CN102761742B - Transform block division methods, transform block divides coding method and the coding/decoding method of parameter - Google Patents

Transform block division methods, transform block divides coding method and the coding/decoding method of parameter Download PDF

Info

Publication number
CN102761742B
CN102761742B CN201210227646.4A CN201210227646A CN102761742B CN 102761742 B CN102761742 B CN 102761742B CN 201210227646 A CN201210227646 A CN 201210227646A CN 102761742 B CN102761742 B CN 102761742B
Authority
CN
China
Prior art keywords
transform block
block
size
parameter
division
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201210227646.4A
Other languages
Chinese (zh)
Other versions
CN102761742A (en
Inventor
郑萧桢
袁媛
于浩平
何芸
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tsinghua University
Huawei Technologies Co Ltd
Original Assignee
Tsinghua University
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tsinghua University, Huawei Technologies Co Ltd filed Critical Tsinghua University
Priority to CN201210227646.4A priority Critical patent/CN102761742B/en
Publication of CN102761742A publication Critical patent/CN102761742A/en
Priority to PCT/CN2013/078613 priority patent/WO2014005500A1/en
Application granted granted Critical
Publication of CN102761742B publication Critical patent/CN102761742B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/119—Adaptive subdivision aspects, e.g. subdivision of a picture into rectangular or non-rectangular coding blocks

Abstract

The present invention provides a kind of transform block division methods, the division coding method and coding/decoding method of transform block.Transform block division methods include：Obtain the minimum value of the size of default transform block；To transform block when carrying out non-square division, if the minimum value of the size in the default transform block wide or high of the transform block, stops dividing the transform block.The embodiment of the present invention when carrying out non-square to transform block and dividing, transform block it is wide or high one of them be equal to the size minimum value of default transform block when, stop dividing.

Description

Transform block division methods, transform block divides coding method and the coding/decoding method of parameter
Technical field
The present embodiments relate to image coding technique, more particularly to a kind of transform block division methods, transform block divides ginseng Several coding methods and coding/decoding method.
Background technology
To reduce the bandwidth shared by transmission of video, it is necessary to carry out coded treatment to video data, wherein, interframe compression side Method is exactly a kind of common video coding technique, and it can treat transmitting video image and be compressed treatment, reduces transmission of video Data.
Interframe compression method is a kind of video coding technique based on estimation, and Video coding end is carried out to video image The process of coding includes：First, image block to be encoded is divided into several equal-sized subimage blocks；Then, for every Individual subimage block, searches for the image block most matched with current subimage block as prediction block in a reference image；Thereafter, by the son The respective pixel values of image block and prediction block are subtracted each other and obtain residual error, and the corresponding residual error of each subimage block that will be obtained is combined one Rise, obtain the residual error of image block；Then by the residual error of image block it is transformed with quantify after the value that obtains carry out entropy code；Finally, The bit stream and motion vector information that entropy code is obtained issue decoding end in the lump, wherein, motion vector information represents current son The alternate position spike of image block and prediction block.Decoding end receive coding side transmission bit stream and motion vector information after, carry out with The opposite processing procedure of coding side, you can obtain the initial data of corresponding image block.At present, in image encoding process, Residual error to image block enters line translation, is to utilize transform block（Also referred to as transformation matrices）Conversion process is carried out to residual error, is converted Coefficient matrix, to remove the redundancy of image block, improves the code efficiency of image block.Because there is difference in the residual error of image block The regularity of distribution, conversion process is carried out to it using the transform block of a certain specific dimensions and does not often reach preferable transform effect, Therefore need to use various sizes of transform block carries out conversion process come the residual error to image block, is imitated with obtaining optimal conversion Really.
, it is necessary to divide transform block in conventional images block cataloged procedure, it is adapted to enter the residual error of image block to obtain The transform block of line translation treatment, wherein, the original dimension of transform block is usually square, i.e. the height and width of transform block are equal.Tradition When method is divided to transform block, be using square division, will transform block be divided into equal square sub- conversion multiple wide and high Block, then carries out conversion process using square transform block to the residual error of image block.
Because image block is when being divided, the non-square dividing modes such as horizontal division, vertical division can be used, and it is this Non-square divides the texture information that can embody image block, now, line translation is entered to the residual error of image block according to square transform block During treatment, square transform block is possible to across the adjacent corresponding residual error of two subimage blocks, due to two neighboring subgraph Can there is the conversion of jumping characteristic in the corresponding residual error of block, when now entering line translation using square transform block, will cause change action Weaken, it is impossible to effectively remove the redundancy of image block, reduce code efficiency.Therefore, prior art it is also proposed a kind of use The method that non-square transform block carries out conversion process to the residual error of image block, wherein, when being divided to transform block, according to The consistent dividing mode of image block, the son conversion that division is obtained is block-shaped consistent with subimage block size, in this manner it is possible to keep away The problem of the redundancy that can not effectively remove image block existed when exempting from using square transform block to the residual noise reduction of image block. Additionally, it is necessary to be encoded to the division parameter of transform block, each transform block needs to compile in partition process is carried out to transform block The division parameter of code conversion block, to represent whether transform block needs to divide.
But, when being divided to conversion, it will usually the minimum value of the size of default transform block, and it is existing using non-side change Change in the transform block partition process of block, when being the minimum value by the size for being divided into default transform block wide and high of transform block, Can just stop divide, so, when being divided to transform block, may exist simultaneously using horizontal division, vertical division or The dividing mode that horizontal division is intersected with vertical division, the coding of transform block parameters is complicated, causes the division logic of transform block to be answered Miscellaneous, the transform block for marking off can also influence the code efficiency of image block, increase the complexity of coding.
The content of the invention
The embodiment of the present invention provides a kind of transform block division methods, and transform block divides coding method and the decoding side of parameter Method, effectively can divide to transform block, improve the code efficiency of image block.
The embodiment of the present invention provides a kind of transform block division methods, including：
Obtain the minimum value of the size of default transform block；
To transform block when carrying out non-square division, if the chi in the default transform block wide or high of the transform block Very little minimum value, then stop dividing the transform block；
Wherein, it refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided.
The embodiment of the present invention separately provides the coding method that a kind of transform block divides parameter, including：
Obtain the minimum value of the size of default transform block；
Relation between the minimum value of the size of size and the default transform block according to transform block, judges whether coding The division parameter of transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
The embodiment of the present invention also provides the coding/decoding method that a kind of transform block divides parameter, it is characterised in that including：
Obtain the minimum value of the size of default transform block；
Relation between the minimum value of the size of size and the default transform block according to transform block, judges whether decoding The division parameter of transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
Additionally, the embodiment of the present invention also provides a kind of transform block divides device, including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Transform block division module, for transform block carry out non-square divide when, if the transform block is wide or high In the minimum value of the size of the default transform block, then stop dividing the transform block；
Wherein, the minimum value of the size of the default transform block is the minimum value wide or high of transform block, or wide and high The minimum value of the average value of sum；It refers to that image block or transform block are divided into wide and division not etc. high that the non-square is divided Method.
The embodiment of the present invention provides the code device that a kind of transform block divides parameter, including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Parameter coding module, between the minimum value of the size according to transform block and the size of the default transform block Relation, judges whether the division parameter of transcoding, coding transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
The embodiment of the present invention provides the decoding apparatus that a kind of transform block divides parameter, including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Parameter decoding module, between the minimum value of the size according to transform block and the size of the default transform block Relation, judges whether the division parameter of decoded transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
The embodiment of the present invention provides a kind of Image Codec, including：
Transform block divides the code device of parameter, for the size according to transform block and the size of the default transform block Relation between minimum value, judges whether the division parameter of transcoding, coding transform block；
Transform block divides the decoding apparatus of parameter, for the size according to transform block and the size of the default transform block Relation between minimum value, judges whether the division parameter of decoded transform block；
Wherein, the division parameter of the transform block is used to represent and transform block is divided or is not divided
Transform block division methods provided in an embodiment of the present invention, transform block divides coding method and the coding/decoding method of parameter, When carrying out non-square to transform block and dividing, if transform block it is wide or high one of them be equal to the size of default transform block most Small value, just stops the division to transform block, improves the code efficiency of image block.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are this hairs Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can be with root Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 a- Fig. 1 d are the structural representation that image block is divided using symmetrical dividing mode；
Fig. 2 a- Fig. 2 d are the structural representation that image block is divided using asymmetric dividing mode；
The scale diagrams of the change block that Fig. 3 a- Fig. 3 c are used for image block；
Fig. 4 is the schematic flow sheet of the transform block division methods that the embodiment of the present invention one is provided；
Fig. 5 is the schematic flow sheet of the transform block division methods that the embodiment of the present invention two is provided；
Fig. 6 is the transform block division methods schematic flow sheet that the embodiment of the present invention three is provided；
Fig. 7 is the coding method schematic flow sheet that the transform block that the embodiment of the present invention four is provided divides parameter；
Fig. 8 is the schematic flow sheet of the coding/decoding method that the transform block that the embodiment of the present invention five is provided divides parameter；
Fig. 9 is that the transform block that the embodiment of the present invention six is provided divides apparatus structure schematic diagram；
Figure 10 is that the transform block that the embodiment of the present invention seven is provided divides apparatus structure schematic diagram；
Figure 11 is that the transform block that the embodiment of the present invention eight is provided divides apparatus structure schematic diagram；
Figure 12 is the structural representation that the transform block that the embodiment of the present invention nine is provided divides device；
Figure 13 is the structural representation of the code device that the transform block that the embodiment of the present invention ten is provided divides parameter；
Figure 14 is the structural representation of the code device that the transform block that the embodiment of the present invention 11 is provided divides parameter；
Figure 15 is the structural representation of the code device that the transform block that the embodiment of the present invention 12 is provided divides parameter；
Figure 16 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 13 is provided divides parameter；
Figure 17 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 14 is provided divides parameter；
Figure 18 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 15 is provided divides parameter；
Figure 19 is the structural representation of the Image Codec that the embodiment of the present invention 16 is provided.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, below in conjunction with the embodiment of the present invention in it is attached Figure, is clearly and completely described to the technical scheme in the embodiment of the present invention, it is clear that described embodiment is the present invention A part of embodiment, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not having The every other embodiment obtained on the premise of making creative work, belongs to the scope of protection of the invention.
For ease of the understanding to embodiment of the present invention technical scheme, first below to the division of image block in Video coding, The division of transform block is described respectively.
The division of image block：
In existing video coding system and decoding standard, such as mobile image expert group（MovingPicture Experts Group, MPEG）, enhancing Video coding (Advanced Video Coding, H.264/AVC), an image block, Or be macro block (macroblock), supermacro block (super-macroblock) etc., several subimage blocks can be divided into, this The size of a little image block is 16x16,16x8,8x16,8x8,8x4,4x8,4x4 etc., and subimage block is transported with these sizes Dynamic to estimate and motion compensation, the coding side of image needs that the code word of identification image block dividing mode is sent to the decoding of image End, knows the dividing mode at Image Coding end so as to the decoding end of image, and according to the dividing mode and motion vector information, really Fixed corresponding prediction block.These subimage blocks are NXM (N and M are whole more than 0 in existing video encoding and decoding standard Number) rectangular block, and N and M have multiple proportion.
Wherein, image block is divided into the usual way of subimage block and has：2Nx2N dividing modes, image block only includes one Individual subimage block, the i.e. image block are not divided into smaller subimage block, as shown in Figure 1a；2NxN dividing modes, by image Block is divided into the big subimage block of upper and lower two grades, as shown in Figure 1 b；Nx2N dividing modes, left and right two is divided into by image block Etc. big subimage block, as illustrated in figure 1 c；NxN dividing modes, four big subimage blocks of grade, such as Fig. 1 d are divided into by image block It is shown.In above-mentioned each dividing mode, N is any positive integer, and above-mentioned dividing mode is symmetrical dividing mode.
Additionally, image block can also use asymmetric dividing mode, as shown in Fig. 2 a to Fig. 2 d, wherein, Fig. 2 a and 2b institutes Show that an image block is divided into upper and lower two anisometric rectangle subimage blocks by dividing mode；Dividing mode shown in Fig. 2 c and 2d One image block is divided into the two anisometric rectangle subimage blocks in left and right.Specifically, Fig. 2 a are shown and are divided using 2NxnU Mode, wherein n=0.5N, in two subimage blocks that the dividing mode is marked off, the subimage block 2Nx0.5N of top, below Subimage block be 2Nx1.5N, wherein, the U in 2NxnU represents image dividing line with respect on the perpendicular bisector of the image block Move, 2NxnU represents that image dividing line is relative and moves n on the perpendicular bisector of the image block, wherein, n=x*N, wherein x more than or wait In 0 and less than or equal to 1；Fig. 2 b show and use 2NxnD dividing modes, wherein n=0.5N, what the dividing mode was marked off In two subimage blocks, the subimage block of top is 2Nx1.5N, and following subimage block is 2Nx0.5N, wherein, in 2NxnD D represents that image dividing line is moved down with respect to the perpendicular bisector of the image block, and 2NxnD represents image dividing line with respect to the image block Perpendicular bisector moves down n, wherein, n=x*N, wherein x are more than or equal to 0 and less than or equal to 1；Fig. 2 c show use NLx2N dividing modes, wherein n=0.5N, in two subimage blocks that the dividing mode is marked off, the subimage block on the left side is 0.5Nx2N, wherein, the L in nLx2N represents that image dividing line is moved to left with respect to the perpendicular bisector of the image block, and nLx2N represents table Diagram moves to left n as dividing line with respect to the perpendicular bisector of the image block, wherein, n=x*N, x are more than or equal to 0 and be less than or wait In 1, the subimage block on the right is 1.5Nx2N；Fig. 2 d are shown using nRx2N dividing modes, wherein n=0.5N, the division side In two subimage blocks that formula is marked off, the subimage block on the left side is 1.5Nx2N, and the subimage block on the right is 0.5Nx2N, R in nRx2N represents that image dividing line is moved to right with respect to the perpendicular bisector of the image block, and nRx2N represents image dividing line Perpendicular bisector with respect to the image block moves to right n, wherein, n=x*N, x are more than or equal to 0 and less than or equal to 1.
In the dividing mode of above-mentioned image block, image block is divided what is vertically arranged by horizontal bar separated time Multiple subimage blocks, the division direction at this time using is horizontal division direction, above-mentioned 2NxN dividing modes, 2NxnU division sides Formula, 2NxnD dividing modes are referred to as horizontal division mode；And pass through vertical division line and be divided into image block to arrange in the horizontal direction Multiple subimage blocks of row are vertical division mode, and the division direction at this time using is vertical division direction, and above-mentioned Nx2N draws The mode of dividing, nLx2N dividing modes, nRx2N dividing modes are referred to as vertical division mode；Simultaneously by horizontal bar separated time and vertical stroke Separated time, four subimage blocks are divided into by image block, and the division direction at this time using is then horizontal and vertical division direction, on NxN dividing modes are stated for horizontal and vertical dividing mode.Wherein horizontal division mode and vertical division mode are referred to as non-square stroke The mode of dividing, each subimage block after division is non-square structure, and horizontal and vertical dividing mode is referred to as square dividing mode, draws Each subimage block after point is square structure.
In image block cataloged procedure, it is necessary to inquire about divide after subimage block prediction block, therefore, above-mentioned image block draw In the mode of dividing, it is also possible to referred to the type of prediction block, the dividing mode of image block is corresponding with prediction block type.
In being divided using the image block of non-square dividing mode, dividing mode can reflect the texture in region residing for image block Information, when image block use level dividing mode, the texture in region residing for image block is more likely to the spy with horizontal texture Levy, when image block is using vertical division mode, the texture in region residing for image block is more likely to the feature with vertical texture. Shape according to the transform block matched with image texture characteristic can further improve the efficiency of transition coding, specifically, When image block use level dividing mode, transform block is horizontal strip shape to adapt to the conversion of the feature with horizontal texture Coding；When image block uses vertical division mode, transform block is vertical sliver shape to adapt to the feature with vertical texture Transition coding.Therefore, non-square dividing mode can be also used when being divided to transform block, transform block is divided into The non-square structure consistent with subimage block, is matched with the textural characteristics with image block, improves the code efficiency of image block.
Transform block is divided：
In Video coding and decoding technique, it is possible to use transform block（That is transformation matrix）Remove the phase of the residual error of image block Guan Xing, that is, remove the redundancy of image block, and to improve code efficiency, the conversion of the data block in image block generally uses two Dimension conversion, i.e., the transformation matrix and its transposed matrix in coding side by the residual information of data block respectively with a NXM are multiplied, phase Multiply that obtain afterwards is conversion coefficient.Above-mentioned steps can be described using below equation：
f=T'×C×T
Wherein, C represents the transposed matrix of the residual information of data block, T and T' representation transformations matrix and transformation matrix, f generations The transform coefficient matrix obtained after the residual information of list data block is transformed.Wherein, transformation matrix can be discrete cosine transform (Discrete Cosine Transform, DCT) matrix, integer transform (Integer Transform) matrix, KL conversion (Karhunen L ò eve Transform, KLT) matrix etc., wherein, KLT can preferably consider image block or image block residual error Texture information.
Above-mentioned treatment is carried out to the residual information of image block to be changed to frequency from spatial domain equivalent to by the residual information of image block Domain, and the transform coefficient matrix obtained after treatment concentrates on low frequency region；Coding side is carried out to the residual information of image block State after conversion, after the treatment such as the transform coefficient matrix to being obtained after conversion is quantified again, entropy code, entropy code is obtained Bit stream is sent to decoding end.
In order that decoding end knows the transformation matrix type and size that coding side is used, generally, coding side can would indicate that The configured information of the transformation matrix that current image block is used is sent to decoding end, and decoding end determines to compile according to above-mentioned configured information The transformation matrix that code end uses, according to the characteristics of transformation matrix（Orthogonality of transformation matrix etc.）, the bit sent to coding side Stream decode obtaining transform coefficient matrix, and transform coefficient matrix is multiplied with transformation matrix and its transposed matrix, can be recovered To the residual information of consistent data block approximate with coding side.Above-mentioned steps can be described using below equation：
C=T × f × T',
Wherein, C represents the transposed matrix of the residual information of data block, T and T' representation transformations matrix and transformation matrix, f generations The transform coefficient matrix that table decoding end is obtained.
Because the residual error of image block there may be the different regularities of distribution, using a certain specific dimensions transformation matrix often The transform effect not reached, accordingly, it would be desirable to the residual error of image block is attempted using various sizes of transform block, therefore, for One image block of 2Nx2N can be the transformation matrix of 2Nx2N using size, it is also possible to using size for NxN transformation matrix, Or size is the transformation matrix of 0.5Nx0.5N, therefore, it is effectively to represent how image block uses various sizes of conversion square Battle array, it is possible to use tree-like identification method, when the conversion block size that identification image block is used, there be for identifying ground floor in code stream Whether image block is the indicating bit of 2Nx2N transformation matrixs using size, if image block is 2Nx2N transformation matrixs using size （As shown in Figure 3 a）, then the indicating bit is 0；If image block does not use 2Nx2N to convert, the indicating bit is 1, and representing needs It is the transformation matrix that 2Nx2N transformation matrixs are further divided into that four sizes are NxN by size, and in the Rotating fields of code stream second Identify whether the transformation matrix that each size is NxN further divides respectively with 4 bits；If image block is used such as Fig. 3 b During shown mapped structure, 4 bits are all 0, represent that the transformation matrix that each size is NxN is not divided further；When During from mapped structure as shown in Figure 3 c, then it is that 1,2 bits are 0 to represent for 0,2 bits to have 2 bits in 4 bits To the size of lower-left and upper right for the transformation matrix of NxN is no longer divided；2 bits are that 1 expression is needed to upper left and bottom right Size for NxN transformation matrix need further divide, obtain size be 0.5Nx0.5N transformation matrix；Then in code stream Represented whether with 4 bits in third layer structure is carried out further to the transformation matrix that the size of upper left is 0.5Nx0.5N Divide, represented whether with 4 bits and the transformation matrix that the size of bottom right is 0.5Nx0.5N is further divided, such as When fruit image block uses mapped structure as shown in Figure 3 c, above-mentioned 4+4 bit is all 0, and expression is not divided further.It is logical Crossing above-mentioned successively mark in code stream can effectively and flexibly represent the transform size that image block and subimage block are used.
As can be seen that when the residual error to image block carries out conversion process, it is necessary to divided to transform block, to obtain figure As the size of the transform block that block is used, optimal transform effect is obtained, eliminate the redundancy of image block, there is provided image block Code efficiency.
When being divided to transform block, transform block can be divided into equal to image block or the size less than image block, and During the coding and decoding of transform block, the division number of plies that generally default transform block is allowed, and default transform block size Minimum value, wherein, it refers to the usable most division numbers of plies of transform block that the divisions number of plies that transform block is allowed is, such as image volume Decoding apparatus and image decoding apparatus preset at most usable two-layer and divide, if the size of transform block initial division is 2Nx2N, Transform block can be divided to the 1st layer by the 0th layer, and the size of transform block is changed into NxN from 2Nx2N.Or, image encoding apparatus can also The configured information of the at most usable division number of plies of mark is write in code stream, this can at most make to notify image decoding apparatus The division number of plies, it is 0.5N that minimum conversion block size is for example preset in image encoding apparatus and image decoding apparatus, If the size of transform block is NxN, when transform block is divided to 0.5Nx0.5N, transform block does not continue to divide.
Above two method also may be used in combination, for example, image encoding apparatus and image decoding apparatus preset at most Three layers of division can be used, the minimum value for presetting the size of transform block is 0.5N.If the size of transform block initial division is 2Nx2N, Then transform block can be divided to the 2nd layer by the 0th layer, and its size is divided into NxN and then is further continued for being divided into 0.5Nx0.5N by 2Nx2N. If the size of transform block initial division is NxN, transform block can be divided to the 1st layer by the 0th layer, and its size is converted by NxN 0.5Nx0.5N, now due to conversion block size reached default transform block size minimum value, now transform block not followed by It is continuous to divide.
In practical application, the division of transform block can use quaternary tree（quadtree）Dividing mode or binary tree（binary tree）Dividing mode is divided, wherein, quad-tree partition refer to transform block by one point for four model split, after division Conversion block size is identical, i.e., the width after all transform blocks are divided is equal, and the height after all transform blocks are divided is equal；Binary tree Dividing mode refers to transform block by the model split being divided into two, and each the conversion block size after division is identical.
Additionally, being to be matched with the texture features of image block, when being divided to transform block, can also use and image block one The dividing mode of cause, i.e., divided using non-square dividing mode to transform block, but, it is existing that non-square is carried out to transform block , it is necessary to when the size on the both sides of transform block to be equal to the minimum value of the size of default transform block, just stop dividing, example during division Such as, when the transform block of 2Mx2M is divided using non-square, 0.5Mx2M is divided into by 2Mx2M, and be divided into by 0.25MxM During 0.25Mx0.25M, it is assumed that when the minimum dimension of default transform block is 0.25M, then the transform block of 0.25MxM may proceed to be divided into 0.25Mx0.25M.The problem of this division methods is that transform block divide the change for obtaining along the side for being not equal to minimum dimension Change the ratio of width to height of the shape of block, i.e. transform block, with division before transform block shape it is inconsistent.It is this it is inconsistent can influence conversion The code efficiency of block, and transform block can be caused to there are three kinds of division directions simultaneously when non-square is divided：Horizontal division direction （Transform block is divided into 0.5Mx2M by 2Mx2M）, horizontal and vertical division direction（Transform block is divided into 0.25MxM by 0.5Mx2M） With vertical division direction（Transform block is divided into 0.25Mx0.25M by 0.25MxM）, this can cause to divide the increasing of logical complexity Plus, and coding side also needs the code efficiency of trial 0.25Mx0.25M transform blocks, can also increase the complexity of coding side.This reality In applying example, when being divided to transform block can using the dividing mode as image block, and use one point for four quaternary tree Mode is divided.
The embodiment of the present invention to using non-square dividing mode transform block is divided when, as long as transform block width or During the minimum value of the high size in default transform block, with regard to stopping dividing transform block, imitated with the coding for improving transform block Rate.The division with specific embodiment to transform block is illustrated below.
Fig. 4 is the schematic flow sheet of the transform block division methods that the embodiment of the present invention one is provided.As shown in figure 4, this implementation Example method may include following steps：
Step 101, the minimum value for obtaining the size for presetting transform block；
Step 102, to transform block when carrying out non-square and dividing, if the chi in default transform block wide or high of transform block Very little minimum value, then stop dividing transform block；
Wherein, preset transform block size minimum value for transform block minimum value wide or high, or transform block width With the minimum value of the average value of sum high；It refers to that image block or transform block are divided into wide and division not etc. high that non-square is divided Method.
The present embodiment be can be applied in the encoding and decoding treatment of image, and transform block can be divided, and be met to figure with obtaining The size of the transform block of conversion process is carried out as the residual error of block, specifically, when image block is divided using non-square, as long as conversion The minimum value of the size in default transform block wide or high of block, just stops the division to changing block, so, the conversion after division Block is wide consistent with shape high, i.e. wide high proportion relationship consistency, and the code efficiency of image block can be improved.
To sum up, transform block division methods provided in an embodiment of the present invention, when non-square division is carried out to transform block, as long as Transform block it is wide or high one of them be equal to default transform block size minimum value, just stop the division to transform block, can make Each conversion after must dividing is block-shaped consistent, improves the transition coding efficiency of image block.
Fig. 5 is the schematic flow sheet of the transform block division methods that the embodiment of the present invention two is provided.As shown in figure 5, this implementation When example is divided to transform block, it may include following steps：
Step 201, judge whether to transform block using non-square divide, be then execution step 203, otherwise, perform step 202；
Step 202, the transform block is divided using square dividing mode, until the size of transform block is equal to default change When changing the minimum dimension of block, stop dividing, terminate；
Step 203, non-square division is carried out to the transform block, judge the transform block after dividing it is wide or high one of them be The minimum value of the no size equal to default transform block, is to stop the division to transform block, is terminated, and otherwise, continues executing with step 203, until the transform block after division it is wide or high one of them be equal to default transform block size minimum value.
It will be understood by those skilled in the art that the original dimension of transform block be it is square, i.e., it is wide equal with height, work as transform block It is to be divided into non-square by square when being divided using non-square, now, the wide of initial transformation block will be divided into original four points One of, or height will be divided into original a quarter, therefore, it is to become by the square precondition for being divided to non-square division The original dimension for changing block is at least 4 times of minimum value of size of default transform block, i.e., only transform block wide or height at least For 4 times of minimum value of the size for presetting transform block when, non-square division just is carried out to transform block, otherwise, using square division.
Specifically, the size for assuming image block is 2Mx2M, the maximum for presetting the size of transform block is 2M, default conversion The minimum value of the size of block is 0.5M, and the initial value of the size of transform block is equal to the size of image block, and transform block allows what is divided It is 3 layers to divide the number of plies, and the dividing mode of image block is Nx2N, and because the initial value of transform block is 2Mx2M, 2M is default transform block 4 times of minimum value 0.5M of size, therefore transform block can use the non-square dividing mode consistent with image block, that is, use The horizontal division mode of Nx2N, non-square division is carried out to transform block.And the original dimension for working as transform block is MxM, conversion is preset The minimum value of the size of block is 0.5M, due to 2 times of the minimum value of size that the size of transform block is default transform block, then now Square division is used to transform block, will the transform block of MxM be divided into the transform block of 0.5Mx0.5M.
In practical application, it is less than due to the full-size of the corresponding transform block of image block or equal to the size of image block, And the maximum of the size of default transform block is likely less than the size of image block, therefore, the corresponding maximum transform block of image block Size is equally likely to the size of image block or the full-size of default transform block.Therefore, judging whether transform block is carried out When non-square is divided, in addition it is also necessary to the relation between the maximum of the size of transform block and the size of pre-set image block is judged, with true Determine the dividing mode of transform block.Specifically, when image block uses horizontal division, the width of transform block is equal to the height of transform block, and becomes During the maximum of the size that the average value for changing the sum wide and high of block is less than or equal to default transform block, using horizontal division side Formula, divides to transform block so that the width of the transform block after division is equal to the width of transform block before dividing, the transform block after division Height less than divide before transform block height；When image block uses vertical division, the width of transform block is equal to the height of transform block, and conversion When the average value of the sum wide and high of block is less than the maximum of the size of default transform block, using vertical division mode, to conversion Block is divided so that the width of the transform block after division less than transform block before dividing height, transform block after division it is high in The height of transform block before dividing；When image block uses horizontal division or vertical division, the wide and height of transform block, and transform block When the average value of sum wide and high is more than the maximum of the size of default transform block, using horizontal division or vertical division so that , less than the width of transform block before dividing, the height of the transform block after division is less than transform block before dividing for the width of the transform block after division It is high.In addition, square dividing mode can be used, transform block is divided.
When image block dividing mode is horizontal division, the width of transform block is equal to the height of the transform block before dividing, and before division Transform block division before average value wide and high be less than or equal to default transform block full-size when, after the division Transform block is converted using non-square, and width of the width equal to the transform block before the division that described non-square is converted, described The height of the height less than the transform block before the division of non-square conversion；
When image block dividing mode is vertical division mode, the width of the transform block before division is equal to the transform block before dividing When the average value wide and high of the transform block before height, and division is less than or equal to default transform block full-size, the division Transform block afterwards is converted using non-square, and the high transform block in before the division of described non-square conversion height, institute State the width of the width less than the transform block before the division of non-square conversion；
When image block dividing mode is horizontal division mode or vertical division mode, the width of the transform block before division is not equal to The average value wide and high of the transform block before the height of the transform block before division, and division is more than default transform block full-size When, the transform block is converted using non-square, and height of the height less than the transform block before the division that described non-square is converted, The width of the width less than the transform block before the division of the non-square conversion.
In each embodiment of the invention described above, when non-square division is carried out to transform block, may also be combined with transform block allows to draw The default division number of plies divided, it is determined whether stopping is divided to transform block.Specifically, the present embodiment is drawn to transform block Timesharing, it is further comprising the steps of：The division number of plies that transform block allows to divide is obtained, wherein, non-square division is carried out to transform block, And in the minimum value of the size in default transform block wide or high of transform block, stop to the division of transform block concretely： Carry out non-square division to transform block, and the division number of plies in transform block divides the number of plies less than default, and transform block wide or height Equal to the size of default transform block minimum value when, stop division to transform block.
Assuming that the size of image block is 2Mx2M, the maximum for presetting the size of transform block is M, the size of default transform block Minimum value be 0.25M, the initial value of the size of transform block is equal to the size of image block, and transform block allows the division number of plies for dividing It it is 4 layers, image block dividing mode is Nx2N.Because the initial value of transform block is 2Mx2M, its size is more than the chi of default transform block Very little maximum, therefore the size of transform block is divided to MxM by 2Mx2M, divides the number of plies and is changed into the 1st layer from the 0th layer；Due to default The size minimum value of transform block is 0.25M, and transform block can continue to divide, and the 2nd layer is divided to by the 1st layer, and its size is become by MxM It is 0.25MxM, the minimum value 0.25M of the size of default transform block is equal to due to the width of transform block, now transform block is not continued to Divide.
Fig. 6 is the transform block division methods schematic flow sheet that the embodiment of the present invention three is provided.Specifically, as shown in fig. 6, originally Embodiment method may include following steps：
The maximum of step 301, the size of the size, the dividing mode of image block and default transform block of acquisition image block Value, the size of described image block is the average value wide and high of image block, and the maximum of the size of the default transform block is change Change the maximum wide or high of block, or the sum wide and high for transform block average value；
Step 302, the size according to image block, and the size of default transform block maximum, obtain the first parameter, institute The first parameter is stated for representing the size of the corresponding maximum transform block of described image block, the size of the transform block is transform block Average value wide and high；
The size of step 303, the dividing mode according to image block, first parameter and the default transform block is most Small value, obtains the second parameter, and second parameter is used to represent whether the size of the corresponding minimum transform block of described image block is big In the minimum value of the size of the default transform block；
Step 304, according to second parameter, judge whether to carry out non-square division to the transform block, be then to perform Step 306, otherwise, performs step 305；
Step 305, the transform block is divided using square dividing mode, until the size of transform block is equal to default change When changing the minimum dimension of block, stop dividing, terminate；
Step 306, non-square division is carried out to the transform block, judge the transform block after dividing it is wide or high one of them be The minimum value of the no size equal to default transform block, is to stop the division to transform block, is terminated, and otherwise, continues executing with step 304, until the transform block after division it is wide or high one of them be equal to default transform block size minimum value.
In the present embodiment, in above-mentioned steps 302, according to the size of image block, and the size of default transform block maximum Value, obtaining the first parameter includes：It is described when the maximum of the size that the size of described image block is more than the default transform block First parameter is equal to the maximum of the size of the default transform block；When the size of described image block is less than the default transform block Size maximum when, first parameter be equal to described image block size；Described in being equal to when the size of described image block During the maximum of the size of default transform block, first parameter is equal to the size of described image block.
In the present embodiment, the size of image block, the size of transform block, the maximum of the size of default transform block and default The minimum value of the size of transform block is adopted with 2 for the logarithm of radix is represented, wherein, in above-mentioned steps 303, state according to image block The minimum value of the size of dividing mode, first parameter and the default transform block, obtaining the second parameter includes：When described The dividing mode of image block is to be divided using non-square, and first parameter value be more than default transform block size minimum When value adds 1, the value of second parameter is set to 1, the value of second parameter is otherwise set to 0.
In the present embodiment, in above-mentioned steps 304, according to second parameter, judge whether to carry out the transform block non- Square division includes：When the value of second parameter is 1, judge to divide the transform block using non-square, otherwise, to institute Transform block is stated using square division；Wherein, the square division refer to by image block or transform block be divided into it is wide and it is high not etc. Division methods.
Below by size for the size partition process of the transform block of the image block of 32x32 as a example by, the embodiment of the present invention More detailed description is done in principle and realization.
If the size of image block is 32x32, the dividing mode of image block is vertical division mode, i.e. image block dividing mode It is Nx2N, nLNx2N, nRx2N one of which, the maximum for presetting the size of transform block is 32, presets the size of transform block most Small value is 4, and the division number of plies that presetting transform block allows is 4 layers.Wherein, the size of image block, the size of transform block, default conversion The maximum and minimum value of the size of block are used with 2 for the logarithm of radix is represented, specifically, the size of image block with Log2CbSize represents that the size of image block is represented by with the relation of log2CbSize：Size=1 of image block<< log2CbSize；The wide and height of transform block represented using parameter log2TrafoHeight and log2TrafoWidth, transform block Height is represented by with the relation of log2TrafoHeight：Height=1 of transform block<<log2TrafoHeight；Transform block it is wide with The relation of log2TrafoWidth is represented by：Transform block it is wide=1<<log2TrafoWidth；The size of default transform block Maximum represents that the maximum of the size of transform block can table with the relation of Log2MaxTrafoSize with Log2MaxTrafoSize It is shown as：Maximum=1 of the size of transform block<<Log2MaxTrafoSize；The minimum value of the size of default transform block with Log2MinTrafoSize represents that the minimum value of the size of transform block is represented by with the relation of Log2MinTrafoSize：Become Change minimum value=1 of the size of block<<Log2MinTrafoSize, therefore, in this example, log2CbSize, The initial value of log2TrafoHeight and log2TrafoWidth is 5, Log2MaxTrafoSize=5, Log2MinTrafoSize=2。
Additionally, the size of transform block can also log2TrafoSize represent, for representing the wide and high average of transform block Value, therefore, the relation of log2TrafoSize and log2TrafoHeight and log2TrafoWidth is represented by： log2TrafoSize=(log2TrafoHeight+log2TrafoHeight)>>1, i.e. log2TrafoSize are The average value of log2TrafoHeight and log2TrafoWidth.
During each size of this example is with the arithmetic operation of exponential representation, "<<" represent shifted left operation, 1<< Log2TrafoHeight is equivalent to 2log2TrafoHeight operations;1<<Log2TrafoWidth is equivalent to 2log2TrafoWidth is operated;1<<Log2TrafoSize is equivalent to 2log2TrafoSize operations；">>" represent right shift Operation, (log2TrafoHeight+log2TrafoHeight)>>1 is equivalent to (log2TrafoHeight+ Log2TrafoHeight)/2 operate.
In this example, when transform block is divided into smaller sub- transform block, the change for converting block size can be by ginseng The plus-minus of number log2TrafoHeight, log2TrafoWidth and log2TrafoSize numerical value is emerged from.For example, current become When the size for changing block is 16x16, the value of log2TrafoHeight is 4 for the value of 4, log2TrafoWidth, The value of log2TrafoSize is 4, and when transform block is divided into the sub- transform block of 16x4, the size of the transform block after division is changed into 16x4, now the value of log2TrafoWidth is constant still for the value of 4, log2TrafoHeight is changed into 2, log2TrafoSize's Value is changed into 3.
In this example, the division of transform block uses one point to be divided for four mode, when transform block is drawn using non-square Timesharing, the length on one side of the transform block after division will be 4 times of another edge lengths, therefore, when carrying out non-side to transform block When shape is divided, will there is following relation in when being wider than high of transform block after division：log2TrafoWidth= Log2TrafoHeight+2, when the width of the transform block after division is less than high, log2TrafoHeight=log2TrafoWidth +2。
When being represented with exponential form due to the size of transform block, the size log2TrafoSize of transform block is width （log2TrafoWidth）And height（log2TrafoHeight）Average value, i.e. log2TrafoSize= (log2TrafoHeight+log2TrafoHeight)>>1.Therefore, when the chi in default transform block wide or high of transform block During very little minimum value, log2TrafoSize=Log2MinTrafoSize+1.
Therefore, during image coding and decoding, non-square division is carried out to transform block, and the value of log2TrafoSize etc. When Log2MinTrafoSize+1, the division to transform block will be stopped；And in the encoding-decoding process of image, it is possible to it is right Transform block carries out square division, and in square division, the value of log2TrafoSize needs to be equal to Log2MinTrafoSize When, transform block can just stop dividing, therefore can introduce variables L og2MinTrafoSizePlus1 in coding and in decoding Judge whether transform block is divided using non-square, when Log2MinTrafoSizePlus1 is equal to 1, transform block can allow to use Non-square is divided, and in coding/decoding system, whether non-square division is carried out to transform block restriction, for example, conversion block size is During certain size, such as 8x8, and transform block minimum dimension be 4 when, cannot again carry out non-square conversion, therefore, meet , it is necessary to when further division is done to transform block, can be divided using non-square in the case of the restriction； When Log2MinTrafoSizePlus1 is equal to 0, conversion does not allow to be divided using non-square, and uses square division, and no matter becomes It is to be divided using non-square or use square division to change block, all in log2TrafoSize=Log2MinTrafoSize+ During Log2MinTrafoSizePlus1, transform block just stops dividing.Wherein, described Log2MinTrafoSizePlus1 is just The second above-mentioned parameter.
In this example, when being divided to transform block, the size for being at least default transform block wide or high of transform block At 4 times of minimum value, just stop dividing transform block.And in the encoding-decoding process of image block, can also pre-set change The maximum of the size of block is changed, therefore, when the size of transform block is more than the maximum of the size of default transform block, transform block is just Needs are divided, and are divided into smaller transform block.
Because the transform block full-size that image block is used is the size equal to image block, therefore, it is determined that transform block is Before no use non-square is divided, in addition it is also necessary to determine the full-size of the corresponding transform block of image block, that is, determine to be drawn using non-square Point the initial division size of transform block.If the initial division size of transform block is Log2MaxTrafoSizeInCu, due to becoming The initial value for changing block is equal to the size log2CbSize of image block, therefore when log2CbSize is more than the size of default transform block During maximum Log2MaxTrafoSize, the value of Log2MaxTrafoSizeInCu is equal to Log2MaxTrafoSize；Otherwise, When log2CbSize is equal to or less than the maximum Log2MaxTrafoSize of the size of default transform block, The value of Log2MaxTrafoSizeInCu is equal to log2CbSize, can be expressed as follows with mathematical form：
Log2MaxTrafoSizeInCu=log2CbSize>Log2MaxTrafoSize?Log2MaxTrafoSize: log2CbSize。
It can thus be appreciated that the original dimension of transform block be at least the minimum value of the size of default transform block 4 times are represented by Log2MaxTrafoSizeInCu>(Log2MinTrafoSize+1), i.e. the value of Log2MaxTrafoSizeInCu is at least (Log2MinTrafoSize+1), because the parameters such as Log2MaxTrafoSizeInCu, Log2MinTrafoSize are with index Form is represented, it is known that now the original dimension of transform block is at least 4 times of the minimum value of the size of default transform block.
Using the precondition that non-square is divided it is that image block employs horizontal division mode or vertical draws due to transform block The mode of dividing.In the present invention except can be with use level dividing mode（2NxN,2NxnU,2NxnD）With vertical division mode （Nx2N,nLx2N,nRx2N）Using the square dividing mode of 2Nx2N and NxN, therefore PART2Nx2N, PART NxN can also be set Representative image block uses 2Nx2N and NxN dividing modes respectively, and transform block is divided using non-square when following condition meets（I.e. The value of Log2MaxTrafoSizeInCu is 1）, otherwise transform block do not use non-square to divide（I.e. The value of Log2MaxTrafoSizeInCu is 0）：
if(nsrqt_enabled_flag&&Log2MaxTrafoSizeInCu>(Log2MinTrafoSize+1)&& PartMode!=PART_2NX2N&&PartMode!=PART_NXN)
Wherein nsrqt enabled flag are the flag bit for allowing to use non-square to divide in encoding-decoding process.
When the size of transform block is more than 4 times of the minimum value of the size of default transform block（That is log2TrafoSize> (Log2MinTrafoSize+Log2MinTrafoSizePlus1)）, Image Coding end need to write transform block and divide parameter, decoding End needs the division parameter of analytic trnasformation block；Otherwise（That is log2TrafoSize==(Log2MinTrafoSize+ Log2MinTrafoSizePlus1)）Coding side is not written into the division parameter of transform block, the division of decoding end not decoded transform block Parameter.Coding side and decoding end embodiment will be described in detail to the coding and decoding that transform block divides parameter below.
Be can be seen that by the second parameter by the example, it is possible to whether determine transform block using square division, and really The size of the transform block after fixed division.It will be understood by those skilled in the art that the second parameter is used for ease of Image Coding An intermediate variable, second parameter can be as needed obtained in practical application, however it is not limited to the second parameter described above Acquisition methods.
Refer to the division to the luminance block of image block, practical application to the division of transform block in the embodiments of the present invention In, the chrominance block in image block is the half of luminance block, therefore, when the chrominance block in image block is divided, also include Following steps：
Obtain the size of the maximum transform block of the corresponding chrominance block of image block；
Whether the size of the corresponding transform block of chrominance block exceedes the corresponding chrominance block of described image block in the transform block of judgement Maximum transform block size, be that the corresponding transform size of chrominance block in the transform block after the division is reset into the figure As the size of the maximum transform block of the corresponding chrominance block of block.
It will be understood by those skilled in the art that the size of the maximum transform block of above-mentioned image block correspondence chrominance block, is exactly The corresponding maximum transform block of image block is the half of the size that image block allows the maximum transform block of the size for using, and works as image When the size of block is less than or equal to the maximum of the size of default transform block, the size of the corresponding maximum transform block of image block is just The size of image block, conversely, the maximum of its size for being equal to default transform block.
Illustrated for ease of being divided to chrominance block, below with video coding and decoding system 4:2:The volume of the video of 0 form In decoding process, the division of chrominance block is illustrated.Wherein, described 4:2:0 form refers to the size of chromatic component in video It is the half of luminance component, if the size of such as video is 416x240, the size of brightness is 416x240, color in piece image The size of degree is 208x120.It is to be encoded in units of block due to video image, therefore 4:2:Color during 0 said shank The size for spending block is always the half of brightness block size, and the size of such as luminance block is 32x32, and the size of chrominance block is 16x16. In above-mentioned transform block encoding-decoding process, the conversion block size of colourity and the conversion block size of brightness follow same relation, Such as when the conversion block size of brightness is 8x8, the conversion block size of colourity is 4x4.
Due to the minimum value of the size of transform block would generally be preset in coding/decoding system, the size of such as default transform block Minimum value is 4, therefore is possible to that colourity can be caused when the conversion block size of colourity halves according to the conversion block size of brightness Conversion block size is less than the minimum value of the size of default transform block, such as when the conversion block size of brightness is 4x4, the change of colourity Block size is changed for 2x2, now the minimum value 4 of the size of the conversion block size already less than default transform block of colourity.In this feelings , it is necessary to be reset to the conversion block size of colourity under condition, the conversion block size of colourity is reset into the transform block chi with brightness It is very little consistent, will its size reset to 4x4.
When transform block is divided using non-square, several extra situations can be further added by.If the size of default transform block Minimum value is 4.Understand when transform block is divided using non-square, minimum conversion block size corresponding to it be 16x4 or 4x16.When the conversion block size of brightness is 16x4 or 4x16, the conversion block size of colourity is 8x2 or 2x8, now needs to check colors The conversion block size of degree is reset.But chrominance block allows the conversion that uses during reset process also needs to consider current image block Block size.
Specifically, when image block size is 32x32, and default maximum conversion block size is 32x32, image block is adopted Use horizontal division mode.Now, the adoptable maximum conversion block size of luminance block is 32, and the change block corresponding to it can be from 32x32 is divided to 16x16 and is further divided to 16x4.The adoptable maximum conversion block size of chrominance block is 16, due to color The conversion block size of degree is halved by the conversion block size of brightness all the time, therefore change block corresponding to it can be divided to from 16x16 16x4 is further divided to 8x2.The minimum value of the size due to 8x2 already less than default transform block is, it is necessary to by its transform block Size is reset.Because the adoptable maximum conversion block size of chrominance block is 16, now can be by the conversion block size of colourity Reset to 16x4.
When image block size is 16x16, and default maximum conversion block size is 32x32, and image block uses level Dividing mode.Now, the adoptable maximum conversion block size of luminance block is 16, and the change block corresponding to it can be drawn from 16x16 Divide to 16x4.The adoptable maximum conversion block size of chrominance block is 8, because the conversion block size of colourity is all the time by the change of brightness Block size is changed to halve, therefore change block corresponding to it can be divided to 8x2 from 8x8.Because 8x2 is already less than default transform block The minimum value of size is reset, it is necessary to be converted block size.Because the adoptable maximum conversion block size of chrominance block is 8, now the conversion block size of colourity cannot be reset to 16x4, therefore the size of chrominance block can only be reset to 8x8.
In sum, it is known that 4:2:In the case of 0 said shank, if transform block uses non-square dividing mode, colourity Conversion block size there is following extra judgement step：Judge whether the size of transform block has reached the transform block chi of minimum It is very little；If the size of transform block has reached the conversion block size of minimum, the corresponding maximum transform block chi of current image block is obtained It is very little；Conversion block size according to chrominance block judges whether its size can exceed the corresponding maximum transform block chi of current image block It is very little, its size is reset into default value if exceeding.
In the present embodiment, in view of when transform block is divided, as long as transform block it is wide or high one of them be equal to default transform block Size minimum value, just stop division to transform block, in video coding and decoding system, can be according to the size of transform block and institute Relation between the minimum value of the size for stating default transform block, is encoded or is decoded to the division parameter of transform block.Below will It is illustrated respectively with different embodiments.
Fig. 7 is the coding method schematic flow sheet that the transform block that the embodiment of the present invention four is provided divides parameter.The present embodiment Coding side that can be in video coding and decoding system, the division parameter to transform block is encoded, specifically, as shown in fig. 7, this reality Applying example may include following steps：
Step 401, the minimum value for obtaining the size for presetting transform block；
Relation between the minimum value of the size of step 402, the size according to transform block and the default transform block, judges Whether the division parameter of transcoding, coding transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
In the present embodiment, above-mentioned steps 402 specifically may include：When transform block is divided using non-square, and the transform block The size in the default transform block wide or high minimum value when, judge do not need transcoding, coding transform block division parameter, it is no Then, the division parameter of transcoding, coding transform block.
In the present embodiment, above-mentioned steps 402 may also comprise：According to the second parameter, judge whether the division of transcoding, coding transform block Parameter, wherein, whether second parameter is used to represent the size of the corresponding minimum transform block of image block more than the default change Change the minimum value of the size of block.
Wherein, the second parameter can be obtained by following steps：
Obtain the maximum of the size of the size, the dividing mode of image block and default transform block of image block；
According to the size of image block, and the size of default transform block maximum, obtain the first parameter, first ginseng Number is used to represent the size of the corresponding maximum transform block of described image block；
The minimum value of the size of dividing mode, first parameter and the default transform block according to image block, obtains Obtain the second parameter.
In the present embodiment, according to the second parameter, judging whether the division parameter of transcoding, coding transform block includes：
When the minimum value that the size of the transform block is more than the size of default transform block adds second parameter, coding becomes The division parameter of block is changed, otherwise, it is not necessary to the division parameter of transcoding, coding transform block.
The acquisition of the second parameter and the first parameter and the invention described above transform block division methods embodiment in the present embodiment In it is identical, will not be repeated here.
In coding/decoding system, it will usually be provided with the maximum of the division number of plies that transform block can be divided, that is, preset and divide The number of plies, therefore, when the division parameter to transform block is encoded, can also include the steps of：Obtaining transform block allows to divide The default division number of plies；When the division number of plies of transform block is equal to the default division number of plies, judge not needing transcoding, coding transform block Division parameter.In transform block cataloged procedure, as long as the size of transform block or division number of plies either condition meet above-mentioned requirements When, i.e., division parameter that need not be again to transform block is encoded, i.e., transform block need not divide parameter, represents that transform block is not required to Divide.
Correspondingly, the embodiment of the present invention also provides the coding/decoding method that a kind of transform block divides parameter.
Fig. 8 is the schematic flow sheet of the coding/decoding method that the transform block that the embodiment of the present invention five is provided divides parameter.Such as Fig. 8 institutes Show, the coding/decoding method of the present embodiment transform block may include following steps：
Step 501, the minimum value for obtaining the size for presetting transform block；
Relation between the minimum value of the size of step 502, the size according to transform block and the default transform block, judges Whether the division parameter of decoded transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
The present embodiment coding/decoding method is actually process corresponding with above-mentioned coding method, specifically, above-mentioned steps 502 In, the relation between the minimum value of the size of size and the default transform block according to transform block judges whether decoded transform The division parameter of block may include：
When transform block using non-square convert, and the transform block the size in default transform block wide or high most During small value, judge not needing the division parameter of decoded transform block, otherwise, the division parameter of decoded transform block.
Additionally, in actual decoding process, above-mentioned steps 502 specifically may include：According to the second parameter, judge whether that decoding becomes The division parameter of block is changed, whether second parameter is used to represent the size of the corresponding minimum transform block of image block more than described pre- If the minimum value of the size of transform block.
The second above-mentioned parameter can be obtained by following steps：Obtain the size of image block, the dividing mode of image block with And the maximum of the size of default transform block；According to the size of image block, and the size of default transform block maximum, obtain First parameter, first parameter is used to represent the size of the corresponding maximum transform block of described image block；According to drawing for image block Divide the minimum value of the size of mode, first parameter and the default transform block, obtain the second parameter.It is described according to second Parameter, judging whether the division parameter of decoded transform block includes：When the size of the transform block is more than the size of default transform block Minimum value when adding second parameter, the division parameter of decoded transform block.
Additionally, the present embodiment coding/decoding method may also include：
Obtain the default division number of plies that transform block allows to divide；
When the division number of plies of transform block is equal to the default division number of plies, judge not needing the division ginseng of decoded transform block Number.
In decoding process, when the division number of plies of transform block has reached the default division number of plies, you can stop to transform block Divide the decoding of parameter.
One of ordinary skill in the art will appreciate that：Realizing all or part of step of above method embodiment can pass through Programmed instruction related hardware is completed, and foregoing program can be stored in a computer read/write memory medium, the program Upon execution, the step of including above method embodiment is performed；And foregoing storage medium includes：ROM, RAM, magnetic disc or light Disk etc. is various can be with the medium of store program codes.
Fig. 9 is that the transform block that the embodiment of the present invention six is provided divides apparatus structure schematic diagram.As shown in figure 9, the present embodiment Transform block divides device and may include：
Dimension acquisition module 1, the minimum value of the size for obtaining default transform block；
Transform block division module 2, for transform block carry out non-square divide when, if the transform block is wide or high In the minimum value of the size of the default transform block, then stop dividing the transform block；
Wherein, the minimum value of the size of the default transform block is the minimum value wide or high of transform block, or wide and high The minimum value of the average value of sum；It refers to that image block or transform block are divided into wide and division not etc. high that the non-square is divided Method.
The step of the present embodiment transform block divides device based on the invention described above transform block division methods, realizes to transform block Division, it implements process reference can be made to the explanation of the invention described above embodiment of the method, will not be repeated here.
Figure 10 is that the transform block that the embodiment of the present invention seven is provided divides apparatus structure schematic diagram.As shown in Figure 10, above-mentioned On the basis of diagram shown in Fig. 9, the present embodiment may also include：
Judge module 3 is divided, for judging whether to carry out non-square division to the transform block.
As shown in Figure 10, the division judge module 3 specifically may include：
Dividing mode acquiring unit 31, the dividing mode for obtaining image block；
First judging unit 32, for using horizontal division when image block, the width of the transform block is equal to the height of transform block, The average value of the sum wide and high of the transform block is less than or equal to the maximum of the size of default transform block, and the transform block 4 times of minimum value of the size for being at least the default transform block wide or high when, judge to use non-square to the transform block Divide so that the width of the transform block after division is equal to the width of transform block before dividing, before the height of the transform block after division is less than dividing The height of transform block；
Second judging unit 33, for using vertical division when image block, the width of the transform block is equal to the height of transform block, The average value of the sum wide and high of the transform block is less than the maximum of the size of default transform block, and the width of the transform block or When height is at least 4 times of the minimum value of the size of the default transform block, judge to divide the transform block using non-square, make The width of the transform block after must dividing is less than the height of transform block before dividing, the high transform block before dividing of the transform block after division It is high；
3rd judging unit 34, for using horizontal division or vertical division, the wide and height of the transform block when image block , the maximum of the size of the average value more than default transform block of the sum wide and high of the transform block, and the transform block 4 times of minimum value of the size for being at least the default transform block wide or high when, judge to use non-square to the transform block Divide so that the width of the transform block after division is less than before dividing less than the width for dividing preceding transform block, the height of the transform block after division The height of transform block.
Figure 11 is that the transform block that the embodiment of the present invention eight is provided divides apparatus structure schematic diagram.Implement with shown in above-mentioned Figure 10 Unlike example, as shown in figure 11, in the present embodiment, described division judge module 3 is specifically included：
First acquisition unit 35, for obtaining the size of image block, the dividing mode of image block and default transform block The maximum of size, the size of described image block is the average value wide and high of image block, the size of the default transform block Maximum for transform block maximum wide and high, or the average value of sum wide and high minimum value；
First parameter acquiring unit 36, for the size according to image block, and the size of default transform block maximum, The first parameter is obtained, first parameter is used to represent the size of the corresponding maximum transform block of described image block, the transform block Size for transform block average value wide and high；
Second parameter acquiring unit 37, for the dividing mode according to image block, first parameter and described default The minimum value of the size of transform block, obtains the second parameter, and second parameter is used to represent the corresponding minimum change of described image block Change block size whether more than the default transform block size minimum value；
Judging unit 38 is divided, for according to second parameter, judging whether that non-square is carried out to the transform block draws Point.
In the present embodiment, the first parameter acquiring unit 36 is default more than described specifically for the size when described image block During the maximum of the size of transform block, first parameter is equal to the maximum of the size of the default transform block；When the figure During the maximum of the size that the size of picture block is less than the default transform block, first parameter is equal to the chi of described image block It is very little；And when the maximum of the size that the size of described image block is equal to the default transform block, first parameter is equal to The size of described image block.
In the present embodiment, the size of image block, the size of transform block, the maximum of the size of default transform block and default The minimum value of the size of transform block is adopted with 2 as the logarithm of radix is represented；And during each size exponential representation, the second above-mentioned ginseng Number acquiring unit 37, is particularly used in when the dividing mode of described image block is to be divided using non-square, and first parameter The minimum value of the value size that is more than default transform block when adding 1, the value of second parameter is set to 1, otherwise by described second The value of parameter is set to 0；Above-mentioned division judging unit 38, is particularly used in when the value of second parameter is 1, judges to allow The transform block is divided using non-square, otherwise, square division is used to the transform block；
Wherein, the square division refers to that image block or transform block are divided into wide and division methods not etc. high.
Figure 12 is the structural representation that the transform block that the embodiment of the present invention nine is provided divides device.In above-mentioned Fig. 9, Tu10Huo On the basis of embodiment illustrated in fig. 11, as shown in figure 12, the present embodiment may also include：
Parameter acquisition module 4 is divided, the default division number of plies divided for obtaining transform block to allow；
The transform block division module 2, being particularly used in carries out non-square division to transform block, and in the transform block Divide the number of plies and divide the number of plies less than default, and during the minimum value of the size in default transform block wide or high of the transform block, Stop the division to the transform block.
Additionally, as shown in figure 12, the present embodiment may also include：
Chromaticity transformation block size acquisition module 5, the size of the maximum transform block for obtaining the corresponding chrominance block of image block；
Chromaticity transformation block resets module 6, the size for judging the corresponding transform block of chrominance block in the transform block after dividing Whether exceed the size of the maximum transform block of the corresponding chrominance block of described image block, be then by color in the transform block after the division The corresponding transform size of degree block resets to the size of the maximum transform block of the corresponding chrominance block of described image block.
The invention described above transform block divides each embodiment of device and implements process reference can be made to the invention described above transform block is drawn Divide the explanation of embodiment of the method, will not be repeated here.
Figure 13 is the structural representation of the code device that the transform block that the embodiment of the present invention ten is provided divides parameter.Such as Figure 13 Shown, the present embodiment code device includes：
Dimension acquisition module 10, the minimum value of the size for obtaining default transform block；
Parameter coding module 20, between the minimum value of the size according to transform block and the size of the default transform block Relation, judge whether the division parameter of transcoding, coding transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
In the present embodiment, above-mentioned parameter coding module 20 is particularly used in when transform block is divided using non-square, and institute During the minimum value of the size in the default transform block wide or high for stating transform block, judge not needing the division of transcoding, coding transform block Parameter, otherwise, the division parameter of transcoding, coding transform block.
Figure 14 is the structural representation of the code device that the transform block that the embodiment of the present invention 11 is provided divides parameter.This reality Apply in example, parameter coding module 20 is particularly used according to the second parameter, judge whether the division parameter of transcoding, coding transform block.Tool Body, as shown in figure 14, the present embodiment may also include：
Second acquisition unit 210, for obtaining the size of image block, the dividing mode of image block and default transform block The maximum of size；
First parameter acquiring unit 211, for the size according to image block, and the size of default transform block maximum Value, obtains the first parameter；
Second parameter acquiring unit 212, for the dividing mode according to image block, first parameter and described default The minimum value of the size of transform block, obtains the second parameter.
In the present embodiment, whether second parameter is used to represent the size of the corresponding minimum transform block of image block more than institute State the minimum value of the size of default transform block；First parameter is used to represent the chi of the corresponding maximum transform block of described image block It is very little.
In the present embodiment, above-mentioned parameter coding module 20 is particularly used in the size when the transform block more than default When the minimum value of the size of transform block adds second parameter, the division parameter of transcoding, coding transform block, otherwise, it is not necessary to transcoding, coding transform The division parameter of block.
Figure 15 is the structural representation of the code device that the transform block that the embodiment of the present invention 12 is provided divides parameter.Upper State on the basis of Figure 13 or embodiment illustrated in fig. 14, as shown in figure 14, the present embodiment may also include：
Parameter acquisition module 30 is divided, the default division number of plies divided for obtaining transform block to allow；
The parameter coding module 20, specifically for when transform block the division number of plies be equal to it is described it is default division the number of plies when, Judgement does not need the division parameter of transcoding, coding transform block.
The code device that the invention described above transform block divides parameter implements process reference can be made to the invention described above transform block The coding method embodiment of parameter is divided, be will not be repeated here.
Figure 16 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 13 is provided divides parameter.As schemed Shown in 16, the present embodiment decoding apparatus may include：
Dimension acquisition module 40, the minimum value of the size for obtaining default transform block；
Parameter decoding module 50, between the minimum value of the size according to transform block and the size of the default transform block Relation, judge whether the division parameter of decoded transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided.
In the present embodiment, above-mentioned parameter decoding module 50 is particularly used in when transform block is converted using non-square, and institute During the minimum value of the size in default transform block wide or high for stating transform block, judge not needing the division ginseng of decoded transform block Number, otherwise, the division parameter of decoded transform block.
Figure 17 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 14 is provided divides parameter.With it is upper State unlike embodiment illustrated in fig. 16 technical scheme, in the present embodiment, parameter decoding module 50, specifically for according to the second ginseng Number, judges whether the division parameter of decoded transform block.Specifically, as shown in figure 17, the present embodiment may also include：
Second acquisition unit 510, for obtaining the size of image block, the dividing mode of image block and default transform block The maximum of size；
First parameter acquiring unit 511, for the size according to image block, and the size of default transform block maximum Value, obtains the first parameter；
Second parameter acquiring unit 512, for the dividing mode according to image block, first parameter and described default The minimum value of the size of transform block, obtains the second parameter.
In the present embodiment, the second above-mentioned parameter is used to represent whether the size of the corresponding minimum transform block of image block is more than The minimum value of the size of the default transform block；The first above-mentioned parameter is used to represent the corresponding maximum transform block of described image block Size.Wherein, parameter decoding module 50, the size for presetting transform block is more than most specifically for the size when the transform block When small value plus second parameter, the division parameter of decoded transform block.
Figure 18 is the structural representation of the decoding apparatus that the transform block that the embodiment of the present invention 15 is provided divides parameter.Upper State on the basis of Figure 16 or embodiment illustrated in fig. 17 technical scheme, as shown in figure 18, the present embodiment may also include：
Parameter acquisition module 60 is divided, the default division number of plies divided for obtaining transform block to allow；
The parameter decoding module 50, is particularly used in when the division number of plies of transform block is equal to the default division number of plies When, judge not needing the division parameter of decoded transform block.
The decoding apparatus that the invention described above transform block divides parameter implement process reference can be made to the invention described above transform block The coding/decoding method embodiment of parameter is divided, be will not be repeated here.
Figure 19 is the structural representation of the Image Codec that the embodiment of the present invention 16 is provided.As shown in figure 19, originally Embodiment Image Codec includes：
Transform block divides the code device 100 of parameter, for the chi of the size according to transform block and the default transform block Relation between very little minimum value, judges whether the division parameter of transcoding, coding transform block；
Transform block divides the decoding apparatus 200 of parameter, for the chi of the size according to transform block and the default transform block Relation between very little minimum value, judges whether the division parameter of decoded transform block；
Wherein, the division parameter of the transform block is used to represent and transform block is divided or is not divided
In the present embodiment, the code device 100 that transform block divides parameter can specifically be provided using the embodiments of the present invention Transform block divide parameter code device, specifically refer to the explanation of Figure 13, Figure 14 or embodiment illustrated in fig. 15；Transform block is drawn The transform block that dividing the decoding apparatus 200 of parameter can specifically be provided using the embodiments of the present invention divides the decoding apparatus of parameter, The explanation of Figure 16, Figure 17 or embodiment illustrated in fig. 18 is specifically referred to, be will not be repeated here.
It will be understood by those skilled in the art that the present embodiment Image Codec carries out its in cataloged procedure to image His processing procedure, for example, decomposed to image block, image block quantified and the process such as entropy code and conventional images are compiled Decoding systematic realizing program is same or like, will not be repeated here.
Finally it should be noted that：Various embodiments above is merely illustrative of the technical solution of the present invention, rather than its limitations；To the greatest extent Pipe has been described in detail with reference to foregoing embodiments to the present invention, it will be understood by those within the art that：Its according to The technical scheme described in foregoing embodiments can so be modified, or which part or all technical characteristic are entered Row equivalent；And these modifications or replacement, the essence of appropriate technical solution is departed from various embodiments of the present invention technology The scope of scheme.

Claims (40)

1. a kind of transform block division methods, it is characterised in that including：
Obtain the minimum value of the size of default transform block；
To transform block when carrying out non-square division, if the size in the default transform block wide or high of the transform block Minimum value, then stop dividing the transform block；
Wherein, it refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided；
Before non-square division is carried out to transform block, also include：Judge whether to carry out non-square division to the transform block；
It is described to judge whether that non-square division is carried out to the transform block to be included：
Obtain the dividing mode of image block；
When image block uses horizontal division, the width of the transform block is equal to the height of transform block, the sum wide and high of the transform block Average value be less than or equal to default transform block size maximum, and the transform block it is wide or high be at least it is described default When the minimum value of the size of transform block adds 2, judge to divide the transform block using non-square so that transform block after division Width is equal to the width of transform block before dividing, the height of the height less than transform block before dividing of the transform block after division；
When image block uses vertical division, the width of the transform block is equal to the height of transform block, the sum wide and high of the transform block Average value be less than default transform block size maximum, and the wide or high of the transform block is at least the default transform block The minimum value of size when adding 2, judge to divide the transform block using non-square so that the width of the transform block after division is less than The width of transform block before dividing, the height of the high transform block before dividing of the transform block after division；
When image block uses horizontal division or vertical division, wide and high, the wide and height of the transform block of the transform block The average value of sum is more than the maximum of the size of default transform block, and the wide or high of the transform block is at least the default change Change the minimum value of the size of block plus when 2, judge to divide the transform block using non-square so that the width of the transform block after division Less than the width of transform block before dividing, the height of the height less than transform block before dividing of the transform block after division；
Wherein, the size of the size of transform block and default transform block with 2 for the logarithm of radix is represented.
2. transform block division methods according to claim 1, it is characterised in that the non-square is divided to be included：2NxN draws Divide method, or 2NxnU division methods, or 2NxnD division methods, or Nx2N division methods, or nLx2N division methods, or nRx2N Division methods.
3. transform block division methods according to claim 1, it is characterised in that also include：
Obtain the default division number of plies that transform block allows to divide；
It is described when carrying out non-square to transform block and divide, if the size in default transform block wide or high of the transform block Minimum value, then stop carrying out division to the transform block including：
Carry out non-square division to transform block, and the division number of plies in the transform block divides the number of plies, and the change less than default During the minimum value of the size in default transform block wide or high for changing block, stop the division to the transform block.
4. transform block division methods according to claim 1, it is characterised in that also include：
Obtain the size of the maximum transform block of the corresponding chrominance block of image block；
Whether the size of the corresponding transform block of chrominance block exceedes the corresponding chrominance block of described image block most in the transform block of judgement The size of big transform block, is that the size of the corresponding transform block of the chrominance block is reset into the corresponding chrominance block of described image block Maximum transform block size.
5. a kind of transform block division methods, it is characterised in that including：
Obtain the minimum value of the size of default transform block；
To transform block when carrying out non-square division, if the size in the default transform block wide or high of the transform block Minimum value, then stop dividing the transform block；
Wherein, it refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided；
Before non-square division is carried out to transform block, also include：Judge whether to carry out non-square division to the transform block；
It is described to judge whether that non-square division is carried out to the transform block to be included：
Obtain the maximum of the size of the size, the dividing mode of image block and default transform block of image block, described image block Size be the average value wide and high of image block, the maximum of the size of the default transform block is the wide and high of transform block Maximum, or the average value of sum wide and high maximum；
According to the size of image block, and the size of default transform block maximum, obtain the first parameter, first parameter is used In the size for representing the corresponding maximum transform block of described image block；
The minimum value of the size of dividing mode, first parameter and the default transform block according to image block, obtains the Whether two parameters, second parameter is used to represent the size of the corresponding minimum transform block of described image block more than the default change Change the minimum value of the size of block；
According to second parameter, judge whether to carry out non-square division to the transform block.
6. transform block division methods according to claim 5, it is characterised in that the size according to image block, and The maximum of the size of default transform block, obtains the first parameter, and first parameter is used to represent that described image block is corresponding most The size of big transform block includes：
When the maximum of the size that the size of described image block is more than the default transform block, first parameter is equal to described The maximum of the size of default transform block；
When the maximum of the size that the size of described image block is less than the default transform block, first parameter is equal to described The size of image block；
When the maximum of the size that the size of described image block is equal to the default transform block, first parameter is equal to described The size of image block, or first parameter is equal to the maximum of the size of the default transform block.
7. transform block division methods according to claim 5 or 6, it is characterised in that the size of described image block, default become The minimum value for changing the maximum of the size of block and the size of default transform block is adopted with 2 as the logarithm of radix is represented；
The minimum value of the size of the dividing mode, first parameter and the default transform block according to image block, obtains Obtaining the second parameter includes：
When the dividing mode of described image block is to be divided using non-square, and the value of first parameter is more than default transform block When the minimum value of size adds 1, the value of second parameter is set to 1, the value of second parameter is otherwise set to 0；
It is described according to second parameter, judge whether to carry out the transform block non-square and divide to include：
When the value of second parameter is 1, judgement allows to divide the transform block using non-square, otherwise, to the change Block is changed using square division；
Wherein, it refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided.
8. transform block division methods according to claim 5, it is characterised in that the non-square is divided to be included：2NxN draws Divide method, or 2NxnU division methods, or 2NxnD division methods, or Nx2N division methods, or nLx2N division methods, or nRx2N Division methods.
9. transform block division methods according to claim 5, it is characterised in that also include：
Obtain the default division number of plies that transform block allows to divide；
It is described when carrying out non-square to transform block and divide, if the size in default transform block wide or high of the transform block Minimum value, then stop carrying out division to the transform block including：
Carry out non-square division to transform block, and the division number of plies in the transform block divides the number of plies, and the change less than default During the minimum value of the size in default transform block wide or high for changing block, stop the division to the transform block.
10. transform block division methods according to claim 5, it is characterised in that also include：
Obtain the size of the maximum transform block of the corresponding chrominance block of image block；
Whether the size of the corresponding transform block of chrominance block exceedes the corresponding chrominance block of described image block most in the transform block of judgement The size of big transform block, is that the size of the corresponding transform block of the chrominance block is reset into the corresponding chrominance block of described image block Maximum transform block size.
A kind of 11. transform blocks divide the coding method of parameter, it is characterised in that including：
Obtain the minimum value of the size of default transform block；
Relation between the minimum value of the size of size and the default transform block according to transform block, judges whether transcoding, coding transform The division parameter of block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided；
Relation between the minimum value of the size of the size and the default transform block according to transform block, judges whether coding The division parameter of transform block includes：
According to the second parameter, judge whether the division parameter of transcoding, coding transform block；
Also include：
Obtain the maximum of the size of the size, the dividing mode of image block and default transform block of image block；
According to the size of image block, and the size of default transform block maximum, obtain the first parameter；
The minimum value of the size of dividing mode, first parameter and the default transform block according to image block, obtains the Two parameters.
12. transform blocks according to claim 11 divide the coding method of parameter, it is characterised in that described according to transform block Size and the default transform block size minimum value between relation, judge whether the division parameter bag of transcoding, coding transform block Include：
When transform block using non-square divide, and the size in the default transform block wide or high of the transform block minimum During value, judge not needing the division parameter of transcoding, coding transform block, otherwise, the division parameter of transcoding, coding transform block.
13. transform blocks according to claim 11 divide the coding method of parameter, it is characterised in that second parameter is used Whether it is more than the minimum value of the size of the default transform block in the size for representing the corresponding minimum transform block of image block；
First parameter is used to represent the size of the corresponding maximum transform block of described image block.
14. transform blocks according to claim 13 divide the coding method of parameter, it is characterised in that described according to image block Size, and the size of default transform block maximum, obtain the first parameter include：
When the maximum of the size that the size of described image block is more than the default transform block, first parameter is equal to described The maximum of the size of default transform block；
When the maximum of the size that the size of described image block is less than the default transform block, first parameter is equal to described The size of image block；
When the maximum of the size that the size of described image block is equal to the default transform block, first parameter is equal to described The size of image block, or first parameter is equal to the maximum of the size of the default transform block.
15. transform blocks according to claim 13 divide the coding method of parameter, it is characterised in that described according to the second ginseng Number, judging whether the division parameter of transcoding, coding transform block includes：
When the minimum value that the size of the transform block is more than the size of default transform block adds second parameter, transcoding, coding transform block Division parameter, otherwise, it is not necessary to the division parameter of transcoding, coding transform block.
16. transform blocks according to claim 11 divide the coding method of parameter, it is characterised in that also include：
Obtain the default division number of plies that transform block allows to divide；
The coding method that the transform block divides parameter also includes：
When the division number of plies of transform block is equal to the default division number of plies, judge not needing the division parameter of transcoding, coding transform block.
A kind of 17. transform blocks divide the coding/decoding method of parameter, it is characterised in that including：
Obtain the minimum value of the size of default transform block；
Relation between the minimum value of the size of size and the default transform block according to transform block, judges whether decoded transform The division parameter of block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided；
Relation between the minimum value of the size of the size and the default transform block according to transform block, judges whether decoding The division parameter of transform block includes：
According to the second parameter, judge whether the division parameter of decoded transform block.
18. transform blocks according to claim 17 divide the coding/decoding method of parameter, it is characterised in that also include：
Obtain the maximum of the size of the size, the dividing mode of image block and default transform block of image block；
According to the size of image block, and the size of default transform block maximum, obtain the first parameter；
The minimum value of the size of dividing mode, first parameter and the default transform block according to image block, obtains the Two parameters.
19. transform blocks according to claim 18 divide the coding/decoding method of parameter, it is characterised in that second parameter is used Whether it is more than the minimum value of the size of the default transform block in the size for representing the corresponding minimum transform block of image block；
First parameter is used to represent the size of the corresponding maximum transform block of described image block.
20. transform blocks according to claim 19 divide the coding/decoding method of parameter, it is characterised in that described according to image block Size, and the size of default transform block maximum, obtain the first parameter include：
When the maximum of the size that the size of described image block is more than the default transform block, first parameter is equal to described The maximum of the size of default transform block；
When the maximum of the size that the size of described image block is less than the default transform block, first parameter is equal to described The size of image block；
When the maximum of the size that the size of described image block is equal to the default transform block, first parameter is equal to described The size of image block, or first parameter is equal to the maximum of the size of the default transform block.
21. transform blocks according to claim 19 divide the coding/decoding method of parameter, it is characterised in that described according to the second ginseng Number, judging whether the division parameter of decoded transform block includes：
When the minimum value that the size of the transform block is more than the size of default transform block adds second parameter, decoded transform block Division parameter.
22. transform blocks according to claim 17 divide the coding/decoding method of parameter, it is characterised in that also include：
Obtain the default division number of plies that transform block allows to divide；
The coding/decoding method that the transform block divides parameter also includes：
When the division number of plies of transform block is equal to the default division number of plies, judge not needing the division parameter of decoded transform block.
A kind of 23. transform blocks divide device, it is characterised in that including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Transform block division module, for transform block carry out non-square divide when, if the transform block is wide or high in institute The minimum value of the size of default transform block is stated, then stops dividing the transform block；
Wherein, the minimum value of the size of the default transform block is the minimum value wide or high of transform block, or sum wide and high Average value minimum value；It refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided
Also include：
Judge module is divided, for judging whether to carry out non-square division to the transform block；
The division judge module includes：
Dividing mode acquiring unit, the dividing mode for obtaining image block；
First judging unit, for using horizontal division when image block, the width of the transform block is equal to the height of transform block, the change Change the sum wide and high of block average value be less than or equal to default transform block size maximum, and the width of the transform block or When the minimum value that height is at least the size of the default transform block adds 2, judge to divide the transform block using non-square so that The width of the transform block after division is equal to the width of transform block before dividing, and the height of the transform block after division is less than transform block before dividing It is high；
Second judging unit, for using vertical division when image block, the width of the transform block is equal to the height of transform block, the change Change the sum wide and high of block average value be less than default transform block size maximum, and the transform block wide or height at least When adding 2 for the minimum value of the size of the default transform block, judge to divide the transform block using non-square so that after division Transform block width of the width less than transform block before dividing, the height of the high transform block before dividing of transform block after division；
3rd judging unit, for using horizontal division or vertical division, wide and high, the institute of the transform block when image block State the sum wide and high of transform block average value be more than default transform block size maximum, and the transform block wide or height When the minimum value of the size of at least described default transform block adds 2, judge to divide the transform block using non-square so that draw The width of the transform block after point is less than the width for dividing preceding transform block, the height of the height less than transform block before dividing of the transform block after division；
Wherein, the size of the size of transform block and default transform block with 2 for the logarithm of radix is represented.
24. transform blocks according to claim 23 divide device, it is characterised in that also include：
Parameter acquisition module is divided, the default division number of plies divided for obtaining transform block to allow；
The transform block division module, specifically for carrying out non-square division to transform block, and in the division layer of the transform block Number divides the number of plies less than default, and during the minimum value of the size in default transform block wide or high of the transform block, it is right to stop The division of the transform block.
25. transform blocks according to claim 23 divide device, it is characterised in that also include：
Chromaticity transformation block size acquisition module, the size of the maximum transform block for obtaining the corresponding chrominance block of image block；
Chromaticity transformation block resets module, and whether the size for judging the corresponding transform block of chrominance block in the transform block after dividing surpasses The size of the maximum transform block of the corresponding chrominance block of described image block is crossed, is then by chrominance block pair in the transform block after the division The size of the transform block answered resets to the size of the maximum transform block of the corresponding chrominance block of described image block.
A kind of 26. transform blocks divide device, it is characterised in that including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Transform block division module, for transform block carry out non-square divide when, if the transform block is wide or high in institute The minimum value of the size of default transform block is stated, then stops dividing the transform block；
Wherein, the minimum value of the size of the default transform block is the minimum value wide or high of transform block, or sum wide and high Average value minimum value；It refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided
Also include：
Judge module is divided, for judging whether to carry out non-square division to the transform block；
The division judge module includes：
First acquisition unit, the size for obtaining the size of image block, the dividing mode of image block and default transform block Maximum, the size of described image block is the average value wide and high of image block, the maximum of the size of the default transform block Be the maximum wide and high of transform block, or the average value of sum wide and high minimum value；
First parameter acquiring unit, for the size according to image block, and the size of default transform block maximum, obtain the One parameter, first parameter is used to represent the size of the corresponding maximum transform block of described image block；
Second parameter acquiring unit, for the dividing mode according to image block, first parameter and the default transform block Size minimum value, obtain the second parameter, second parameter be used for represent the corresponding minimum transform block of described image block Whether size is more than the minimum value of the size of the default transform block；
Judging unit is divided, for according to second parameter, judging whether to carry out non-square division to the transform block.
27. transform blocks according to claim 26 divide device, it is characterised in that the first parameter acquiring unit, specific to use When the maximum of the size for being more than the default transform block when the size of described image block, first parameter is equal to described pre- If the maximum of the size of transform block；When the maximum of the size that the size of described image block is less than the default transform block, First parameter is equal to the size of described image block；And when the size of described image block is equal to the chi of the default transform block During very little maximum, first parameter is equal to the size of described image block.
28. transform blocks according to claim 26 or 27 divide device, it is characterised in that the size of described image block, pre- If the minimum value of the size of the maximum of the size of transform block and default transform block is adopted with 2 as the logarithm of radix is represented；
Second parameter acquiring unit, specifically for being to be divided using non-square when the dividing mode of described image block, and institute When stating the value of the first parameter more than the minimum value of the size for presetting transform block plus 1, the value of second parameter is set to 1, otherwise The value of second parameter is set to 0；
The division judging unit, specifically for when the value of second parameter is 1, judgement allows to use the transform block Non-square is divided, and otherwise, square division is used to the transform block；
Wherein, it refers to that image block or transform block are divided into wide and division methods not etc. high that the non-square is divided.
29. transform blocks according to claim 26 divide device, it is characterised in that also include：
Parameter acquisition module is divided, the default division number of plies divided for obtaining transform block to allow；
The transform block division module, specifically for carrying out non-square division to transform block, and in the division layer of the transform block Number divides the number of plies less than default, and during the minimum value of the size in default transform block wide or high of the transform block, it is right to stop The division of the transform block.
30. transform blocks according to claim 26 divide device, it is characterised in that also include：
Chromaticity transformation block size acquisition module, the size of the maximum transform block for obtaining the corresponding chrominance block of image block；
Chromaticity transformation block resets module, and whether the size for judging the corresponding transform block of chrominance block in the transform block after dividing surpasses The size of the maximum transform block of the corresponding chrominance block of described image block is crossed, is then by chrominance block pair in the transform block after the division The size of the transform block answered resets to the size of the maximum transform block of the corresponding chrominance block of described image block.
A kind of 31. transform blocks divide the code device of parameter, it is characterised in that including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Parameter coding module, for the pass between the minimum value of the size according to transform block and the size of the default transform block System, judges whether the division parameter of transcoding, coding transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided；
The parameter coding module, specifically for according to the second parameter, judging whether the division parameter of transcoding, coding transform block；
Also include：
Second acquisition unit, the size for obtaining the size of image block, the dividing mode of image block and default transform block Maximum；
First parameter acquiring unit, for the size according to image block, and the size of default transform block maximum, obtain the One parameter；
Second parameter acquiring unit, for the dividing mode according to image block, first parameter and the default transform block Size minimum value, obtain the second parameter.
32. transform blocks according to claim 31 divide the code device of parameter, it is characterised in that the parameter coding mould Block, specifically for when transform block using non-square divide, and the transform block the chi in the default transform block wide or high During very little minimum value, judge not needing the division parameter of transcoding, coding transform block, otherwise, the division parameter of transcoding, coding transform block.
33. transform blocks according to claim 31 divide the code device of parameter, it is characterised in that second parameter is used Whether it is more than the minimum value of the size of the default transform block in the size for representing the corresponding minimum transform block of image block；
First parameter is used to represent the size of the corresponding maximum transform block of described image block.
34. transform blocks according to claim 31 divide the code device of parameter, it is characterised in that the parameter coding mould Block, when adding second parameter specifically for being more than the minimum value of size of default transform block when the size of the transform block, compiles The division parameter of code conversion block, otherwise, it is not necessary to the division parameter of transcoding, coding transform block.
35. transform blocks according to claim 31 divide the code device of parameter, it is characterised in that also include：
Parameter acquisition module is divided, the default division number of plies divided for obtaining transform block to allow；
The parameter coding module, specifically for when the division number of plies of transform block is equal to the default division number of plies, judging not Need the division parameter of transcoding, coding transform block.
A kind of 36. transform blocks divide the decoding apparatus of parameter, it is characterised in that including：
Dimension acquisition module, the minimum value of the size for obtaining default transform block；
Parameter decoding module, for the pass between the minimum value of the size according to transform block and the size of the default transform block System, judges whether the division parameter of decoded transform block；
The division parameter of the transform block is used for expression and transform block is divided or is not divided；
The parameter decoding module, specifically for according to the second parameter, judging whether the division parameter of decoded transform block.
37. transform blocks according to claim 36 divide the decoding apparatus of parameter, it is characterised in that also include：
Second acquisition unit, the size for obtaining the size of image block, the dividing mode of image block and default transform block Maximum；
First parameter acquiring unit, for the size according to image block, and the size of default transform block maximum, obtain the One parameter；
Second parameter acquiring unit, for the dividing mode according to image block, first parameter and the default transform block Size minimum value, obtain the second parameter.
38. transform block according to claim 37 divides the decoding apparatus of parameter, it is characterised in that second parameter is used Whether it is more than the minimum value of the size of the default transform block in the size for representing the corresponding minimum transform block of image block；
First parameter is used to represent the size of the corresponding maximum transform block of described image block.
39. transform block according to claim 38 divides the decoding apparatus of parameter, it is characterised in that parameter decoding module, When minimum value specifically for being more than the size of default transform block when the size of the transform block adds second parameter, decoding becomes Change the division parameter of block.
40. transform blocks according to claim 36 divide the decoding apparatus of parameter, it is characterised in that also include：
Parameter acquisition module is divided, the default division number of plies divided for obtaining transform block to allow；
The parameter decoding module, specifically for when the division number of plies of transform block is equal to the default division number of plies, judging not Need the division parameter of decoded transform block.
CN201210227646.4A 2012-07-03 2012-07-03 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter Active CN102761742B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201210227646.4A CN102761742B (en) 2012-07-03 2012-07-03 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter
PCT/CN2013/078613 WO2014005500A1 (en) 2012-07-03 2013-07-01 Transform block division method and transform block division parameter encoding/ decoding method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210227646.4A CN102761742B (en) 2012-07-03 2012-07-03 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter

Publications (2)

Publication Number Publication Date
CN102761742A CN102761742A (en) 2012-10-31
CN102761742B true CN102761742B (en) 2017-06-06

Family

ID=47056046

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210227646.4A Active CN102761742B (en) 2012-07-03 2012-07-03 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter

Country Status (2)

Country Link
CN (1) CN102761742B (en)
WO (1) WO2014005500A1 (en)

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102761742B (en) * 2012-07-03 2017-06-06 华为技术有限公司 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter
CN102970532A (en) * 2012-11-13 2013-03-13 鸿富锦精密工业（深圳）有限公司 Image cutting system and method
CN102970535A (en) * 2012-11-13 2013-03-13 鸿富锦精密工业（深圳）有限公司 Image cutting system and method
CN102970534A (en) * 2012-11-13 2013-03-13 鸿富锦精密工业（深圳）有限公司 Image cutting system and method
CN102970533A (en) * 2012-11-13 2013-03-13 鸿富锦精密工业（深圳）有限公司 Image cutting system and method
WO2015015681A1 (en) * 2013-07-31 2015-02-05 パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ Image coding method, and image coding device
CN107566848B (en) 2016-06-30 2020-04-14 华为技术有限公司 Method and device for coding and decoding
KR20180110061A (en) * 2016-12-26 2018-10-08 닛본 덴끼 가부시끼가이샤 An image encoding method, an image decoding method, a video encoding apparatus, an image decoding apparatus,
US10536724B2 (en) * 2016-12-26 2020-01-14 Nec Corporation Video encoding method, video decoding method, video encoding device, video decoding device, and program
EP3383044A1 (en) * 2017-03-27 2018-10-03 Thomson Licensing Reducing redundancy between tree types
EP3383045A1 (en) 2017-03-27 2018-10-03 Thomson Licensing Multiple splits prioritizing for fast encoding
KR20200132979A (en) * 2018-04-01 2020-11-25 엘지전자 주식회사 Image processing method and apparatus therefor
WO2019234613A1 (en) 2018-06-05 2019-12-12 Beijing Bytedance Network Technology Co., Ltd. Partition tree with partition into 3 sub-blocks by horizontal and vertical splits
CN111050184A (en) * 2018-10-15 2020-04-21 华为技术有限公司 Image block dividing method and device
CN111327894B (en) * 2018-12-15 2022-05-17 华为技术有限公司 Block division method, video coding and decoding method and video coder and decoder
CN115714861A (en) * 2018-12-16 2023-02-24 华为技术有限公司 Video decoder and corresponding method
KR20230013160A (en) 2019-06-19 2023-01-26 엘지전자 주식회사 Image encoding/decoding method, apparatus and method for transmitting bitstream using adaptive size limitation of chroma block
CN112135147B (en) * 2019-06-24 2023-02-28 杭州海康威视数字技术股份有限公司 Encoding method, decoding method and device

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101389026A (en) * 2007-09-12 2009-03-18 索尼株式会社 Image coding apparatus and image coding method
CN101500161A (en) * 2008-01-31 2009-08-05 华为技术有限公司 Inter-frame prediction method and apparatus based on adaptive block transformation
CN101686395A (en) * 2005-07-21 2010-03-31 Lg电子株式会社 Method of encoding and decoding video signals

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101232625A (en) * 2008-02-26 2008-07-30 吉林大学 High efficient multidimensional video stream encoding and decoding method
EA029414B1 (en) * 2009-04-08 2018-03-30 Шарп Кабусики Кайся Video frame encoding apparatus and video frame decoding apparatus
KR101484280B1 (en) * 2009-12-08 2015-01-20 삼성전자주식회사 Method and apparatus for video encoding by motion prediction using arbitrary partition, and method and apparatus for video decoding by motion compensation using arbitrary partition
US20110176611A1 (en) * 2010-01-15 2011-07-21 Yu-Wen Huang Methods for decoder-side motion vector derivation
CN102761742B (en) * 2012-07-03 2017-06-06 华为技术有限公司 Transform block division methods, transform block divides coding method and the coding/decoding method of parameter

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101686395A (en) * 2005-07-21 2010-03-31 Lg电子株式会社 Method of encoding and decoding video signals
CN101389026A (en) * 2007-09-12 2009-03-18 索尼株式会社 Image coding apparatus and image coding method
CN101500161A (en) * 2008-01-31 2009-08-05 华为技术有限公司 Inter-frame prediction method and apparatus based on adaptive block transformation

Also Published As

Publication number Publication date
CN102761742A (en) 2012-10-31
WO2014005500A1 (en) 2014-01-09

Similar Documents

Publication Publication Date Title
CN102761742B (en) Transform block division methods, transform block divides coding method and the coding/decoding method of parameter
CN102970526B (en) A kind of method obtaining transform block size and module
CN102957907B (en) The acquisition methods of a kind of transform block positional information and module
TWI504209B (en) Method for decoding pictures
CN101389021B (en) Video encoding/decoding method and apparatus
CN105681794B (en) Moving image decoding apparatus and dynamic image decoding method
CN101068364B (en) Video encoder and graph processing unit
CN103918273B (en) It is determined that the method for the binary code word for conversion coefficient
CN104469378A (en) Coding position data for the last non-zero transform coefficient in a coefficient group
CN102685487B (en) Image coding and decoding methods, image coding and decoding equipment and network system
CN107836117A (en) Method for video and the conditional y-bend tree block segmenting structure of image coding and decoding
CN104469384A (en) Transposing a block of transform coefficients, based upon an intra-prediction mode
CN113784124B (en) Block matching encoding and decoding method for fine division using multi-shape sub-blocks
CN104853211A (en) Image compression method and apparatus employing various forms of reference pixel storage spaces
CN103067716A (en) Encoding and decoding methods and encoding and decoding device of range image
CN104704838B (en) two-dimensional block image coding
CN109246430B (en) Virtual reality 360-degree video fast intra prediction and CU partition advance decision
CN104581159A (en) Methods for determining block partition modes during video encoding and decoding and relevant devices
CN102065293B (en) Image compression method based on space domain predictive coding
CN110383837A (en) Method for video processing and equipment
CN110505488B (en) Image coding or decoding method for expanding prediction pixel array
CN106791859A (en) Method for video coding and video encoder
CN103299638B (en) Self-adaptive scanning device and scan method thereof
CN105812824B (en) A kind of video encoding method and device
CN107197273B (en) The coding method of image and device

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C41 Transfer of patent application or patent right or utility model
TA01 Transfer of patent application right

Effective date of registration: 20170220

Address after: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen

Applicant after: HUAWEI Technologies Co., Ltd.

Applicant after: Tsinghua University

Address before: 518129 Bantian HUAWEI headquarters office building, Longgang District, Guangdong, Shenzhen

Applicant before: HUAWEI Technologies Co., Ltd.

GR01 Patent grant
GR01 Patent grant