference picture in the motion vector of the encoding block of same position, lookup obtains setting quantity, correspondence figure to be encoded As the predicted motion vector of block motion vector, predicted composition motion vector set；

In rejecting the predicted motion vector set, in setting the predicted motion vector of quantity, the predicted motion vector for repeating.

3. method according to claim 1, it is characterised in that in the fortune of the reference prediction encoding block from image block to be encoded In moving vector, lookup obtains setting quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition fortune After trend duration set, by the image block motion vector to be encoded respectively with the predicted motion vector set in prediction Motion vector is contrasted, and is judged in the predicted motion vector set, if is existed and is moved with the image block to be encoded Before vectorial identical predicted motion vector, the method also includes：

According to the reference key of the image block motion vector to be encoded, and the prediction fortune in the predicted motion vector set The reference key of moving vector, zooms in and out treatment to the predicted motion vector in the predicted motion vector set respectively, makes institute State the predicted motion vector in predicted motion vector set and the side-play amount phase represented by the image block motion vector to be encoded Together.

4. method according to claim 1, it is characterised in that the coding and the image block motion vector phase to be encoded The sequence number index of same predicted motion vector, including：

Generation is characterized in the predicted motion vector set, is existed and is referred to the image block motion vector identical to be encoded The identification information of motion vector；

The identification information is encoded, and encodes the sequence with the image block motion vector identical predicted motion vector to be encoded Number index.

5. method according to claim 1, it is characterised in that the prediction in the predicted motion vector set Motion vector, differential coding treatment is carried out to the image block motion vector to be encoded, obtains the image block motion to be encoded The coding of vector, including：

The X-component of the predicted motion vector in the predicted motion vector set, is calculated the image block to be encoded The X-component difference of motion vector；

The Y-component of the predicted motion vector in the predicted motion vector set, is calculated the image block to be encoded The Y-component difference of motion vector；

To the X-component difference, the sequence number index of the correspondence X-component difference, the Y-component difference and the correspondence Y-component The sequence number index of difference, carries out coded treatment, obtains the coding of the image block motion vector to be encoded.

6. method according to claim 5, it is characterised in that the prediction in the predicted motion vector set The X-component of motion vector, is calculated the X-component difference of the image block motion vector to be encoded, including：

Generation First ray, comprising first predicted motion vector in the predicted motion vector set in the First ray X-component；

The X-component of the remaining predicted motion vector in the predicted motion vector set is traveled through, a predicted motion is often traversed During the X-component of vector, the X-component of the predicted motion vector in the X-component of the predicted motion vector and the First ray is calculated Distance；If the distance of the X-component of the predicted motion vector in the X-component of the predicted motion vector and the First ray is all More than 1, then the X-component of the predicted motion vector is added into the First ray；

From the First ray, lookup obtains minimum with the absolute difference of the X-component of the image block motion vector to be encoded Predicted motion vector X-component, pre- minimum with the absolute difference of the X-component of the image block motion vector to be encoded of record Survey sequence number index of the X-component of motion vector in the First ray；

According to the X-component of the minimum predicted motion vector of the absolute difference with the X-component of the image block motion vector to be encoded, And the X-component of the image block motion vector to be encoded, the X-component for being calculated the image block motion vector to be encoded is poor Point.

7. method according to claim 6, it is characterised in that the prediction in the predicted motion vector set The Y-component of motion vector, is calculated the Y-component difference of the image block motion vector to be encoded, including：

By the Y-component of the predicted motion vector in the predicted motion vector set, the second sequence and the 3rd sequence are divided into；Its In, the X-component of the corresponding predicted motion vector of Y-component of the predicted motion vector in second sequence, with first sequence In row, the X-component phase with the predicted motion vector of the absolute difference minimum of the X-component of the image block motion vector to be encoded Together；The X-component of the corresponding predicted motion vector of Y-component of the predicted motion vector in the 3rd sequence, with first sequence In row, the X-component of minimum predicted motion vector is not with the absolute difference of the X-component of the image block motion vector to be encoded Together；

The 4th sequence is generated, comprising first Y-component of predicted motion vector in second sequence in the 4th sequence；

The Y-component of remaining predicted motion vector in second sequence is traveled through, a Y for predicted motion vector points is often traversed During amount, the distance of the Y-component of predicted motion vector in the Y-component of the predicted motion vector and the 4th sequence is calculated；Such as Really the distance of the Y-component of the predicted motion vector in the Y-component of the predicted motion vector and the 4th sequence is both greater than 1, then The Y-component of the predicted motion vector is added into the 4th sequence；

The 5th sequence is generated, comprising first Y-component of predicted motion vector in the 3rd sequence in the 5th sequence；

The Y-component of remaining predicted motion vector in the 3rd sequence is traveled through, a Y for predicted motion vector points is often traversed During amount, the distance of the Y-component of predicted motion vector in the Y-component of the predicted motion vector and the 5th sequence is calculated；Such as Really the distance of the Y-component of the predicted motion vector in the Y-component of the predicted motion vector and the 5th sequence is both greater than 1, then The Y-component of the predicted motion vector is added into the 5th sequence；

From the 4th sequence and the 5th sequence, lookup obtains the Y-component with the image block motion vector to be encoded The minimum predicted motion vector of absolute difference Y-component, record exhausted with the Y-component of the image block motion vector to be encoded To the sequence where the Y-component of the minimum predicted motion vector of value difference, and Y points with the image block motion vector to be encoded Sequence index in sequence of the Y-component of the minimum predicted motion vector of the absolute difference of amount where it；

According to the Y-component of the minimum predicted motion vector of the absolute difference with the Y-component of the image block motion vector to be encoded, And the Y-component of the image block motion vector to be encoded, the Y-component for being calculated the image block motion vector to be encoded is poor Point.

8. method according to claim 7, it is characterised in that described poor to the X-component difference, the correspondence X-component The sequence number index of the sequence number index, the Y-component difference and the correspondence Y-component difference that divide, carries out coded treatment, obtains described The coding of image block motion vector to be encoded, including：

Absolute value sum, the absolute value of the X-component difference, the X point to the X-component difference and the Y-component difference The prediction that the absolute difference of the X-component of amount difference, the Y-component difference and the image block motion vector to be encoded is minimum is transported Sequence number index of the X-component of moving vector in the First ray, characterize and the Y-component of the image block motion vector to be encoded The minimum predicted motion vector of absolute difference Y-component where sequence information and the image block to be encoded move to Sequence index in sequence of the Y-component of the minimum predicted motion vector of the absolute difference of the Y-component of amount where it is entered respectively Row coded treatment.

9. a kind of motion vector code device, it is characterised in that including：

Prediction processing unit, for from the motion vector of the reference prediction encoding block of image block to be encoded, lookup to be set Quantity, the predicted motion vector of the image block motion vector to be encoded, predicted composition motion vector set；

Judging unit, for by the image block motion vector to be encoded respectively with the predicted motion vector set in prediction Motion vector is contrasted, and is judged in the predicted motion vector set, if is existed and is moved with the image block to be encoded Vectorial identical predicted motion vector；

First coding unit, judges in the predicted motion vector set for the judging unit, exists and waits to compile with described During code image block motion vector identical predicted motion vector, coding is predicted with the image block motion vector identical to be encoded The sequence number index of motion vector, as the coding of the image block motion vector to be encoded；

Second coding unit, judges in the predicted motion vector set for the judging unit, is treated in the absence of with described During coded picture block motion vector identical predicted motion vector, predicted motion in the predicted motion vector set to Amount, differential coding treatment is carried out to the image block motion vector to be encoded, obtains the image block motion vector to be encoded Coding.

10. device according to claim 9, it is characterised in that ginseng of the prediction processing unit from image block to be encoded Examine in the motion vector of predictive coding block, lookup obtains setting quantity, the prediction fortune of the image block motion vector to be encoded Moving vector, during predicted composition motion vector set, specifically for：

According to efficient video coding HEVC coding methods, from the motion vector of the spatial domain adjacent encoder block of image block to be encoded, with And in time domain reference picture in the motion vector of the encoding block of same position, lookup obtains setting quantity, correspondence figure to be encoded As the predicted motion vector of block motion vector, predicted composition motion vector set；In rejecting the predicted motion vector set, if In the predicted motion vector of fixed number amount, the predicted motion vector for repeating.

11. devices according to claim 9, it is characterised in that described device also includes：

Scaling processing unit, for the reference key according to the image block motion vector to be encoded, and the predicted motion The reference key of the predicted motion vector in vector set, respectively to the predicted motion vector in the predicted motion vector set Treatment is zoomed in and out, makes predicted motion vector and the image block motion vector to be encoded in the predicted motion vector set Represented side-play amount is identical.

12. devices according to claim 9, it is characterised in that the first coding unit coding and the figure to be encoded When being indexed as the sequence number of block motion vector identical predicted motion vector, specifically for：

Generation is characterized in the predicted motion vector set, is existed and is referred to the image block motion vector identical to be encoded The identification information of motion vector；Encode the identification information, and coding and the image block motion vector identical to be encoded The sequence number index of predicted motion vector.

13. devices according to claim 9, it is characterised in that second coding unit, including：

First component calculation unit, for the X-component of the predicted motion vector in the predicted motion vector set, calculates Obtain the X-component difference of the image block motion vector to be encoded；

Second component computing unit, for the Y-component of the predicted motion vector in the predicted motion vector set, calculates Obtain the Y-component difference of the image block motion vector to be encoded；

Differential coding unit, for poor to the X-component difference, the sequence number index of the correspondence X-component difference, the Y-component Divide and the sequence number of the correspondence Y-component difference is indexed, carry out coded treatment, obtain the volume of the image block motion vector to be encoded Code.

14. devices according to claim 13, it is characterised in that first component calculation unit is according to the prediction fortune The X-component of the predicted motion vector in trend duration set, is calculated the X-component difference of the image block motion vector to be encoded When, specifically for：

15. devices according to claim 14, it is characterised in that the second component computing unit is according to the prediction fortune The Y-component of the predicted motion vector in trend duration set, is calculated the Y-component difference of the image block motion vector to be encoded When, specifically for：

16. devices according to claim 15, it is characterised in that the differential coding unit is to the X-component difference, right The sequence number for answering the X-component difference is indexed, the sequence number of the Y-component difference and the correspondence Y-component difference is indexed, and is encoded Treatment, when obtaining the coding of the image block motion vector to be encoded, specifically for：