CN108111833A

CN108111833A - For the method, apparatus and system of stereo video coding-decoding

Info

Publication number: CN108111833A
Application number: CN201611043145.5A
Authority: CN
Inventors: 黄敦笔; 张磊; 杜武平
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2016-11-24
Filing date: 2016-11-24
Publication date: 2018-06-01

Abstract

This application discloses a kind of method, including：N number of sequence of video images of carrying three-dimensional video-frequency is obtained, N is the integer more than or equal to 2；Based on N number of sequence of video images, 1 the second image sequence of the first image sequence and N is determined；By 1 the second coding image sequences generation three-dimensional video-frequency bit stream of described first image sequence and N；Wherein, coding mode includes used by being encoded for the second image sequence：Inter-sequence prediction coding mode；The inter-sequence prediction coding mode refers to, predictive coding is carried out using the image in described first image sequence as reference frame, to the block of pixels in second image sequence.

Description

For the method, apparatus and system of stereo video coding-decoding

Technical field

This application involves three-dimensional video-frequency technical fields, and in particular to method, apparatus, system for stereo video coding-decoding And machine readable media.

Background technology

As industrial technology is constantly progressive, 3D TVs and three-dimensional film (3D films) are increasingly popularized, some consumers are not Meet again and obtain for the sense organ of traditional two dimension (2D) video content information and entertain demand, and tend to three-dimensional video-frequency and more force Genuine effect experience.Three-dimensional video-frequency compares traditional two-dimensional video, can more show the sense of depth and stereovision of scenic picture, because This reduces reality with more telepresenc and authenticity using three-dimensional video-frequency technology, and which represent the weights of virtualized reality technology Want direction.

Three-dimensional video-frequency technology utilizes the binocular parallax principle of people, synchronously shot with video-corder under the conditions of Same Scene using video camera to Then few two sequence of video images generate three-dimensional video-frequency bit stream by coded treatment, and store on a storage medium or logical It crosses network and is sent to recipient.When needing to play three-dimensional video-frequency, to being read from storage medium or receive from network vertical Reduction generation stereo video signals, are sent to stereo video display and show, pass through binocular sense after the decoded processing of volumetric video bit stream Parallax on official, so that beholder gets longitudinal depth information of scene and experiences three-dimensional sense.

For the prior art when at least two sequence of video images are carried out coded treatment, generally use ITU is H.264/AVC Or the coding standards such as ISO MPEG-H HEVC carry out absolute coding to each sequence of video images respectively, cause life after coding Into the usual data volume of three-dimensional video-frequency bit stream it is larger, so as to propose larger challenge to memory space or network bandwidth.With logical It crosses exemplified by network transmission three-dimensional video-frequency bit stream, due to needing to occupy more network bandwidth, can not be met the requirements in network bandwidth In the case of can because network congestion packet loss three-dimensional video-frequency playing process be caused interim card or higher delay occur, influence user's Viewing experience.

The content of the invention

The application provides a kind of method, including：

N number of sequence of video images of carrying three-dimensional video-frequency is obtained, N is the integer more than or equal to 2；

Based on N number of sequence of video images, the first image sequence and N-1 the second image sequences are determined；

By described first image sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；

Wherein, coding mode includes used by being encoded for the second image sequence：Inter-sequence prediction coding mode； The inter-sequence prediction coding mode refers to, using the image in described first image sequence as reference frame, to second image Block of pixels in sequence carries out predictive coding.

Description of the drawings

Fig. 1 is a kind of flow chart of the embodiment for method that the application provides；

Fig. 2 is one of the image that left eye and right eye are corresponded respectively in sequence of video images provided by the embodiments of the present application Example；

Fig. 3 is the schematic diagram provided by the embodiments of the present application using upper and lower coding mode generation sequence of subsampled images；

Fig. 4 is the process flow that the image to be encoded provided by the embodiments of the present application in the second image sequence is encoded Figure；

Fig. 5 is a kind of schematic diagram of the embodiment for device that the application provides；

Fig. 6 is the flow chart of the embodiment for another method that the application provides；

Fig. 7 is the schematic diagram of the embodiment for another device that the application provides；

Fig. 8 is the schematic diagram for the exemplary system that the application provides；

Fig. 9 is a kind of schematic diagram of the embodiment for system that the application provides.

Specific embodiment

Many details are elaborated in the following description in order to fully understand the application.But the application can Much to implement different from other manner described here, those skilled in the art can be in the feelings without prejudice to the application intension Similar popularization is done under condition, therefore, the application from following public specific implementation limitation.

The technical solution of the application can have different change embodiment or alternate embodiments, and this specification will be tied The example that attached drawing provides is closed specific embodiment is described in detail.It will be understood, however, to one skilled in the art, that this theory The purpose of bright book is not configured to technical scheme being defined in the particular implementation of this disclosure, but in order to cover Lid have altered embodiment, equivalent embodiments and the alternate embodiments consistent with technical scheme.

In the present specification to the reference of " embodiment ", " the present embodiment " or " example embodiment " etc., indicate described Embodiment can include specific feature, structure or characteristic, but do not require that each embodiment is required for including this specifically Feature, structure or characteristic.In addition, when describing specific feature, structure or characteristic with reference to one embodiment, in this field skill In the knowledge of art personnel, this feature can be implemented with reference to other embodiment, (no matter whether it is detailed for structure or characteristic Thin description).

Embodiments herein can be realized according to software, hardware, firmware or its combination or other modes.The application's Embodiment can also be achieved to be stored in impermanency or permanent machine readable media (such as：Computer-readable medium) on finger Order, described instruction can be read or be performed by one or more processors.Machine readable media includes any by machine readable shape Formula stores or storage device, mechanism or other physical arrangements of transmission information.For example, machine readable media can include read-only deposit Reservoir (ROM), random access storage device (RAM), magnetic disk storage medium, optical storage media, flash memory device and other.

In the attached drawing provided in this specification, some structures or method characteristic be typically according to specific arrangement mode and/ Or order is shown.However, it should be understood that these specific arrangement modes and/or order are not required in that.In some realities It applies in example, these features can carry out tissue according to the arrangement mode different from attached drawing and/or order.In addition, in a certain attached drawing Comprising structure or method characteristic, be not meant in all embodiments will include this feature, in some embodiments, can It can be combined with not including this feature or this feature with other features.

In order to make it easy to understand, first the technical solution of the application is briefly described.

N number of sequence of video images of carrying three-dimensional video-frequency is typically synchronously to be shot with video-corder for Same Scene, by different angle Sequence of video images, N is greater than the integer equal to 2.For N number of sequence of video images, one of video image sequence Picture frame in row is usually respectively provided with corresponding picture frame, i.e., in other sequence of video images：Synchronization is from difference The picture frame that angle is shot with video-corder, these picture frames have identical temporal information, in the specific implementation, can use the time of picture frame Stamp mark.

H.264/AVC or ISO MPEG-H for the prior art when being encoded for three-dimensional video-frequency, generally use ITU The coding standards such as HEVC carry out absolute coding to each sequence of video images respectively, cause the three-dimensional video-frequency bit stream obtained after coding Data volume it is larger, higher requirement is proposed to memory space and network transmission bandwidth.

And in fact, adhere to the picture frame with same or similar temporal information of different video image sequence separately, due to It is shot with video-corder at the same or similar moment for Same Scene, therefore with stronger correlation, to realize further number Possibility is provided according to compressed encoding.Based on above-mentioned consideration, the technical solution of the application is determining first based on sequence of video images After image sequence and N-1 the second image sequences, during to the second coding image sequences, introduce different from tradition The inter-sequence prediction coding mode of intra-prediction code mode and inter prediction encoding pattern, i.e.,：In to the second image sequence Block of pixels encoded during, can using the image in the first image sequence as with reference to frame carry out predictive coding, from And can significantly promote the data compression rate in cataloged procedure, reduce the three-dimensional video-frequency bit stream of coding generation to memory space or The occupancy of person's network bandwidth.

In the technical program, the predictive coding refers to that there is the spies of certain relevance according between discrete signal Point, using front one or more signal estimation next signal, and to one kind that the difference of actual value and predicted value is encoded Coding techniques.The block of pixels refers to what is be made of more than one adjacent pixel of a pixel in image or space Image block, such as：8 × 8 block of pixels are exactly the image block being made of 8 rows adjacent in image, 8 row pixels.

In technical scheme, the number for carrying the sequence of video images of three-dimensional video-frequency is N, and N is greater than equal to 2 Integer.In the examples below for convenience of description, emphasis carries out the implementation process of technical scheme by taking N=2 as an example Explanation.

In the following, a kind of embodiment of method of the application offer is described in detail.It please refers to Fig.1, is the application A kind of flow chart of the embodiment of the method provided.Described method includes following steps：

Step 101, two sequence of video images for obtaining carrying three-dimensional video-frequency.

This step obtains two sequence of video images of carrying three-dimensional video-frequency, and described two sequence of video images are usually distinguished Sequence of video images to correspond to left eye and the sequence of video images corresponding to right eye.Fig. 2 is referred to, it illustrates adhere to separately Two stereoscopic video images sequences, there are the two field pictures of same time information, wherein (a) is the image corresponding to left eye, (b) it is the image corresponding to right eye.

When it is implemented, two sequence of video images shot with video-corder respectively by two photographic devices can be obtained, can also lead to It crosses the multimedia resource file for reading storing stereoscopic video and performs corresponding transcoding operation, obtains two sequence of video images. Two sequence of video images include a series of images frame, and the figure shot with video-corder in two sequence of video images in synchronization respectively As frame have identical temporal information, for example, with identical timestamp.

In the specific implementation, this step can also pass through photographic device master while two sequence of video images are obtained The dynamic information reported or by the analysis to video image, obtains image parameter information, including：The resolution and figure of image The color format of picture.Wherein, resolution includes picture traverse w and picture altitude h, and unit is pixel；Color format typically refers to, The representation of image color space, such as RGB or YUV are described.

Step 102, based on described two sequence of video images, determine the first image sequence and the second image sequence.

This step further determines that two image sequences on the basis of step 101 obtains two sequence of video images, this Two image sequences are the input signals that step 103 is encoded, for the two sequence of video images phases obtained with step 101 It distinguishes, the image sequence that this step is determined is referred to as the first image sequence and the second image sequence.

As simple and practicable embodiment, can directly using corresponding to the sequence of video images of left eye as the first image Sequence, using corresponding to the sequence of video images of right eye as the second image sequence.Same reason will can also directly correspond to The sequence of video images of right eye as the first image sequence, using corresponding to the sequence of video images of left eye as the second image sequence Row, it is also possible.

Preferably, in order to promote compression rates, the first image sequence and the are determined again the present embodiment provides first down-sampled The preferred embodiment of two image sequences.It specifically, can be according to default down-sampled mode, in two sequence of video images Every two field picture carry out respectively it is down-sampled, then selected in two sequence of video images after down-sampled first image sequence and Second image sequence.

It is down-sampled described in the present embodiment to refer to, by being sampled to image, the image processing method to reduce number of pixels Formula.Default down-sampled mode includes：Interlacing is down-sampled, every arranging, down-sampled or interpolation is down-sampled.It is down-sampled for interlacing, Every two field picture in two sequence of video images can be handled as follows：Only retain the even number line in image or only protect Stay the odd-numbered line in image；For every arrange it is down-sampled, then can only retain even column or odd column in image；For interpolation It is down-sampled, new pixel column can be generated or by by adjacent two pixel by the way that adjacent two pixel columns interpolation smoothing is calculated Row interpolation smoothing, which calculates, generates new pixel column, and the image with new pixel column or pixel column synthesis after down-sampled.Using drop Sample mode, can be before step 103 be encoded, by the first image sequence as input and the number of the second image sequence It is greatly decreased according to amount, such as：Using above-mentioned listed down-sampled mode, data volume can be reduced half, it is whole so as to be promoted Data compression rate.After completing down-sampled operation, selected in two sequence of video images after down-sampled first image sequence and Second image sequence.

Preferably, in order to promote compression rates, the present embodiment also provides one kind based on sequence of subsampled images preferably Embodiment.Specifically, this step can be according to default pretreatment coding mode, according to two sequence of video images generation Sequence of subsampled images；According to the corresponding fractionation mode of the pretreatment coding mode, by every frame figure in sequence of subsampled images As splitting into two images with same time information, so as to obtain two image sequences；It is chosen from two image sequences First image sequence, and using another image sequence as the second image sequence.

The pretreatment coding mode includes：Interlacing coding mode, upper and lower coding mode, left and right coding mode etc. or Checkerboard codes pattern.

The interlacing coding mode can be, identical to having in a sequence of video images and another sequence of video images The two field pictures of temporal information are handled as follows：By the wherein even number line of the odd-line pixels of a two field picture and another two field picture Pixel interlacing intersects, and synthesizes a frame new images.

The coding mode up and down can be, identical to having in a sequence of video images and another sequence of video images The two field pictures of temporal information are handled as follows：By the odd-line pixels of a wherein two field picture either even rows or phase Pixel column that adjacent rows interpolation obtains simultaneously synthesizes first object image, by the even rows or odd-line pixels of another two field picture Or the obtained pixel column of adjacent row interpolation and the second target image is synthesized, and first object image and the second icon image are erected Directly it is spliced into a frame new images.

The left and right coding mode is similar with upper and lower coding mode, and difference lies in generation first object image and the second mesh It is that image column is handled during logo image, and finally by first object image and the horizontal splicing of the second icon image Into a frame new images.

Further, it is also possible to coding mode, such as checker board pattern are pre-processed using other, i.e.,：With same time information The distribution of pixel in two field pictures in post synthesis in new images meets the form that is interspersed of the black and white lattice in chessboard.

This step is in the sequence of video images obtained to step 101, every two first according to default pretreatment coding mode The image that frame has same time information is all spliced accordingly or synthesis processing, so as to obtain a new image sequence, Since new image sequence derives from the sequence of video images of step 101 acquisition, but number of pixels reduces half, therefore claims For sequence of subsampled images.Then, according to fractionation mode corresponding with pretreatment coding mode (for example, coding mode pair up and down Mode should be split up and down, and left and right coding mode corresponds to left and right fractionation mode), every two field picture in sequence of subsampled images is split Into two images with same time information, so as to obtain two image sequences；Finally, is selected from two image sequences One image sequence, and using another image sequence as the second image sequence.

Below by taking upper and lower coding mode as an example, the embodiment of above-mentioned generation sequence of subsampled images is illustrated.Please Referring to Fig. 3, wherein, the resolution of every two field picture in sequence of video images is w × h, and (a) is the video image corresponding to left eye Image in sequence, (b) are the image for having same time information corresponding in the sequence of video images of right eye and (a), (c) new images to be spliced into vertically using upper and lower coding mode.LR_n, RR_nThe line n pixel in (a) and (b) is represented respectively. (c) the row pixel value VR in_n' can be calculated by equation 1 below, formula 2 or formula 3.

VR_n'=VR_2n；--- --- formula 1

VR_n'=VR_2n+1；--- --- formula 2

VR_n'=α × VR_2n+(1-α)×VR_2n+1；--- --- formula 3

Wherein, V corresponds to 0,1,2,3 .. etc. integers of L or R, n.In equation 3, α is the power for carrying out interpolation calculation Weight coefficient when α values take 0.5, represents that even number line and odd-numbered line equally sample, α>When 0.5, even number line is sampled prior to odd-numbered line, α <When 0.5, odd-numbered line is sampled prior to even number line.

It, can be according to the corresponding fractionation side of upper and lower coding mode after obtaining sequence of subsampled images using aforesaid way Formula splits every two field picture in sequence of subsampled images, such as：It, can be according to heavy line institute for (c) in Fig. 3 The horizontal center line shown is split, and obtains two images with same time information, the resolution of each image for w × (h/2)。

Using the embodiment described above based on sequence of subsampled images, flexible synthesis fractionation side can be provided Formula for example, after synthesizing sequence of subsampled images using checker board pattern, can be split according to the chequered with black and white mode of chessboard, It can also be split according to left and right or upper and lower mode, there is further flexibility so as to input to the image sequence of step 103, Different compression ratios can be obtained in a particular application.Wherein, compared with other pretreatment coding modes, due to encoding up and down Pattern and the image sequence of left and right coding mode generation include the independent image texture letter corresponding to left eye or right eye respectively Breath, i.e.,：There is the image context information of more horn of plenty, be therefore preferred pretreatment coding mode.

In the specific implementation, each image in the first image sequence and the second image sequence can retain video image sequence The temporal information of respective image in row, and with can be with the index of unique mark oneself.By the processing of this step, the first figure As sequence and the second image sequence still carry stereo video signals, and equally have the data related between two sequences Property, therefore in subsequent step 103, can between calling sequence predictive coding pattern, so as to utilize the first image sequence and second Correlation between image sequence promotes data compression rate.

Described first image sequence and the second coding image sequences are generated three-dimensional video-frequency bit stream by step 103.

This step generates three-dimensional video-frequency bit stream to the first image sequence and the second coding image sequences.It is carried in the present embodiment In the technical solution of confession, since the first image sequence is as the primary image referred in an encoding process for the second image sequence Sequence, therefore usually can start cataloged procedure prior to the second image sequence.

To the first image sequence, industrial tradition DPCM coding standards may be employed and encoded, for example, using ITU.264/ The coding standards such as AVC, ISO MPEG-H HEVC are encoded.

Requirement of the video encoding standard to coded sequence be typically：First picture frame is usually to belong to intraframe coding type Picture frame, in interleave the picture frames of several interframe encode types, after immediately another intraframe coding type picture frame.Using This mode can provide random access demand, to belonging to the picture frame of intraframe coding type, not depend on other picture frame contents I.e. decodable code generates this image frame signal；And the coding compression efficiency of interframe encode type is higher than intraframe coding type.Two An image sets are formed between the picture frame of a intraframe coding type.For the picture frame of intraframe coding type, generally use frame Intraprediction encoding pattern encodes block of pixels therein, for the picture frame of interframe encode type, then may be employed in frame Predictive coding pattern or inter prediction encoding pattern encode block of pixels therein.

This step encodes the first image sequence using aforesaid way, this part processing procedure and the prior art regard Frequency coding mode is identical, and details are not described herein again.

In the cataloged procedure to the second image sequence, on the basis for the various predictive coding patterns that the prior art provides On, also introduce inter-sequence prediction coding mode.The inter-sequence prediction coding mode refers to, with the figure in the first image sequence As carrying out predictive coding for reference frame, to the block of pixels in the second image sequence.Below to the cataloged procedure of the second image sequence It elaborates.

The type of coding of every two field picture in second image sequence may be employed the mode identical with the first image sequence into Row is set, then can be according to the type of coding wherein per two field picture (in frame during being encoded to the second image sequence Type of coding or interframe encode type), according to corresponding order to every two field picture (i.e.：Image to be encoded described below) It is encoded, the coding mode that when coding uses can include inter-sequence prediction coding mode.

The processing procedure encoded to the image to be encoded in the second image sequence comprises the following steps 103-1 to 103- 4, it is described in detail with reference to Fig. 4.

Step 103-1, according to the pixel block size set according to predetermined manner, image to be encoded is divided into and multiple waits to compile The block of pixels of code.

The pixel block size usually represents that n can be preset fixed numbers with n × n, such as：N=64.

Preferably, the present embodiment provides according to the corresponding image analytic degree parameter of sequence of video images, block of pixels ruler is set Very little preferred embodiment.I.e.：Pixel block size n × n can be based on the image analytic degree ginseng obtained in step 101 in the present embodiment Number is made adaptive dynamic and is adjusted.Such as：When image analytic degree meets W × H≤640 × 480, n=32 can be set；Work as figure When meeting W × H >=3840 × 2160 as resolution, n=128 can be set；When W × H value 640 × 480 and 3840 × When between 2160, n=64 can be set.It, can be with the video figure of efficient process difference resolution using this preferred embodiment As sequence or even super-resolution degree, such as 4K, 8K.

Step 103-2, judge whether need encoded pixels block in image to be encoded, if so, step 103-3 is performed, Otherwise, the coding of coded image is treated in end.

Usually according to order from left to right, from top to bottom, the block of pixels to be encoded treated successively in coded image carries out Coding.This step judges whether not carry out the block of pixels of coded treatment also in the order described above, if so, then performing step 103- 3, otherwise terminate to treat the coding of coded image.

Step 103-3, according to the type of coding of the image to be encoded, from including the corresponding of inter-sequence prediction coding mode It is the coding mode that block of pixels selection to be encoded meets preset condition in coding mode set.Wherein, it is described to meet preset condition Coding mode include：Meet the coding mode of the rate-distortion optimization model based on cost minimization.

In the following, the type of coding from image to be encoded is：Two aspects of intraframe coding type or interframe encode type, to choosing The processing procedure for selecting coding mode illustrates.

(1) image to be encoded belongs to intraframe coding type.

For intraframe coding type, corresponding coding mode set includes：Intra-prediction code mode, inter-sequence prediction Coding mode.

The present embodiment devises a kind of rate-distortion optimization model based on cost minimization, this mould for intraframe coding type On the one hand type will consider the contribution of inter-sequence prediction coding mode, still further aspect consider conventional intra prediction coding mode (with Lower abbreviation intra-prediction code mode) contribution, the optimization effect of selection rate distortion from the two, i.e.,：Choose Coding cost Minimum coding mode.

For intraframe coding type, inter-sequence prediction coding mode be using the first associated images in the first image sequence as Reference frame is encoded, and first associated images refer to the image that temporal information is identical with image to be encoded or is close, For example, it may be there is the image of same time information with image to be encoded.Since image to be encoded and the first associated images are Based on the image that Same Scene is shot at the same or similar moment, therefore body region, is utilized there are larger correlation This characteristic helps to eliminate the signal redundancy between image, so as to improve compression ratio.

Therefore, based on the rate-distortion model of cost minimization as shown in equation 4 below, represent for intraframe coding class Type, the result RDO of rate-distortion optimization_IntraIt is intraframe predictive coding cost Cost_anchorWith inter-sequence prediction Coding cost Cost_inter-viewIn smaller value.

RDO_Intra=Min (Cost_anchor,Cost_inter-view)；--- ----formula 4

The process of coding mode is selected including following 1) to 3) for block of pixels to be encoded based on above-mentioned model：

1) the first cost encoded using intra-prediction code mode to the block of pixels to be encoded is calculated.

Calculate the first cost (i.e.：Cost_anchor) factor include：Residual error data and use intraframe predictive coding introduce Code check etc..The calculating process is same as the prior art, and details are not described herein again.

2) calculate using the first associated images in the first image sequence as reference frame, using inter-sequence prediction coding mode pair The second cost that the block of pixels to be encoded is encoded.

Second cost is calculated (i.e. using equation 5 below：Cost_inter-view)：

Cost_inter-view=SAD+ λ_motion×Bits_mv；--- -- formula 5

Second cost and absolute error and (Sum of Absolute Differences, be abbreviated as SAD), λ_motionWith Bits_mvIt is related.Wherein, SAD represents block of pixels to be encoded and in reference frame, i.e.,：In first associated images, that searches for With the distortion size between block of pixels, i.e.,：The value of corresponding pixel points seeks the absolute value summation after difference between block, and calculation formula is for example following Shown in formula 6, wherein Cur_ijIt is the pixel value of pixel (i, j) in block of pixels to be encoded, Ref_ijIt is pixel in matched pixel block The pixel value of point (i, j).λ_motionIt for corresponding LaGrange parameter, is provided by formula 7 and formula 8, wherein QP joins for quantization Number.Bits_mvRepresent motion vector (Motion Vector) occupied digit.

It should be noted that the present embodiment is described to search for matched pixel block or similar description in reference frame Sentence typically refers to search for matched pixel block in the reconstructed image after encoding for reference frame.Hereafter repeat no more.

3) by the corresponding coding mode of minimum value in first cost and the second cost, selected as block of pixels to be encoded The coding mode selected.

1) and 2) minimum value is selected from the first cost and the second cost calculated, and by the corresponding coding mode of minimum value As the coding mode selected for block of pixels to be encoded.That is, the coding mode being finally selected, be generate less residual error data and The coding mode of code check.

(2) image to be encoded belongs to interframe encode type.

For interframe encode type, corresponding coding mode set includes：Intra-prediction code mode, inter-prediction are compiled Pattern and inter-sequence prediction coding mode.

It is similar with (one), for interframe encode type, the rate-distortion model based on cost minimization of the present embodiment design As shown in equation 9 below, wherein, Cost_anchorIt is the Least-cost value using intraframe predictive coding and inter prediction encoding, Cost_inter-viewIt is the cost for the inter-sequence prediction coding that the present embodiment technical solution introduces.

RDO_Inter=Min (Cost_anchor,Cost_inter-view)；--- -- formula 9

The process of coding mode is selected including following 1) to 4) for block of pixels to be encoded based on above-mentioned model：

1) third generation valency encoded using intra-prediction code mode to the block of pixels to be encoded is calculated.

2) the forth generation valency encoded using inter prediction encoding pattern to the block of pixels to be encoded is calculated.

1) and 2) above-mentioned processing procedure is calculated respectively using in frame and inter prediction encoding pattern generates respectively the 3rd Cost and forth generation valency, the processing procedure belong to the prior art, and details are not described herein again.

3) calculate using the second associated images in the first image sequence as reference frame, using inter-sequence prediction coding mode pair The 5th cost that the block of pixels to be encoded is encoded.

Second associated images include：There is image, the temporal information of same time information with the image to be encoded The image or temporal information for being not later than image to be encoded are later than earlier than the image and temporal information of the image to be encoded The image of the image to be encoded.

It is similar with inter prediction encoding of the prior art to calculate the process of the 5th cost, difference, which essentially consists in reference frame, is It is chosen from the first image sequence.

In the specific implementation, the mode of single directional prediction may be employed.Correspondingly, the second associated images can include：Have The image of same time information, the calculation of the 5th cost are identical with calculation shown in formula 5 in (one)；Second associated diagram As that can also include：Temporal information is not later than the image of image to be encoded, such as：The corresponding temporal information of image to be encoded is T4, can be selected when calculating five costs temporal information in the first image sequence be not later than t4, such as temporal information be respectively t3, 4 images of t2, t1 and t0 are respectively adopted the calculation of above-mentioned formula 5, calculate the Coding cost of block of pixels to be encoded, and Therefrom selecting minimum value, the corresponding image of minimum value is corresponding single directional prediction reference frame as the 5th cost.

It in the specific implementation, can also be by the way of bi-directional predicted.Correspondingly, the second associated images are at least two frames, It can include：Temporal information is later than the figure of the image to be encoded earlier than the image and temporal information of the image to be encoded Picture.In this case, when calculating five costs using the calculation shown in above-mentioned formula 5, weighted calculation may be employed in SAD As a result, Bits_mvFor the length of corresponding sports vector.Such as：The corresponding temporal information of image to be encoded is t4, selects the first image The image at t0 and t8 moment is used as bi-directional predicted with reference to frame progress in sequence, respectively obtains two sad values and two movement arrows Amount, then, can be using the result of described two SAD weighted sums as formula 5 when calculating five costs using above-mentioned formula 5 In SAD, using the occupied total bit of described two motion vectors as the Bits in formula 5_mv, so as to which sequence be calculated Between predictive coding the 5th cost.

4) by the corresponding coding mode of minimum value in the three, the four, the 5th cost, as the pixel to be encoded The coding mode of block selection.

From the third generation valency 1), 2) He 3) calculated, select minimum value in forth generation valency and the 5th cost, and by minimum value pair The coding mode answered is as the coding mode selected for block of pixels to be encoded.That is, the coding mode being finally selected is to generate more The coding mode of few residual error data and code check.

In embodiments illustrated above, predictive coding pattern intraframe coding type calling sequence, by being based on Associated images in one image sequence carry out motion prediction (i.e.：Search for matched pixel block), it can effectively reduce residual error data Amount of coded data plays the role of compressed data；Predictive coding pattern interframe encode type calling sequence, can equally subtract The amount of coded data of few residual error data, it is particularly more frequent in scene conversion ratio, alternatively, under grand movement background, between sequence Predictive coding can receive gain at double compared with prior art so that data compression rate is obviously improved.

Step 103-4, the block of pixels to be encoded is encoded using selected coding mode, then goes to step 103- 2 perform.

The process that encoded pixels block is encoded is treated, mainly residual error data is encoded, with amount of compressed data, and Coding result is write in three-dimensional video-frequency bit stream.In the specific implementation, can also by corresponding Coding and description information, such as：It compiles The information such as pattern, reference frame index, motion vector are write after being encoded according to default mode in three-dimensional video-frequency bit stream, for It is used during decoding.After completing to the coded treatment of the block of pixels to be encoded, step 103-2 execution is gone to, until will be to be encoded All block of pixels codings to be encoded in image finish.

More than, it is that the block of pixels in image to be encoded selects corresponding encoded after predictive coding pattern between calling sequence The pattern and process encoded is described.

Preferably, the process of the second cost or the 5th cost is being calculated for the block of pixels to be encoded in image to be encoded In, the matched pixel block of the block of pixels to be encoded is searched in the reconstructed image of the first associated images or the second associated images When, according in for the image encoding process to be encoded, the head that search matched pixel block obtains in the reconstructed image The coordinate of a motion vector and the block of pixels to be encoded, determines the search origin coordinates in the reconstructed image, and from Described search origin coordinates starts to search for the matched pixel block.Such as：In the image to be encoded progress to intraframe coding type During coding, for the first time in the reconstructed image of the first associated images by search for the motion vector that matched pixel block obtains for (Δ x, Δ y), and current block of pixels to be encoded coordinate (such as：The coordinate value in its upper left corner) it is (x, y), then it is described in order to be directed to The cost of predictive coding between the block of pixels sequence of calculation to be encoded searches for matched pixel block in the reconstructed image of the first associated images When, it can be with (x+ Δs x, y+ Δ y) is scanned for as initiating searches coordinate.In this way, movement can be substantially speeded up The convergence rate of search finds matched pixel block with the time as few as possible.

In addition, the present embodiment also provides the preferred embodiment for merging Coding and description information.Specifically, to the second image During image to be encoded in sequence is encoded, if the reference frame phase there are two or more contiguous pixels blocks With and motion vector it is identical, then the Coding and description information of the contiguous pixels block can be merged into one group of write-in stereopsis In frequency bit stream.Such as：There are continuous 8 block of pixels all to employ inter-sequence prediction coding mode, and motion vector all for (Δ x, Δ y), reference frame index are all 100, then the description information of these block of pixels can be merged into similar " coding mode=sequence Between predictive coding merging patterns, motion vector=(Δ x, Δ y), reference frame index=100, the form of continuous block number=8 ", and It writes in three-dimensional video-frequency bit stream.In this way, the compression ratio of stereo scopic video coding can be further improved.

It should be noted that in the specific implementation, it, can be by right during being encoded to the first image sequence The decoding and reconstituting process of encoded image exports the decoded picture buffer list (Decoded based on the first image sequence Picture Buffer List, DPB queue), that is, reconstructed image queue after the coding of the first image sequence, for the second figure It is used as sequence carries out inter-sequence prediction coding, carries out corresponding motion prediction and motion compensation calculations.

So far, the cataloged procedure of the first image sequence and the second image sequence is described respectively.It is being embodied When, according to the needs of practical application scene, it can be finished in the first coding image sequences and restart the coding of the second image sequence Process, can also in the first coding image sequences to a certain extent, such as：After completing the coding of default number of images, start the The cataloged procedure of two image sequences, it is also possible.

Correspondingly, on the basis of being encoded to the first image sequence and the second image sequence, not similar shape can be generated The three-dimensional video-frequency bit stream of formula.It for example, can be by the bit stream generated for the first coding image sequences and for the second image sequence The bit stream of coding generation is sequentially output, forms three-dimensional video-frequency bit stream；The position that can also will be generated for the first video sequence coding Stream is with being directed to the bit stream of the second coding image sequences generation, interweave generation three-dimensional video-frequency bit stream according to predetermined manner.Perform decoding One root of operation performs corresponding decoding operate according to the form of the three-dimensional video-frequency bit stream, it is possible to reduction carrying three-dimensional video-frequency Two-path video image sequence.

It wherein, will be for the bit stream of the first image sequence example coding generation and the bit stream for the generation of the second coding image sequences The mode to form three-dimensional video-frequency bit stream is sequentially output, it, can be only for the Video Decoder for only supporting traditional decoding process The bit stream of first image sequence is decoded, is included in the first image sequence corresponding to the video image of left eye or right eye In the case of independent texture information, then two-dimensional video image output can be obtained, is realized compatible with the prior art.

In addition, in specific implementation process, for the ease of performing a side of decoding operate correctly to three-dimensional video-frequency bit stream Decoding, relevant parameter can be consulted in advance by implementing the both sides of coding and decoding operation, including：Image analytic degree, color mode, Pixel block size N, can also include down-sampled coefficient or stereo scopic video coding pattern relevant with sequence of subsampled images, The parameters such as weight α write three-dimensional video-frequency after above-mentioned parameter information can also be encoded according to the form that both sides arrange by coding staff It is used in bit stream, for decoding side.

So far, by above-mentioned steps 101-103, the embodiment of method provided in this embodiment is described. When it is implemented, the three-dimensional video-frequency bit stream of coding generation can be write in storage medium, such as：It is stored in the form of a file hard On disk；Or the recipient of three-dimensional video-frequency bit stream is sent to by network, plays three-dimensional video-frequency so that recipient decodes.

It should be noted that implementations described above is performed for two sequence of video images that step 101 obtains Encoding operation, in other embodiments, N can be the integer more than 2, i.e.,：2 or more can be obtained in step 101 to regard Frequency image sequence in this case, still may be referred to the image in the first selected image sequence, to other N-1 the Image in two image sequences is encoded, and can equally realize the technical solution of the application, and obtains corresponding advantageous effect.

In conclusion method provided in this embodiment, due to considering the correlation of stereo video signals, based on video After image sequence determines the first image sequence and the second image sequence, during being encoded to the second image sequence, draw Inter-sequence prediction coding mode is entered, therefore can have been further improved under conditions of equal quality three-dimensional video-frequency bit stream is obtained Compression ratio reduces the data volume of the three-dimensional video-frequency bit stream of generation, so as to reduce to the occupancy of memory space or reduction pair The occupancy of network bandwidth, avoid the occurrence of due to caused by network congestion packet loss the problems such as three-dimensional video-frequency playing process interim card, effectively Promote the application experience that three-dimensional video-frequency plays product.

In the above-described embodiment, provide a method, corresponding, the application also provides a kind of device.Please Referring to Fig. 5, for a kind of schematic diagram of the embodiment of device of the application.Implement since device embodiment is substantially similar to method Example, so describing fairly simple, the relevent part can refer to the partial explaination of embodiments of method.Device described below is implemented Example is only schematical.

The device of the present embodiment, including：Sequence of video images acquiring unit 501, for obtaining the N number of of carrying three-dimensional video-frequency Sequence of video images, N are the integer more than or equal to 2；Image sequence determination unit 502, for being based on N number of video image sequence Row determine the first image sequence and N-1 the second image sequences；Coding image sequences unit 503, for by described first image Sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；Wherein, encoded and adopted for the second image sequence Coding mode includes：Inter-sequence prediction coding mode；The inter-sequence prediction coding mode refers to, with described first image Image in sequence carries out predictive coding for reference frame, to the block of pixels in second image sequence.

Optionally, the N be 2, the sequence of video images acquiring unit, specifically for obtain correspond respectively to left eye and Two sequence of video images of right eye.

Optionally, described image sequence determination unit is regarded specifically for selecting one from N number of sequence of video images Frequency image sequence is as described first image sequence, using other each sequence of video images as the N-1 the second images Each sequence in sequence.

Optionally, described image sequence determination unit includes：

Down-sampled subelement, for according to default down-sampled mode, to every two field picture in each sequence of video images It carries out respectively down-sampled；

Sequence selection subelement, for selecting a sequence of video images in N number of sequence of video images after down-sampled As described first image sequence, using other each sequence of video images after down-sampled as the N-1 the second images Each sequence in sequence.

Optionally, described image sequence determination unit includes：

Sub-sampling sequence generates subelement, for according to default pretreatment coding mode, according to N number of video image Sequence generates sequence of subsampled images；

Sub-sampling sequence splits subelement, for pre-processing the corresponding fractionation mode of coding mode according to described, by described in Every two field picture in sequence of subsampled images splits into N number of image with same time information, so as to obtain N number of image sequence Row；

Image sequence selects subelement, for from N number of image sequence, selects the first image sequence, and by other Image sequence is as the N-1 the second image sequences.

Optionally, described image sequential coding unit includes：

First coding image sequences subelement, for being encoded to the first image sequence；

Second coding image sequences subelement, for being encoded to N-1 the second image sequences；

The second coding image sequences subelement, including：

Block of pixels divides subelement, for according to the pixel block size set according to predetermined manner, image to be encoded to be drawn It is divided into multiple block of pixels to be encoded；

Loop control subelement for being directed to each block of pixels to be encoded, calls following model selection subelement successively Coded treatment is carried out with block of pixels coded sub-units；

Model selection subelement for the type of coding according to the image to be encoded, is encoded from including inter-sequence prediction It is the coding mode that block of pixels selection to be encoded meets preset condition in the corresponding encoded set of modes of pattern；

Block of pixels coded sub-units, for being encoded using selected coding mode to the block of pixels to be encoded.

Optionally, the model selection subelement, specifically for the type of coding according to the image to be encoded, from including It is that block of pixels selection to be encoded meets based on cost minimization in the corresponding encoded set of modes of inter-sequence prediction coding mode The coding mode of rate-distortion optimization model.

Optionally, the model selection subelement includes the following subelement for intraframe coding type：

First cost computation subunit carries out the block of pixels to be encoded using intra-prediction code mode for calculating First cost of coding；

Second cost computation subunit for calculating using the first associated images in the first image sequence as reference frame, is adopted The second cost encoded with inter-sequence prediction coding mode to the block of pixels to be encoded；

First mode selects subelement, for by the corresponding coding mould of minimum value in first cost and the second cost Formula, as the coding mode for the block of pixels selection to be encoded.

Optionally, the model selection subelement includes the following subelement for interframe encode type：

Third generation valency computation subunit carries out the block of pixels to be encoded using intra-prediction code mode for calculating The third generation valency of coding；

Forth generation valency computation subunit carries out the block of pixels to be encoded using inter prediction encoding pattern for calculating The forth generation valency of coding；

5th cost computation subunit for calculating using the second associated images in the first image sequence as reference frame, is adopted The 5th cost encoded with inter-sequence prediction coding mode to the block of pixels to be encoded；

Second mode selects subelement, for by the corresponding coding mould of minimum value in the three, the four, the 5th cost Formula, as the coding mode for the block of pixels selection to be encoded.

Optionally, the block of pixels division subelement is specifically used for, according to the corresponding image solution of the sequence of video images Analysis degree parameter sets the pixel block size, and image to be encoded is divided into multiple block of pixels according to the pixel block size.

Optionally, the second cost computation subunit or the 5th cost computation subunit, in the first associated images or When the matched pixel block of the block of pixels to be encoded is searched in the reconstructed image of two associated images, according to for described to be encoded The obtained first motion vector of matched pixel block is searched in image encoding process, in the reconstructed image and described waits to compile The coordinate of code block of pixels, determines the search origin coordinates in the reconstructed image, and is searched since described search origin coordinates Suo Suoshu matched pixel blocks.

Optionally, the second image coded sub-units are during coding, if there are two or more companies The reference frame of continuous block of pixels is identical and motion vector is identical, then the Coding and description information of the contiguous pixels block is merged into one group It writes in the three-dimensional video-frequency bit stream.

Optionally, described device further includes：

Storage unit, the three-dimensional video-frequency bit stream for described image sequential coding unit to be generated are write in storage medium； Alternatively,

Transmitting element, the three-dimensional video-frequency bit stream for described image sequential coding unit to be generated are sent to recipient.

Corresponding with a kind of method that the application provides, the application also provides another method.Fig. 6 is refer to, for this Apply for the flow chart of the embodiment of another method provided, the present embodiment part identical with above method embodiment step is not Repeat again, below emphasis difference is described.Another method that the application provides includes：

Step 601 obtains three-dimensional video-frequency bit stream to be decoded.

This step may be employed different modes and obtain three-dimensional video-frequency bit stream to be decoded, including can be from storage medium Middle reading three-dimensional video-frequency bit stream, such as：It is read from the file for storing three-dimensional video-frequency bit stream；Can also by network receive by The three-dimensional video-frequency bit stream that sender sends.

Step 602 obtains the first image sequence and the second image sequence by decoding from the three-dimensional video-frequency bit stream.

In the case of the coding stream of the first image sequence and the second image sequence is included in three-dimensional video-frequency bit stream successively, The decoding operate for the first image sequence can be first carried out, decoding is finished or decoded to a certain extent, starts for second The decoding operate of image sequence；If three-dimensional video-frequency bit stream is the bit stream of intertexture form, can first be separated according to predetermined manner For the first image sequence and the coding stream of the second image sequence, i.e. isolate the first image sequence from three-dimensional video-frequency bit stream Then row bit stream and the second image sequence bit stream perform decoding operate according still further to aforesaid way.

Wherein, the first image sequence bit stream is standard bit streams, prior art decoding may be employed, details are not described herein again.Pin To the second image sequence bit stream, then can first determine with obtaining the relevant parameters of sequence of video images, the parameters Can be it is that encoding and decoding both sides consult in advance or according to carried in three-dimensional video-frequency bit stream parameter information decoding after It arrives.

Then using the parameter information, each image in the second image sequence is generated by following decoding process：Solution The bit stream for corresponding to described image in the second image sequence bit stream is analysed, the corresponding coding of wherein each block of pixels is obtained by decoding Pattern and residual error data, then according to the coding mode of each block of pixels and residual error data, rebuild respective pixel block, and with reconstruction Block of pixels synthesis described image, such as according to from left to right, from top to bottom order synthesize described image.

Wherein, it is the block of pixels of inter-sequence prediction coding mode for coding mode, it is also necessary to the ginseng obtained according to decoding Frame index information is examined, is used as using the respective image that output has been decoded in the first image sequence and is rebuild with reference to frame.Specifically, When reconstruction belongs to the block of pixels of intraframe coding type, used reference frame includes：In the first image sequence and pixel The affiliated image of block has the image of same time information；When reconstruction belongs to the block of pixels of interframe encode type, used reference Frame include the first image sequence in hypograph：With the affiliated image of the block of pixels have same time information image, when Between information be not later than the affiliated image of the block of pixels image or temporal information earlier than the affiliated image of the block of pixels figure Picture and temporal information are later than the image of the affiliated image of the block of pixels.

By above-mentioned processing procedure, the first image sequence and the second image sequence are just generated.

Step 603 obtains the two of carrying three-dimensional video-frequency according to the first image sequence and the second image sequence of the acquisition A sequence of video images.

Mode is corresponding used by determining two image sequences according to two sequence of video images during with encoding, this step Corresponding manner, the first image sequence exported according to step 602 and the second image sequence can also be used to obtain carrying stereopsis Two sequence of video images of frequency.It is exemplified below several ways.

The first image sequence and the second image sequence that step 602 can be exported are as two sequence of video images In each sequence；It can also be according to down-sampled mode used by coding, to described first image sequence and second figure Sampling processing is risen as every two field picture progress in sequence is corresponding, then by the first image sequence and the second image after liter sampling Sequence is respectively as each sequence in two sequence of video images；For coding when using the conjunction based on sequence of subsampled images And for fractionation mode, this step then can by elder generation according to coding when used fractionation in a manner of, for described first image sequence Corresponding union operation is performed with image in the second image sequence, that there is same time information, obtains sub-sampled images sequence The sequence of subsampled images then according to used stereo scopic video coding pattern during coding, is accordingly divided into two by row Image sequence to get to it is described carrying three-dimensional video-frequency two sequence of video images.

Two sequence of video images obtained by above-mentioned processing procedure, typically respectively regarding corresponding to left eye and right eye Frequency image sequence.The two sequence of video images can be sent to stereo video display and show, by the parallax on binocular sense organ, So that beholder gets longitudinal depth information of scene and experiences three-dimensional sense.

So far, by above-mentioned steps 601-603, the embodiment of another method of the application offer is retouched It states.The first image sequence and second image sequence are obtained by decoding in the above-described embodiment, for three-dimensional video-frequency It, can also be in the same way to corresponding bit stream in the case that bit stream includes two or more second image sequences It is decoded, to obtain each second image sequence, and finally obtains the sequence of video images of respective number.

By foregoing description as can be seen that a kind of described method of embodiment of the method provided before using the application It on the basis of being encoded, is decoded using method provided in this embodiment, enables to the three-dimensional video-frequency after coding by just True ground decoded back, so as to while stereo video data compression ratio is promoted, ensure the normal play of three-dimensional video-frequency.

It is corresponding foregoing provide the embodiment of another method of the application, the another of the application is provided below A kind of embodiment of device.Fig. 7 is refer to, is the schematic diagram of the embodiment of another device of the application.Since device is real It applies example and is substantially similar to embodiment of the method, so describing fairly simple, referring to the part explanation of embodiment of the method in place of correlation .Device embodiment described below is only schematical.

The device of the present embodiment, including：Three-dimensional video-frequency bit stream acquiring unit 701, for obtaining three-dimensional video-frequency to be decoded Bit stream；Three-dimensional video-frequency bit stream decoding unit 702 obtains the first image sequence for passing through decoding from the three-dimensional video-frequency bit stream With N-1 the second image sequences；Wherein, each image in the second image sequence is obtained by following decoding process：Root Belong to according to the corresponding encoded pattern entrained by the second image sequence bit stream in the three-dimensional video-frequency bit stream and residual error data, reconstruction Each block of pixels of described image, and synthesize described image with the block of pixels rebuild；What the second image sequence bit stream carried Coding mode includes：Inter-sequence prediction coding mode；Sequence of video images generation unit 703, for according to the acquisition One image sequence and N-1 the second image sequences obtain N number of sequence of video images of carrying three-dimensional video-frequency.

Optionally, the N is 2；The three-dimensional video-frequency bit stream decoding unit, specifically for passing through decoding from the stereopsis The first image sequence and second image sequence are obtained in frequency bit stream；The sequence of video images generation unit, is specifically used for Two sequence of video images of carrying three-dimensional video-frequency are obtained according to the first image sequence of the acquisition and the second image sequence.

Optionally, the sequence of video images generation unit, specifically for described first image sequence and the N-1 is a Second image sequence is respectively as each sequence in N number of sequence of video images.

Optionally, the sequence of video images generation unit, including：

Sampling subelement is risen, for according to down-sampled mode used by coding, to described first image sequence and institute It states every two field picture in the second image sequence and carries out corresponding liter sampling processing；

Image sequence generates subelement, for that will rise the first image sequence after sampling and N-1 the second image sequences point Not as each sequence in N number of sequence of video images.

Optionally, the sequence of video images generation unit, including：

Sub-sampling sequence also atomic unit, for according to coding when used fractionation mode, for described first image Image in sequence and the second image sequence, having same time information performs corresponding union operation, obtains sub-sampling figure As sequence；

Sequence of video images divide subelement, for according to coding when used stereo scopic video coding pattern, will described in Sequence of subsampled images be accordingly divided into N number of image sequence to get to it is described carrying three-dimensional video-frequency N number of sequence of video images.

Optionally, the three-dimensional video-frequency bit stream decoding unit belongs to intraframe coding according to the reconstruction of inter-sequence prediction coding mode During the block of pixels of type, used reference frame includes：It is in first image sequence, with the affiliated image of the block of pixels have phase With the image of temporal information.

Optionally, the three-dimensional video-frequency bit stream decoding unit belongs to interframe encode according to the reconstruction of inter-sequence prediction coding mode During the block of pixels of type, used reference frame include the first image sequence in hypograph：

There is the image of same time information with the affiliated image of the block of pixels；Alternatively,

Temporal information is not later than the image of the affiliated image of the block of pixels；Alternatively,

Temporal information is later than figure belonging to the block of pixels earlier than the image and temporal information of the affiliated image of the block of pixels The image of picture.

Optionally, the three-dimensional video-frequency bit stream acquiring unit, specifically for reading three-dimensional video-frequency bit stream from storage medium, Alternatively, receive the three-dimensional video-frequency bit stream sent by sender.

Fig. 8 is referred to, for a kind of schematic diagram for exemplary system that the application provides.As shown in figure 8, as shown in figure 8, it is System 800 include the devices 801 (in the present embodiment be known as stereo scopic video coding device) that a kind of above-mentioned device embodiment provided, And the device 802 (being known as three-dimensional video-frequency decoding apparatus in the present embodiment) that above-mentioned another device embodiment is provided.

The stereo scopic video coding device 801 includes：Sequence of video images acquiring unit 801-1, image sequence determine list First 801-2 and coding image sequences unit 801-3, the function of wherein each unit refer to a kind of device embodiment provided before In explanation, details are not described herein again.The three-dimensional video-frequency decoding apparatus 802 includes：Three-dimensional video-frequency bit stream acquiring unit 802-1, Three-dimensional video-frequency bit stream decoding unit 802-2 and sequence of video images generation unit 802-3, the function of wherein each unit refer to it Explanation in another device embodiment of preceding offer, details are not described herein again.

In the specific implementation, the stereo scopic video coding device 801 and the three-dimensional video-frequency decoding apparatus 802 can be distinguished It is deployed on different electronic equipments, the electronic equipment includes PC or mobile computing device etc., the mobile computing Equipment can include but is not limited to：Laptop, tablet computer, mobile phone, and/or other smart machines etc..It is three-dimensional Video coding apparatus 801, can be by sequence of video images that photographic device is shot with video-corder or more comprising three-dimensional video-frequency by reading The sequence of video images that media resource file and transcoding obtain, encoded operation generation three-dimensional video-frequency bit stream, and pass through network It sends, wherein, the coding mode used in cataloged procedure includes：Inter-sequence prediction coding mode；Three-dimensional video-frequency decoding dress Three-dimensional video-frequency bit stream can be received from network and perform corresponding decoding operate by putting 802, so as to reduce carrying three-dimensional video-frequency Sequence of video images, so that corresponding stereo video display is shown.Since stereo scopic video coding device 801 is encoded Inter-sequence prediction coding mode is employed in journey, therefore can effectively promote stereo video data compression ratio, is reduced to Netowrk tape Wide occupancy, the situation for avoiding the occurrence of network congestion packet loss are smooth so as to be obtained for three-dimensional video-frequency decoder 802 by decoding Three-dimensional video-frequency provides safeguard.

In the specific implementation, the stereo scopic video coding device 801 and the three-dimensional video-frequency decoding apparatus 802 can also portions It is deployed on same electronic equipment, the electronic equipment includes PC or mobile computing device etc., and the mobile computing is set It is standby to include but is not limited to：Laptop, tablet computer, mobile phone, and/or other smart machines etc..Stereopsis Frequency code device 801 can write the three-dimensional video-frequency bit stream for employing the generation of inter-sequence prediction coding mode in storage medium, For example, it is stored in the form of a file on hard disk；When needing to play three-dimensional video-frequency, three-dimensional video-frequency decoding apparatus 802 then can be with Three-dimensional video-frequency bit stream is read from storage medium, such as：It reads, and passes through from the fixed disk file for storing three-dimensional video-frequency bit stream The sequence of video images of corresponding decoding operate reduction carrying three-dimensional video-frequency, for playing.Due to stereo scopic video coding device 801 Inter-sequence prediction coding mode is employed in an encoding process, therefore can effectively promote stereo video data compression ratio, is reduced To the occupancy of storage medium, memory space is saved.

In addition, present invention also provides a kind of embodiments of system, Fig. 9 is refer to, it illustrates one that the application provides The schematic diagram of kind system embodiment.

Wherein, system 900 can include：Processor 901, the system control unit 902 coupled with the processor, with institute The Installed System Memory (System Memory) 903 of system control unit coupling is stated, is coupled with the system control unit non-volatile Memory (NonVolatile Memory-NVM) or storage device 904 and the network coupled with the system control unit Interface 905.

The processor 901 can include at least one processor, and each processor can be single core processor or more Core processor.The processor 901 can include general processor and application specific processor (for example, graphics processor, using processing Device, baseband processor etc.) any combination.

The system control unit 902 can include any corresponding interface controller, be the processor 901 in extremely A few processor and/or, any equipment to communicate with the system control unit 902 or component provide interface.

The system control unit 902 can include at least one memory control that interface is provided for the Installed System Memory 903 Device processed.The Installed System Memory 903 can be used for loading and store data and/or instruction.The Installed System Memory 903 can include appointing What volatile memory, for example, dynamic random access memory (dynamic random access memory-DRAM).

The nonvolatile storage or storage device 904 can including at least one tangible, permanent computer Medium is read, the computer-readable medium is used to store data and/or instruction.The nonvolatile storage or storage device 904 Can include any type of nonvolatile storage, for example, flash memory (flash memory) and/or, any non-volatile memory is set It is standby, for example, at least one hard disk drive (hard disk drive-HDD), at least one CD drive, and/or extremely Few digital versatile disc (digital versatile disk-DVD) driver.

The Installed System Memory 903 and the nonvolatile storage or storage device 904 can be stored with instruction 907 respectively Temporary copy and persistence copy.

When the instruction in described instruction 907 is performed by least one processor in processor 901, can cause described System 900 performs following operation：N number of sequence of video images of carrying three-dimensional video-frequency is obtained, N is the integer more than or equal to 2；It is based on N number of sequence of video images determines the first image sequence and N-1 the second image sequences；By described first image sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；Wherein, carry out encoding used volume for the second image sequence Pattern includes：Inter-sequence prediction coding mode.

Alternatively, the system 900 is caused to perform following operation：Obtain three-dimensional video-frequency bit stream to be decoded；By decoding from The first image sequence and N-1 the second image sequences are obtained in the three-dimensional video-frequency bit stream；According to the first image of the acquisition Sequence and N-1 the second image sequences obtain N number of sequence of video images of carrying three-dimensional video-frequency；Wherein, in the second image sequence Each image obtained by following decoding process：The second image sequence bit stream institute in the three-dimensional video-frequency bit stream The corresponding encoded pattern and residual error data of carrying, reconstruction belong to each block of pixels of described image, and are closed with the block of pixels rebuild Into described image；The coding mode that the second image sequence bit stream carries includes：Inter-sequence prediction coding mode.

The network interface 905 can include the transceiver that wireless interface is provided for the system 900, the system 900 Can the communication of across a network be realized by wireless interface and/or communicated with other equipment.The network interface 905 can include Any hardware and/or firmware.The network interface 905 can include providing multi input, the mutiple antennas of multi output wireless interface. In the specific implementation, the network interface 905 can be network adapter, wireless network adapter, telephone modem, And/or radio modem.

In the specific implementation, at least one processor in the processor 901 can be with the system control unit 902 In the control logic of at least one controller be packaged together.In the specific implementation, it is at least one in the processor 901 Processor can be packaged together with the control logic of at least one controller in the system control unit 902, form system System encapsulation (System in Package-SiP).In the specific implementation, at least one processor in the processor 901 can It is integrated on the same chip with the control logic at least one controller in the system control unit 902.Specific During implementation, at least one processor in the processor 901 can be at least one control in the system control unit 902 The control logic of device processed integrates on the same chip, forms system on chip (System on Chip-SoC).

The system 900 can include input/output (I/O) equipment 906.The input-output apparatus 906 can include The user interface interacted for user and the system 900 and/or the periphery interacted for peripheral assembly and the system 900 Component interface.

In various embodiments, user interface can include but is not limited to：Display (for example, liquid crystal display, Touch-screen display etc.), loud speaker, microphone, at least one photographic device (such as camera, and/or video camera), flash lamp, And keyboard.

In various embodiments, peripheral component interface can include but is not limited to：Nonvolatile memory port, Audio jack and power interface.

In various embodiments, the system 900 can be deployed in the electronics such as PC, mobile computing device and set Standby upper, the mobile computing device can include but is not limited to：Laptop, tablet computer, mobile phone, and/or its His smart machine etc..In various embodiments, the system 900 can include more or fewer components and/or difference Framework.

This specification can include a variety of example embodiments disclosed below.

In example embodiment 1, a kind of method can include：Obtain N number of sequence of video images of carrying three-dimensional video-frequency, N To be more than or equal to 2 integer；Based on N number of sequence of video images, the first image sequence and N-1 the second image sequences are determined Row；By described first image sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；Wherein, for the second figure Coding mode includes used by being encoded as sequence：Inter-sequence prediction coding mode；The inter-sequence prediction coding mode Refer to, predicted using the image in described first image sequence as reference frame, to the block of pixels in second image sequence Coding.

In example embodiment 2, the N described in example embodiment 1 is 2, and N number of sequence of video images is respectively to correspond to The sequence of video images of left eye and right eye.

In example embodiment 3, described in any embodiment of example embodiment 1-2 based on N number of video image sequence Row determine the first image sequence and N-1 the second image sequences, including：One is selected from N number of sequence of video images to regard Frequency image sequence is as described first image sequence, using other each sequence of video images as the N-1 the second images Each sequence in sequence.

In example embodiment 4, described in any embodiment of example embodiment 1-3 based on N number of video image sequence Row determine the first image sequence and N-1 the second image sequences, further include：According to default down-sampled mode, to each video Every two field picture in image sequence carries out down-sampled respectively；It is described that a video figure is selected from N number of sequence of video images As sequence is as described first image sequence, using other each sequence of video images as a second image sequences of the N-1 In each sequence, including：A sequence of video images is selected as described first in N number of sequence of video images after down-sampled Image sequence, using other each sequence of video images after down-sampled as the N-1 the second image sequences in each sequence Row.

In example embodiment 5, described in any embodiment of example embodiment 1-4 based on N number of video image sequence Row determine the first image sequence and N-1 the second image sequences, including：According to default pretreatment coding mode, according to described N number of sequence of video images generates sequence of subsampled images；According to the corresponding fractionation mode of the pretreatment coding mode, by described in Every two field picture in sequence of subsampled images splits into N number of image with same time information, so as to obtain N number of image sequence Row；From N number of image sequence, the first image sequence is selected, and using other image sequences as the N-1 the second images Sequence.

In example embodiment 6, any embodiment of example embodiment 1-5 uses following steps in the second image sequence Image to be encoded encoded：According to the pixel block size set according to predetermined manner, image to be encoded is divided into multiple Block of pixels to be encoded；Successively to each block of pixels to be encoded, following encoding operation is performed：According to the image to be encoded Type of coding is that block of pixels selection to be encoded meets in advance from the corresponding encoded set of modes including inter-sequence prediction coding mode If the coding mode of condition；The block of pixels to be encoded is encoded using selected coding mode.

In example embodiment 7, the coding mode for meeting preset condition described in any embodiment of example embodiment 1-6 Including：Meet the coding mode of the rate-distortion optimization model based on cost minimization.

In example embodiment 8, any embodiment of example embodiment 1-7 is for intraframe coding type, the corresponding volume Pattern set includes：Intra-prediction code mode, inter-sequence prediction coding mode；The coding according to image to be encoded Type is that block of pixels selection to be encoded meets default item from the corresponding encoded set of modes including inter-sequence prediction coding mode The coding mode of part, including：Calculate the first generation encoded using intra-prediction code mode to the block of pixels to be encoded Valency；It calculates using the first associated images in the first image sequence as reference frame, is treated using inter-sequence prediction coding mode to described The second cost that encoded pixels block is encoded；By the corresponding coding mode of minimum value in first cost and the second cost, As the coding mode for the block of pixels selection to be encoded.

In example embodiment 9, the first associated images described in any embodiment of example embodiment 1-8, including：With institute Stating image to be encoded has the image of same time information.

In example embodiment 10, any embodiment of example embodiment 1-9 is described corresponding pre- for interframe encode type If coding mode set includes：Intra-prediction code mode, inter prediction encoding pattern and inter-sequence prediction coding mould Formula；The type of coding according to image to be encoded, from the corresponding encoded set of modes including inter-sequence prediction coding mode Meet the coding mode of preset condition for block of pixels to be encoded selection, including：It calculates using intra-prediction code mode to described The third generation valency that block of pixels to be encoded is encoded；It calculates and the block of pixels to be encoded is carried out using inter prediction encoding pattern The forth generation valency of coding；It calculates using the second associated images in the first image sequence as reference frame, is encoded using inter-sequence prediction The 5th cost that pattern encodes the block of pixels to be encoded；Minimum value in three, the four, the 5th cost is corresponded to Coding mode, as the coding mode for the block of pixels to be encoded selection.

In example embodiment 11, the second associated images described in any embodiment of example embodiment 1-10 include：With The image to be encoded has the image of same time information；Alternatively, temporal information is not later than the image of the image to be encoded； Alternatively, temporal information is later than the image of the image to be encoded earlier than the image and temporal information of the image to be encoded.

In example embodiment 12, pixel is set according to predetermined manner described in any embodiment of example embodiment 1-11 Block size, including：According to the corresponding image analytic degree parameter of the sequence of video images, the pixel block size is set.

In example embodiment 13, any embodiment of example embodiment 1-12 is calculating the second cost or the 5th cost During, the matching picture of the block of pixels to be encoded is searched in the reconstructed image of the first associated images or the second associated images During plain block, obtained according to matched pixel block is searched in for the image encoding process to be encoded, in the reconstructed image First motion vector and the block of pixels to be encoded coordinate, determine the search origin coordinates in the reconstructed image, And the matched pixel block is searched for since described search origin coordinates.

In example embodiment 14, any embodiment of example embodiment 1-13 is to be encoded in the second image sequence During image is encoded, if the reference frame there are two or more contiguous pixels blocks is identical and motion vector phase Together, then the Coding and description information of the contiguous pixels block is merged into one group of write-in three-dimensional video-frequency bit stream.

In example embodiment 15, any embodiment of example embodiment 1-14 is by described first image sequence and N-1 After a second coding image sequences generation three-dimensional video-frequency bit stream, including：The three-dimensional video-frequency bit stream is write in storage medium； Alternatively, send the three-dimensional video-frequency bit stream to recipient.

In example embodiment 16, a kind of device can include：Sequence of video images acquiring unit, it is vertical for obtaining carrying N number of sequence of video images of volumetric video, N are the integer more than or equal to 2；Image sequence determination unit, for being based on described N number of regard Frequency image sequence determines the first image sequence and N-1 the second image sequences；Coding image sequences unit, for by described One image sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；Wherein, compiled for the second image sequence Coding mode includes used by code：Inter-sequence prediction coding mode；The inter-sequence prediction coding mode refers to, with described Image in one image sequence carries out predictive coding for reference frame, to the block of pixels in second image sequence.

In example embodiment 17, the N described in example embodiment 16 is 2, the sequence of video images acquiring unit, specifically For obtaining two sequence of video images for corresponding respectively to left eye and right eye.

In example embodiment 18, the image sequence determination unit described in any embodiment of example embodiment 16-17, tool Body from N number of sequence of video images for selecting a sequence of video images as described first image sequence, by other Each sequence of video images is respectively as each sequence in the N-1 the second image sequences.

In example embodiment 19, the image sequence determination unit bag described in any embodiment of example embodiment 16-18 It includes：Down-sampled subelement, for according to default down-sampled mode, to every two field picture in each sequence of video images respectively into Row is down-sampled；Sequence selection subelement, for selecting a sequence of video images in N number of sequence of video images after down-sampled As described first image sequence, using other each sequence of video images after down-sampled as the N-1 the second images Each sequence in sequence.

In example embodiment 20, the image sequence determination unit bag described in any embodiment of example embodiment 16-19 It includes：Sub-sampling sequence generates subelement, for according to default pretreatment coding mode, according to N number of sequence of video images Generate sequence of subsampled images；Sub-sampling sequence splits subelement, for according to the corresponding fractionation of the pretreatment coding mode Every two field picture in the sequence of subsampled images is split into N number of image with same time information, so as to obtain by mode N number of image sequence；Image sequence selects subelement, for from N number of image sequence, selecting the first image sequence, and will Other image sequences are as the N-1 the second image sequences.

In example embodiment 21, the coding image sequences unit bag described in any embodiment of example embodiment 16-20 It includes：First coding image sequences subelement, for being encoded to the first image sequence；Second coding image sequences subelement, For being encoded to N-1 the second image sequences；The second coding image sequences subelement, including：Block of pixels division Unit, for according to the pixel block size set according to predetermined manner, image to be encoded to be divided into multiple pixels to be encoded Block；Loop control subelement for being directed to each block of pixels to be encoded, calls following model selection subelement and pixel successively Block coded sub-units carry out coded treatment；Model selection subelement, for the type of coding according to the image to be encoded, from bag Include in the corresponding encoded set of modes of inter-sequence prediction coding mode is that block of pixels to be encoded selects the coding for meeting preset condition Pattern；Block of pixels coded sub-units, for being encoded using selected coding mode to the block of pixels to be encoded.

In example embodiment 22, the model selection subelement described in any embodiment of example embodiment 16-21, specifically For the type of coding according to the image to be encoded, from the corresponding encoded set of modes including inter-sequence prediction coding mode The coding mode for meeting the rate-distortion optimization model based on cost minimization is selected for block of pixels to be encoded.

In example embodiment 23, the model selection subelement described in any embodiment of example embodiment 16-22 includes For the following subelement of intraframe coding type：First cost computation subunit, for calculating using intra-prediction code mode The first cost encoded to the block of pixels to be encoded；Second cost computation subunit, for calculating with the first image sequence The first associated images in row are reference frame, and the block of pixels to be encoded is encoded using inter-sequence prediction coding mode Second cost；First mode selects subelement, for by the corresponding coding mould of minimum value in first cost and the second cost Formula, as the coding mode for the block of pixels selection to be encoded.

In example embodiment 24, the model selection subelement described in any embodiment of example embodiment 16-23 includes For the following subelement of interframe encode type：Third generation valency computation subunit, for calculating using intra-prediction code mode The third generation valency encoded to the block of pixels to be encoded；Forth generation valency computation subunit, for calculating using inter-prediction The forth generation valency that coding mode encodes the block of pixels to be encoded；5th cost computation subunit, for calculating with The second associated images in one image sequence are reference frame, using inter-sequence prediction coding mode to the block of pixels to be encoded into 5th cost of row coding；Second mode selects subelement, for minimum value in the three, the four, the 5th cost to be corresponded to Coding mode, as the coding mode for the block of pixels to be encoded selection.

In example embodiment 25, the block of pixels division subelement tool described in any embodiment of example embodiment 16-24 Body is used for, and according to the corresponding image analytic degree parameter of the sequence of video images, sets the pixel block size, and according to described Image to be encoded is divided into multiple block of pixels by pixel block size.

In example embodiment 26, the second cost computation subunit described in any embodiment of example embodiment 16-25 Or the 5th cost computation subunit, the picture to be encoded is searched in the reconstructed image of the first associated images or the second associated images During the matched pixel block of plain block, according in for the image encoding process to be encoded, the search in the reconstructed image The first motion vector and the coordinate of the block of pixels to be encoded obtained with block of pixels, determines in the reconstructed image Origin coordinates is searched for, and the matched pixel block is searched for since described search origin coordinates.

In example embodiment 27, the second image coded sub-units described in any embodiment of example embodiment 16-26 During coding, if the reference frame there are two or more contiguous pixels blocks is identical and motion vector is identical, The Coding and description information of the contiguous pixels block is merged into one group of write-in three-dimensional video-frequency bit stream.

In example embodiment 28, any embodiment of example embodiment 16-27 further includes：Storage unit, for by institute In the three-dimensional video-frequency bit stream write-in storage medium for stating the generation of coding image sequences unit；Alternatively, transmitting element, for by the figure As the three-dimensional video-frequency bit stream that sequential coding unit generates is sent to recipient.

In example embodiment 29, a kind of method can include：Obtain three-dimensional video-frequency bit stream to be decoded；By decoding from The first image sequence and N-1 the second image sequences are obtained in the three-dimensional video-frequency bit stream, N is the integer more than or equal to 2；According to The first image sequence and N-1 the second image sequence of the acquisition obtain N number of sequence of video images of carrying three-dimensional video-frequency；Its In, each image in the second image sequence is obtained by following decoding process：According in the three-dimensional video-frequency bit stream Corresponding encoded pattern and residual error data, reconstruction entrained by second image sequence bit stream belong to each block of pixels of described image, And synthesize described image with the block of pixels rebuild；The coding mode that the second image sequence bit stream carries includes：It is pre- between sequence Survey coding mode.

In example embodiment 30, the N described in example embodiment 29 is 2, and N number of sequence of video images is respectively to correspond to In left eye and the sequence of video images of right eye.

In example embodiment 31, described in any embodiment of example embodiment 29-30 according to the first of the acquisition Image sequence and N-1 the second image sequences obtain N number of sequence of video images of carrying three-dimensional video-frequency, including：By described first Image sequence and the N-1 the second image sequences are respectively as each sequence in N number of sequence of video images.

In example embodiment 32, described in any embodiment of example embodiment 29-31 according to the first of the acquisition Image sequence and N-1 the second image sequences obtain N number of sequence of video images of carrying three-dimensional video-frequency, further include：According to coding Used down-sampled mode carries out every two field picture in described first image sequence and second image sequence corresponding Liter sampling processing；It is described using described first image sequence and the N-1 the second image sequences as N number of video Each sequence in image sequence, including：Using the first image sequence after liter sampling and N-1 the second image sequences as institute State each sequence in N number of sequence of video images.

In example embodiment 33, described in any embodiment of example embodiment 29-32 according to the first of the acquisition Image sequence and N-1 the second image sequences obtain N number of sequence of video images of carrying three-dimensional video-frequency, including：During according to coding Used fractionation mode, for figure in described first image sequence and the second image sequence, that there is same time information As performing corresponding union operation, sequence of subsampled images is obtained；It, will according to used stereo scopic video coding pattern during coding The sequence of subsampled images be accordingly divided into N number of image sequence to get to it is described carrying three-dimensional video-frequency N number of video image sequence Row.

In example embodiment 34, any embodiment of example embodiment 29-33 is according to inter-sequence prediction coding mode weight When building the block of pixels for belonging to intraframe coding type, used reference frame includes：In the first image sequence and block of pixels Affiliated image has the image of same time information.

In example embodiment 35, any embodiment of example embodiment 29-34 is according to inter-sequence prediction coding mode weight When building the block of pixels for belonging to interframe encode type, used reference frame include the first image sequence in hypograph：With institute Stating the affiliated image of block of pixels has the image of same time information；Alternatively, temporal information is not later than the affiliated image of the block of pixels Image；Alternatively, temporal information is later than the block of pixels institute earlier than the image and temporal information of the affiliated image of the block of pixels Belong to the image of image.

In example embodiment 36, stereopsis to be decoded is obtained described in any embodiment of example embodiment 29-35 Frequency bit stream, including：Three-dimensional video-frequency bit stream is read from storage medium；Alternatively, receive the three-dimensional video-frequency bit stream sent by sender.

In example embodiment 37, a kind of device can include：Three-dimensional video-frequency bit stream acquiring unit, it is to be decoded for obtaining Three-dimensional video-frequency bit stream；Three-dimensional video-frequency bit stream decoding unit obtains first for passing through decoding from the three-dimensional video-frequency bit stream Image sequence and N-1 the second image sequences, N are the integer more than or equal to 2；Wherein, each image in the second image sequence It is obtained by following decoding process：It is corresponding entrained by the second image sequence bit stream in the three-dimensional video-frequency bit stream Coding mode and residual error data, reconstruction belong to each block of pixels of described image, and synthesize described image with the block of pixels rebuild； The coding mode that the second image sequence bit stream carries includes：Inter-sequence prediction coding mode；Sequence of video images generation is single Member, for obtaining N number of video of carrying three-dimensional video-frequency according to the first image sequence and N-1 of the acquisition the second image sequence Image sequence.

In example embodiment 38, the N described in example embodiment 37 is 2；The three-dimensional video-frequency bit stream decoding unit, specifically For obtaining the first image sequence and second image sequence from the three-dimensional video-frequency bit stream by decoding；The video figure As sequence generating unit, carrying stereopsis is obtained specifically for the first image sequence according to the acquisition and the second image sequence Two sequence of video images of frequency.

In example embodiment 39, the sequence of video images generation described in any embodiment of example embodiment 37-38 is single Member, specifically for using described first image sequence and the N-1 the second image sequences as N number of video image sequence Each sequence in row.

In example embodiment 40, the sequence of video images generation described in any embodiment of example embodiment 37-39 is single Member, including：Sampling subelement is risen, for according to down-sampled mode used by coding, to described first image sequence and institute It states every two field picture in the second image sequence and carries out corresponding liter sampling processing；Image sequence generates subelement, is adopted for that will rise The first image sequence and N-1 the second image sequence after sample is respectively as each sequence in N number of sequence of video images.

In example embodiment 41, the sequence of video images generation described in any embodiment of example embodiment 37-40 is single Member, including：Sub-sampling sequence also atomic unit, for according to coding when used fractionation mode, for described first image Image in sequence and the second image sequence, having same time information performs corresponding union operation, obtains sub-sampling figure As sequence；Sequence of video images divide subelement, for according to coding when used stereo scopic video coding pattern, by the son Sequence of subsampled images be accordingly divided into N number of image sequence to get to it is described carrying three-dimensional video-frequency N number of sequence of video images.

In example embodiment 42, the three-dimensional video-frequency bitstream decoding list described in any embodiment of example embodiment 37-41 When member belongs to the block of pixels of intraframe coding type according to the reconstruction of inter-sequence prediction coding mode, used reference frame includes：The Image in one image sequence, that there is same time information with the affiliated image of the block of pixels.

In example embodiment 43, the three-dimensional video-frequency bitstream decoding list described in any embodiment of example embodiment 37-42 When member belongs to the block of pixels of interframe encode type according to the reconstruction of inter-sequence prediction coding mode, used reference frame includes first In image sequence with hypograph：There is the image of same time information with the affiliated image of the block of pixels；Alternatively, temporal information It is not later than the image of the affiliated image of the block of pixels；Alternatively, temporal information earlier than the affiliated image of the block of pixels image and Temporal information is later than the image of the affiliated image of the block of pixels.

In example embodiment 44, the three-dimensional video-frequency bit stream described in any embodiment of example embodiment 37-43 obtains single Member, specifically for reading three-dimensional video-frequency bit stream from storage medium, alternatively, receiving the three-dimensional video-frequency bit stream sent by sender.

In example embodiment 45, a kind of machine readable media can be with store instruction, and described instruction, which is read out by the processor, to be held During row, the method described in any embodiment of example embodiment 1-15 is performed.

In example embodiment 46, a kind of machine readable media can be with store instruction, and described instruction, which is read out by the processor, to be held During row, the method described in any embodiment of example embodiment 29-36 is performed.

In example embodiment 47, a kind of system can include：Processor and memory；The memory is used to deposit Storage instruction, described instruction by the processor when reading execution, described in any embodiment of execution example embodiment 1-15 Method.

In example embodiment 48, a kind of system can include：Processor and memory；The memory is used to deposit Storage instruction, described instruction by the processor when reading execution, described in any embodiment of execution example embodiment 29-36 Method.

Although the application is disclosed as above with preferred embodiment, it is not for limiting the application, any this field skill Art personnel are not being departed from spirit and scope, can make possible variation and modification, therefore the guarantor of the application Shield scope should be subject to the scope that the application claim is defined.

Claims

A kind of 1. method, which is characterized in that including：

N number of sequence of video images of carrying three-dimensional video-frequency is obtained, N is the integer more than or equal to 2；

Based on N number of sequence of video images, the first image sequence and N-1 the second image sequences are determined；

By described first image sequence and N-1 the second coding image sequences generation three-dimensional video-frequency bit streams；

Wherein, coding mode includes used by being encoded for the second image sequence：Inter-sequence prediction coding mode；It is described Inter-sequence prediction coding mode refers to, using the image in described first image sequence as reference frame, to second image sequence In block of pixels carry out predictive coding.
2. according to the method described in claim 1, it is characterized in that, the N is 2；N number of sequence of video images is respectively pair It should be in left eye and the sequence of video images of right eye.
3. according to the method described in claim 1, it is characterized in that, described determine first based on N number of sequence of video images Image sequence and N-1 the second image sequences, including：

A sequence of video images is selected from N number of sequence of video images as described first image sequence, other are each Sequence of video images is respectively as each sequence in the N-1 the second image sequences.
4. according to the method described in claim 3, it is characterized in that, described determine first based on N number of sequence of video images Image sequence and N-1 the second image sequences, further include：

According to default down-sampled mode, every two field picture in each sequence of video images is carried out respectively down-sampled；

It is described that a sequence of video images is selected from N number of sequence of video images as described first image sequence, by it His each sequence of video images respectively as each sequence in the N-1 the second image sequences, including：It is N number of after down-sampled A sequence of video images is selected in sequence of video images as described first image sequence, by other each videos after down-sampled Image sequence is respectively as each sequence in the N-1 the second image sequences.
5. according to the method described in claim 1, it is characterized in that, described based on N number of sequence of video images, first is determined Image sequence and N-1 the second image sequences, including：

According to default pretreatment coding mode, sequence of subsampled images is generated according to N number of sequence of video images；

According to the corresponding fractionation mode of the pretreatment coding mode, every two field picture in the sequence of subsampled images is split Into N number of image with same time information, so as to obtain N number of image sequence；

From N number of image sequence, the first image sequence is selected, and using other image sequences as the N-1 the second figures As sequence.
6. according to the method described in claim 1, it is characterized in that, using following steps to be encoded in the second image sequence Image is encoded：

According to the pixel block size set according to predetermined manner, image to be encoded is divided into multiple block of pixels to be encoded；

Successively to each block of pixels to be encoded, following encoding operation is performed：

According to the type of coding of the image to be encoded, from the corresponding encoded set of modes including inter-sequence prediction coding mode Meet the coding mode of preset condition for block of pixels to be encoded selection；Using selected coding mode to the block of pixels to be encoded into Row coding.
7. according to the method described in claim 6, it is characterized in that, the coding mode for meeting preset condition includes：Meet The coding mode of rate-distortion optimization model based on cost minimization.
8. the method according to the description of claim 7 is characterized in that for intraframe coding type, the corresponding encoded set of patterns Conjunction includes：Intra-prediction code mode, inter-sequence prediction coding mode；

The type of coding according to image to be encoded, from the corresponding encoded set of modes including inter-sequence prediction coding mode Meet the coding mode of preset condition for block of pixels to be encoded selection, including：

Calculate the first cost encoded using intra-prediction code mode to the block of pixels to be encoded；

It calculates using the first associated images in the first image sequence as reference frame, is treated using inter-sequence prediction coding mode to described The second cost that encoded pixels block is encoded；

By the corresponding coding mode of minimum value in first cost and the second cost, selected as the block of pixels to be encoded Coding mode.
9. according to the method described in claim 8, it is characterized in that, first associated images, including：With the figure to be encoded As having the image of same time information.
10. the method according to the description of claim 7 is characterized in that for interframe encode type, the corresponding pre-arranged code mould Formula set includes：Intra-prediction code mode, inter prediction encoding pattern and inter-sequence prediction coding mode；

The type of coding according to image to be encoded, from the corresponding encoded set of modes including inter-sequence prediction coding mode Meet the coding mode of preset condition for block of pixels to be encoded selection, including：

Calculate the third generation valency encoded using intra-prediction code mode to the block of pixels to be encoded；

Calculate the forth generation valency encoded using inter prediction encoding pattern to the block of pixels to be encoded；

It calculates using the second associated images in the first image sequence as reference frame, is treated using inter-sequence prediction coding mode to described The 5th cost that encoded pixels block is encoded；

By the corresponding coding mode of minimum value in the three, the four, the 5th cost, selected as the block of pixels to be encoded Coding mode.
11. according to the method described in claim 10, it is characterized in that, second associated images include：

There is the image of same time information with the image to be encoded；Alternatively,

Temporal information is not later than the image of the image to be encoded；Alternatively,

Temporal information is later than the image of the image to be encoded earlier than the image and temporal information of the image to be encoded.
12. according to the method described in claim 6, it is characterized in that, according to predetermined manner set pixel block size, including：Root According to the corresponding image analytic degree parameter of the sequence of video images, the pixel block size is set.
13. the method according to claim 8 or 10, which is characterized in that calculating the mistake of the second cost or the 5th cost Cheng Zhong searches for the matched pixel block of the block of pixels to be encoded in the reconstructed image of the first associated images or the second associated images When, according in for the image encoding process to be encoded, the head that search matched pixel block obtains in the reconstructed image The coordinate of a motion vector and the block of pixels to be encoded, determines the search origin coordinates in the reconstructed image, and from Described search origin coordinates starts to search for the matched pixel block.
14. it according to the method described in claim 6, it is characterized in that, is carried out to the image to be encoded in the second image sequence It, will if the reference frame there are two or more contiguous pixels blocks is identical and motion vector is identical during coding The Coding and description information of the contiguous pixels block is merged into one group of write-in three-dimensional video-frequency bit stream.
15. according to the method described in claim 1, it is characterized in that, by described first image sequence and N-1 the second images After sequential coding generation three-dimensional video-frequency bit stream, including：

The three-dimensional video-frequency bit stream is write in storage medium；Alternatively,

The three-dimensional video-frequency bit stream is sent to recipient.
16. a kind of device, which is characterized in that including：

Sequence of video images acquiring unit, for obtaining N number of sequence of video images of carrying three-dimensional video-frequency, N is more than or equal to 2 Integer；

Image sequence determination unit for being based on N number of sequence of video images, determines the first image sequence and N-1 second Image sequence；

Coding image sequences unit, for described first image sequence and N-1 the second coding image sequences to be generated stereopsis Frequency bit stream；Wherein, coding mode includes used by being encoded for the second image sequence：Inter-sequence prediction coding mode； The inter-sequence prediction coding mode refers to, using the image in described first image sequence as reference frame, to second image Block of pixels in sequence carries out predictive coding.
17. device according to claim 16, which is characterized in that the N is 2, the sequence of video images acquiring unit, Two sequence of video images of left eye and right eye are corresponded respectively to specifically for obtaining.
18. device according to claim 16, which is characterized in that described image sequence determination unit, specifically for from institute It states and a sequence of video images is selected in N number of sequence of video images as described first image sequence, by other each video images Sequence is respectively as each sequence in the N-1 the second image sequences.
19. device according to claim 18, which is characterized in that described image sequence determination unit includes：

Down-sampled subelement, for according to default down-sampled mode, distinguishing every two field picture in each sequence of video images It carries out down-sampled；

Sequence selection subelement, for selected in N number of sequence of video images after down-sampled a sequence of video images as Described first image sequence, using other each sequence of video images after down-sampled as the N-1 the second image sequences In each sequence.
20. device according to claim 16, which is characterized in that described image sequence determination unit includes：

Sub-sampling sequence generates subelement, for according to default pretreatment coding mode, according to N number of sequence of video images Generate sequence of subsampled images；

Sub-sampling sequence splits subelement, for according to the corresponding fractionation mode of the pretreatment coding mode, the son to be adopted Every two field picture in sampled images sequence splits into N number of image with same time information, so as to obtain N number of image sequence；

Image sequence selects subelement, for from N number of image sequence, selects the first image sequence, and by other images Sequence is as the N-1 the second image sequences.
21. device according to claim 16, which is characterized in that described image sequential coding unit includes：

First coding image sequences subelement, for being encoded to the first image sequence；

Second coding image sequences subelement, for being encoded to N-1 the second image sequences；

The second coding image sequences subelement, including：

Block of pixels divides subelement, for according to the pixel block size set according to predetermined manner, image to be encoded to be divided into Multiple block of pixels to be encoded；

Loop control subelement for being directed to each block of pixels to be encoded, calls following model selection subelement and picture successively Plain block coded sub-units carry out coded treatment；

Model selection subelement, for the type of coding according to the image to be encoded, from including inter-sequence prediction coding mode Corresponding encoded set of modes in be coding mode that block of pixels to be encoded selection meets preset condition；

Block of pixels coded sub-units, for being encoded using selected coding mode to the block of pixels to be encoded.
22. device according to claim 21, which is characterized in that the model selection subelement, specifically for according to institute The type of coding of image to be encoded is stated, is picture to be encoded from the corresponding encoded set of modes including inter-sequence prediction coding mode Plain block selection meets the coding mode of the rate-distortion optimization model based on cost minimization.
23. device according to claim 22, which is characterized in that the model selection subelement is included for intraframe coding The following subelement of type：

First cost computation subunit encodes the block of pixels to be encoded using intra-prediction code mode for calculating The first cost；

Second cost computation subunit, for calculating using the first associated images in the first image sequence as reference frame, using sequence The second cost that predictive coding pattern encodes the block of pixels to be encoded between row；

First mode select subelement, for will the corresponding coding mode of minimum value in first cost and the second cost, work For the coding mode for the block of pixels selection to be encoded.
24. device according to claim 22, which is characterized in that the model selection subelement is included for interframe encode The following subelement of type：

Third generation valency computation subunit encodes the block of pixels to be encoded using intra-prediction code mode for calculating Third generation valency；

Forth generation valency computation subunit encodes the block of pixels to be encoded using inter prediction encoding pattern for calculating Forth generation valency；

5th cost computation subunit, for calculating using the second associated images in the first image sequence as reference frame, using sequence The 5th cost that predictive coding pattern encodes the block of pixels to be encoded between row；

Second mode select subelement, for will the corresponding coding mode of minimum value in the three, the four, the 5th cost, work For the coding mode for the block of pixels selection to be encoded.
25. device according to claim 21, which is characterized in that the block of pixels division subelement is specifically used for, according to The corresponding image analytic degree parameter of the sequence of video images, sets the pixel block size, and according to the pixel block size Image to be encoded is divided into multiple block of pixels.
26. the device according to claim 23 or 24, which is characterized in that the second cost computation subunit or the 5th generation Valency computation subunit searches for of the block of pixels to be encoded in the reconstructed image of the first associated images or the second associated images During with block of pixels, according in for the image encoding process to be encoded, in the reconstructed image search for matched pixel block Obtained first motion vector and the coordinate of the block of pixels to be encoded determines the search starting in the reconstructed image Coordinate, and the matched pixel block is searched for since described search origin coordinates.
27. device according to claim 16, which is characterized in that the second image coded sub-units are in the process of coding In, if the reference frame there are two or more contiguous pixels blocks is identical and motion vector is identical, by the continuous image The Coding and description information of plain block is merged into one group of write-in three-dimensional video-frequency bit stream.
28. device according to claim 16, which is characterized in that further include：

Storage unit, the three-dimensional video-frequency bit stream for described image sequential coding unit to be generated are write in storage medium；Alternatively,

Transmitting element, the three-dimensional video-frequency bit stream for described image sequential coding unit to be generated are sent to recipient.
A kind of 29. method, which is characterized in that including：

Obtain three-dimensional video-frequency bit stream to be decoded；

Obtain from the three-dimensional video-frequency bit stream the first image sequence and N-1 the second image sequences by decoding, N be more than etc. In 2 integer；

N number of video figure of carrying three-dimensional video-frequency is obtained according to the first image sequence and N-1 of the acquisition the second image sequence As sequence；

Wherein, each image in the second image sequence is obtained by following decoding process：According to the three-dimensional video-frequency position Corresponding encoded pattern entrained by the second image sequence bit stream and residual error data, reconstruction in stream belong to each picture of described image Plain block, and synthesize described image with the block of pixels rebuild；The coding mode that the second image sequence bit stream carries includes：Sequence Between predictive coding pattern.
30. according to the method for claim 29, which is characterized in that the N is 2；N number of sequence of video images is respectively Corresponding to left eye and the sequence of video images of right eye.
31. according to the method for claim 29, which is characterized in that first image sequence and N- according to the acquisition 1 the second image sequence obtains N number of sequence of video images of carrying three-dimensional video-frequency, including：

In using described first image sequence and the N-1 the second image sequences as N number of sequence of video images Each sequence.
32. according to the method for claim 29, which is characterized in that first image sequence and N- according to the acquisition 1 the second image sequence obtains N number of sequence of video images of carrying three-dimensional video-frequency, further includes：

According to down-sampled mode used by coding, to every frame in described first image sequence and second image sequence Image carries out corresponding liter sampling processing；

It is described using described first image sequence and the N-1 the second image sequences as N number of sequence of video images In each sequence, including：It N number of is regarded using the first image sequence after liter sampling and N-1 the second image sequences as described Each sequence in frequency image sequence.
33. according to the method for claim 29, which is characterized in that first image sequence and N- according to the acquisition 1 the second image sequence obtains N number of sequence of video images of carrying three-dimensional video-frequency, including：

According to coding when used fractionation mode, for it is in described first image sequence and the second image sequence, have phase Image with temporal information performs corresponding union operation, obtains sequence of subsampled images；

According to used stereo scopic video coding pattern during coding, the sequence of subsampled images is accordingly divided into N number of image Sequence to get to it is described carrying three-dimensional video-frequency N number of sequence of video images.
34. according to the method for claim 29, which is characterized in that belonged to according to the reconstruction of inter-sequence prediction coding mode in frame During the block of pixels of type of coding, used reference frame includes：Image tool in first image sequence, affiliated with the block of pixels There is the image of same time information.
35. according to the method for claim 29, which is characterized in that interframe is belonged to according to the reconstruction of inter-sequence prediction coding mode During the block of pixels of type of coding, used reference frame include the first image sequence in hypograph：

There is the image of same time information with the affiliated image of the block of pixels；Alternatively,

Temporal information is not later than the image of the affiliated image of the block of pixels；Alternatively,

Temporal information is later than the affiliated image of the block of pixels earlier than the image and temporal information of the affiliated image of the block of pixels Image.
36. according to claim 29-35 any one of them methods, which is characterized in that described to obtain three-dimensional video-frequency to be decoded Bit stream, including：

Three-dimensional video-frequency bit stream is read from storage medium；Alternatively,

Receive the three-dimensional video-frequency bit stream sent by sender.
37. a kind of device, which is characterized in that including：

Three-dimensional video-frequency bit stream acquiring unit, for obtaining three-dimensional video-frequency bit stream to be decoded；

Three-dimensional video-frequency bit stream decoding unit obtains the first image sequence and N- for passing through decoding from the three-dimensional video-frequency bit stream 1 the second image sequence, N are the integer more than or equal to 2；Wherein, each image in the second image sequence is by following solution Code process obtains：Corresponding encoded pattern entrained by the second image sequence bit stream in the three-dimensional video-frequency bit stream and residual Difference data, reconstruction belong to each block of pixels of described image, and synthesize described image with the block of pixels rebuild；Second image The coding mode that sequence bit stream carries includes：Inter-sequence prediction coding mode；

Sequence of video images generation unit, for being obtained according to the first image sequence and N-1 of the acquisition second image sequence To N number of sequence of video images of carrying three-dimensional video-frequency.
38. the device according to claim 37, which is characterized in that the N is 2；The three-dimensional video-frequency bit stream decoding unit, The first image sequence and second image sequence are obtained from the three-dimensional video-frequency bit stream specifically for passing through decoding；It is described to regard It is vertical to obtain carrying specifically for the first image sequence according to the acquisition and the second image sequence for frequency image sequence generation unit Two sequence of video images of volumetric video.
39. the device according to claim 37, which is characterized in that the sequence of video images generation unit is specifically used for Using described first image sequence and the N-1 the second image sequences as each sequence in N number of sequence of video images Row.
40. the device according to claim 37, which is characterized in that the sequence of video images generation unit, including：

Sampling subelement is risen, for according to down-sampled mode used by coding, to described first image sequence and described the Every two field picture in two image sequences carries out corresponding liter sampling processing；

Image sequence generates subelement, makees respectively for that will rise the first image sequence after sampling and N-1 the second image sequences For each sequence in N number of sequence of video images.
41. the device according to claim 37, which is characterized in that the sequence of video images generation unit, including：

Sub-sampling sequence also atomic unit, for according to coding when used fractionation mode, for described first image sequence Corresponding union operation is performed with image in the second image sequence, that there is same time information, obtains sub-sampled images sequence Row；

Sequence of video images divides subelement, for according to used stereo scopic video coding pattern during coding, the son to be adopted Sampled images sequence be accordingly divided into N number of image sequence to get to it is described carrying three-dimensional video-frequency N number of sequence of video images.
42. the device according to claim 37, which is characterized in that the three-dimensional video-frequency bit stream decoding unit is according between sequence When predictive coding mode reconstruction belongs to the block of pixels of intraframe coding type, used reference frame includes：In first image sequence , with the affiliated image of the block of pixels have same time information image.
43. the device according to claim 37, which is characterized in that the three-dimensional video-frequency bit stream decoding unit is according between sequence When predictive coding mode reconstruction belongs to the block of pixels of interframe encode type, used reference frame is included in the first image sequence With hypograph：

There is the image of same time information with the affiliated image of the block of pixels；Alternatively,

Temporal information is not later than the image of the affiliated image of the block of pixels；Alternatively,

Temporal information is later than the affiliated image of the block of pixels earlier than the image and temporal information of the affiliated image of the block of pixels Image.
44. according to claim 37-43 any one of them devices, which is characterized in that the three-dimensional video-frequency bit stream acquiring unit, Specifically for reading three-dimensional video-frequency bit stream from storage medium, alternatively, receiving the three-dimensional video-frequency bit stream sent by sender.
45. a kind of machine readable media, which is characterized in that the machine readable media store instruction, described instruction is by processor When reading execution, such as claim 1-15 any one of them methods are performed.
46. a kind of machine readable media, which is characterized in that the machine readable media store instruction, described instruction is by processor When reading execution, such as claim 29-36 any one of them methods are performed.
47. a kind of system, which is characterized in that including：

Processor；

Memory, for storing instruction, described instruction perform claim 1-15 such as and appoint when reading execution by the processor Method described in one.
48. a kind of system, which is characterized in that including：

Processor；

Memory, for storing instruction, described instruction perform such as claim 29-36 when reading execution by the processor Any one of them method.