CN106375764A - Directional intra prediction and block copy prediction combined video intra coding method - Google Patents

Abstract

The invention discloses a directional intra prediction and block copy prediction combined video intra coding method. According to the method, a coding block is segmented into two sub-blocks in a flexible block segmentation mode, so prediction content in the same block can contain both local information and non-local information; and prediction of each block with least cost is decided through RDO (Rate Distortion Optimization). Moreover, rapid RDO and complete RDO are combined, and on the premise of not influencing the coding performance, the coding complexity is reduced. The general effect is that the complexity of a decoding side is basically not increased, and the higher compression efficiency is obtained.

Description

A kind of bonding position prediction and block replicate the video intra-frame encoding method predicted

Technical field

The present invention relates to technical field of video coding, more particularly, to a kind of bonding position prediction and block replicate the video of prediction Inner frame coding method.

Background technology

In recent years, with the fast development of the Internet, the application demand being directed to video in the Internet gets more and more, and regards The data volume of frequency is very big, wants video and can transmit in band-limited the Internet it may first have to the problem solving is exactly to regard Frequency compressed encoding problem.

The video encoding standard formulated broadly falls into mixed video coding framework.So-called hybrid video coding, typically by Following sections form: prediction (prediction), conversion (transform), quantization (quantization) and entropy code (entropy coding).Wherein prediction is generally divided into infra-frame prediction and inter prediction.Regarding of intra prediction mode can only be selected Frequency frame is referred to as i frame, both can be referred to as p frame or b frame using the frame of video that intra prediction mode can also use inter-frame forecast mode. Intra prediction mode, by the use of being in same frame with present encoding block but the pixel having rebuild out is as reference, removes space superfluous Remaining.Inter-frame forecast mode, then by the use of the pixel of other reconstructed frames as reference, removes time redundancy.General inter prediction Accuracy rate higher, but the first frame of video or the Stochastic accessing frame arranging for random access must be i frames, can only make Use infra-frame prediction.Lifting the accuracy rate of infra-frame prediction further, thus improving i frame compression efficiency, being very urgent in Video coding Demand.

At present, the newest standards of generic video coding are high efficiency video coding (hevc).hevc Infra-frame prediction have 35 kinds of patterns.Wherein pattern 0 is represented and is predicted using planar mode, and 1 expression is predicted using dc mode, and 2 to 34 expression use direction extrapolation patterns.The common feature of these patterns is the pixel of reconstruction based on present encoding block next-door neighbour Carry out extrapolation, hereafter will be referred to collectively as direction prediction (directional intra prediction, dip).

Additionally, there being a few thing to propose other intra-frame prediction methods, such as:

(y.ye and m.karczewicz, " improved is h.264intra for the extrapolation prediction of combination both direction coding based on bi-directional intra prediction,directional transform,and adaptive coefficient scanning,"image processing,2008.icip 2008.15th ieee international conference on,san diego,ca,2008,pp.2116-2119.).But, the side in hevc To the prediction of prediction or combination both direction weighting, only consider the local correlations in frame of video, and do not consider non local correlation Property.

Prediction (liu, d., sun, x., wu, f., ＆zhang, y.q. (2008) .edge- based on edge (edge) oriented uniform intra prediction.image processing,ieee transactions on,17 (10),1827-1836.).Prediction (t.k.tan, c.s.boon, and based on template matching (template matching) y.suzuki,“intra prediction bytemplate matching,”in image processing, 2006ieeeinternational conference on,oct 2006,pp.1693–1696.).But, pre- based on edge Survey or the prediction based on template matching, decoding end complexity is higher.

Replicate prediction (the siu-leong yu and christos of (intra block copy) based on intra block chrysafis,“intra-predictionusing intra-macroblock motion compensation,” oct.2006,us patent7,120,196.).It is current volume that intra block replicates Predicting Technique in the reconstruction regions of present frame Code block looks for a most like reference block as prediction, to the space motion between found most like reference block and current block Vector is encoded；It is highly suitable for the situation having repetition texture pattern to occur in a frame.This technology is follow-up in hevc Extended version: be used, because often going out in this kind of video in screen content coding (screen content coding, scc) The pattern of the regular shape of some repetitions existing, such as character etc..Block duplication Predicting Technique effect in scc is fine, but in nature Effect on driving birds is not good in video.This is primarily due to, and natural video frequency shooting angle is complicated, and noise is big, even if there is the texture repeating, It is mostly irregularly shaped, little and square coding unit alignment, therefore cannot search similar reference block.

Content of the invention

It is an object of the invention to provide a kind of bonding position prediction and block replicate the video intra-frame encoding method of prediction, can Consider the local correlations in frame of video, and non local dependency can be considered, also will ensure that decoding end complexity is relatively low simultaneously.

The purpose of the present invention is achieved through the following technical solutions:

A kind of bonding position prediction and block replicate the video intra-frame encoding method predicted, comprising:

Respectively present encoding block is divided using the predefined segmentation template comprising multiple classified types, each Plant classified types and all present encoding block is divided into two sub-blocks；

All sub-blocks are carried out with intra block and replicates ibc prediction: first determine the hunting zone of ibc prediction, recycle search model Each reference block in enclosing carries out ibc prediction to all sub-blocks, determines the optimal motion vector of each sub-block, and records every height The forecast error that block is obtained by optimal motion vector prediction；

Dip prediction is carried out to all sub-blocks, obtains the optimum prediction direction of each sub-block, and record each sub-block by optimum The forecast error that prediction direction prediction obtains；

After obtaining forecast error under ibc prediction with dip prediction for all sub-blocks, by quick rate-distortion optimization rdo Optimal dividing and corresponding prediction combination is determined in all classified types；

Compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, dip is carried out to present encoding block The Coding cost of prediction, and the Coding cost of optimal dividing and corresponding prediction combination, select the minimum prediction of Coding cost Pattern is as the final predictive mode of present encoding block.

Further, described multiple classified types include:

Cover horizontal direction, vertical direction, diagonally opposed from upper right to lower-left, and diagonal from left to bottom right Totally 28 kinds of classified types in direction, wherein, all there is the division class of 7 kinds of classified types, horizontal direction and vertical direction in each direction Type is referred to as rectangular partition, and diagonally opposed from upper right to lower-left is referred to as with from diagonally opposed classified types left to bottom right For triangle division；

Or, including 7 kinds of classified types that upper left is l type and upper right is 7 kinds of classified types of l type, and both divides Mode is all that present encoding block is divided into two sub-blocks: a sub-block is square block, and another sub-block is two pieces of adjacent squares The l type block of shape block composition.

Further, if described multiple classified types are to cover horizontal direction, vertical direction, right from upper right to lower-left Angular direction, and from diagonally opposed totally 28 kinds of classified types left to bottom right, then intra block duplication is carried out to all sub-blocks During ibc prediction, determine the optimal motion vector of each sub-block, and record each sub-block by optimal motion vector prediction obtain pre- Survey error to include:

7 kinds of divisions of horizontal direction are designated as h1～h7,7 kinds of divisions of vertical direction are designated as v1～v7；From upper right to a left side Under diagonally opposed 7 kind division be designated as t1～t7, the diagonally opposed 7 kind division from left to bottom right is designated as d1～d7；

After each in hunting zone is completed, obtain the minimum of each sub-block in every kind of division with reference to block search Forecast error and corresponding optimal motion vector, carry out following labelling:

In horizontal direction, the optimal motion vector mv of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mv_ibc_ respectively H1_1 and mv_ibc_h1_2, corresponding forecast error is designated as sad_ibc_h1_1 and sad_ibc_h1_2；The like, the 7th kind The optimal motion vector mv of upper and lower two sub-blocks of classified types h7 is designated as mv_ibc_h7_1 and mv_ibc_h7_2, accordingly respectively Forecast error is designated as sad_ibc_h7_1 and sad_ibc_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_ respectively V1_1 and mv_ibc_v1_2, corresponding forecast error is designated as sad_ibc_v1_1 and sad_ibc_v1_2；The like, the 7th kind Classified types v7 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_v7_1 and mv_ibc_v7_2 respectively, accordingly Forecast error is designated as sad_ibc_v7_1 and sad_ibc_v7_2；

The optimal motion arrow of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right in diagonally opposed from upper right to lower-left Amount mv is designated as mv_ibc_t1_1 and mv_ibc_t1_2 respectively, and corresponding forecast error is designated as sad_ibc_t1_1 and sad_ibc_ t1_2；The like, the optimal motion vector mv of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as mv_ respectively Ibc_t7_1 and mv_ibc_t7_2, corresponding forecast error is designated as sad_ibc_t7_1 and sad_ibc_t7_2；

The optimal motion arrow of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left from left to bottom right diagonally opposed Amount mv is designated as mv_ibc_d1_1 and mv_ibc_d1_2 respectively, and corresponding forecast error is designated as sad_ibc_d1_1 and sad_ibc_ d1_2；The like, the optimal motion vector mv of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as mv_ respectively Ibc_d7_1 and mv_ibc_d7_2, corresponding forecast error is designated as sad_ibc_d7_1 and sad_ibc_d7_2.

Further, described optimum prediction direction dip prediction being carried out to all sub-blocks, obtaining each sub-block, and record Each sub-block is included by the forecast error that optimum prediction direction prediction obtains:

Dip prediction is and determines optimum prediction direction in 35 kinds of infra-frame predictions of hevc, in the predictive value in every kind of direction After generation, calculate the forecast error of each sub-block between prediction block and present encoding block；When 35 prediction direction have been traveled through After one-tenth, obtain the forecast error of minimum of each sub-block and corresponding intra prediction direction in every kind of division, carried out as follows Labelling:

In horizontal direction, the optimum prediction direction of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mode_dip_ respectively H1_1 and mode_dip_h1_2, corresponding forecast error is designated as sad_dip_h1_1 and sad_dip_h1_2；The like, the 7th The optimum prediction direction planting upper and lower two sub-blocks of classified types h7 is designated as mode_dip_h7_1 and mode_dip_h7_2, phase respectively The forecast error answered is designated as sad_dip_h7_1 and sad_dip_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimum prediction direction of two sub-blocks be designated as mode_dip_ respectively V1_1 and mode_dip_v1_2, corresponding forecast error is designated as sad_dip_v1_1 and sad_dip_v1_2；The like, the 7th Kind of classified types h7 about the optimum prediction direction of two sub-blocks be designated as mode_dip_v7_1 and mode_dip_v7_2, phase respectively The forecast error answered is designated as sad_dip_v7_1 and sad_dip_v7_2；

The optimum prediction side of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right in diagonally opposed from upper right to lower-left To being designated as mode_dip_t1_1 and mode_dip_t1_2 respectively, corresponding forecast error is designated as sad_dip_t1_1 and sad_ dip_t1_2；The like, the optimum prediction direction of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as respectively Mode_dip_t7_1 and mode_dip_t7_2, corresponding forecast error is designated as sad_dip_t7_1 and sad_dip_t7_2；

The optimum prediction side of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left from left to bottom right diagonally opposed To being designated as mode_dip_d1_1 and mode_dip_d1_2 respectively, corresponding forecast error is designated as sad_dip_d1_1 and sad_ dip_d1_2；The like, the optimum prediction direction of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as respectively Mode_dip_d7_1 and mode_dip_d7_2, corresponding forecast error is designated as sad_dip_d7_1 and sad_dip_d7_2.

Further, described determine optimal dividing and corresponding in all classified types by quick rate-distortion optimization rdo Prediction combination include:

In the horizontal direction, vertical direction, diagonally opposed from upper right to lower-left and diagonally opposed each from left to bottom right From 7 kinds of classified types in determine the minimum division of overall cost and prediction combination, its process is as follows:

In the horizontal direction in 7 kinds of classified types of h1～h7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_h\_i\_1=sad\_ibc\_hi\_1+sad\_dip\_hi\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_hi\_1,mode\_dip\_hi\_2)\\ c\_h\_i\_2=sad\_ibc\_hi\_2+sad\_dip\_hi\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_hi\_2,mode\_dip\_hi\_1)\end{array};$

In above formula, i=1,2 ..., 7；λ is Lagrange multiplier, to be determined by the parameter that encoder is specified, bits Function is the estimation that mv and dip direction is carried out expend needed for entropy code with bit number；Find out c_h_i_1, in c_h_i_2 Little combination, is designated as c_h_k1_p1, and corresponding predictive mode is designated as (k1, p1), and k1 is a number in 1 to 7, indicates division Type is hk1；P1 is a number in 1 to 2, indicates prediction compound mode, and 1 refers to that left-hand component carries out ibc prediction, the right portion Divide and carry out dip prediction, otherwise 2；

In 7 kinds of classified types of vertical direction v1～v7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_v\_i\_1=sad\_ibc\_vi\_1+sad\_dip\_vi\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_vi\_1,mode\_dip\_vi\_2)\\ c\_v\_i\_2=sad\_ibc\_vi\_2+sad\_dip\_vi\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_vi\_2,mode\_dip\_vi\_1)\end{array};$

Find out c_v_i_1 from above formula, in c_v_i_2, minimum combination, is designated as c_v_k2_p2, corresponding predictive mode It is designated as (k2, p2), k2 is a number in 1 to 7, indicating classified types is vk2；P2 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper part carries out ibc prediction, and lower part carries out dip prediction, otherwise 2；

In 7 kinds of divisions of the diagonally opposed t1～t7 of upper right to lower-left, have 14 kinds of combinations, its corresponding cost is respectively For:

$\begin{array}{c}c\_t\_i\_1=sad\_ibc\_ti\_1+sad\_dip\_ti\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ti\_1,mode\_dip\_ti\_2)\\ c\_t\_i\_2=sad\_ibc\_ti\_2+sad\_dip\_ti\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ti\_2,mode\_dip\_ti\_1)\end{array};$

Find out c_t_i_1 from above formula, in c_t_i_2, minimum combination, is designated as c_t_k3_p3, corresponding predictive mode It is designated as (k3, p3), k3 is a number in 1 to 7, indicating classified types is tk3；P3 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper left carries out ibc prediction, and lower right-most portion carries out dip prediction, otherwise 2；

In 7 kinds of divisions of diagonally opposed d1～d7 left to bottom right, have 14 kinds of combinations, its corresponding cost is respectively For:

$\begin{array}{c}c\_d\_i\_1=sad\_ibc\_di\_1+sad\_dip\_di\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_di\_1,mode\_dip\_di\_2)\\ c\_d\_i\_2=sad\_ibc\_di\_2+sad\_dip\_di\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_di\_2,mode\_dip\_di\_1)\end{array}$

Find out c_d_i_1 from above formula, in c_d_i_2, minimum combination, is designated as c_d_k4_p4, corresponding predictive mode It is designated as (k4, p4), k4 is a number in 1 to 7, indicating classified types is dk4；P4 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper right portion carries out ibc prediction, and bottom left section carries out dip prediction, otherwise 2.

Further, described compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, to current Encoding block carries out the Coding cost of dip prediction, and the Coding cost of optimal dividing and corresponding prediction combination, selects coding generation The minimum predictive mode of valency includes as the final predictive mode of present encoding block:

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For horizontal direction, by predictive mode (k1, p1), pre- according to corresponding motion vector mv_ibc_hk1_p1, frame in Survey directionGenerate the prediction block of current block, the volume calculating through complete rdo with classified types hk1 Code cost is designated as c_h；Wherein,

For vertical direction, by predictive mode (k2, p2), pre- according to corresponding motion vector mv_ibc_vk2_p2, frame in Survey directionGenerate the prediction block of current block, the volume calculating through complete rdo with classified types vk2 Code cost is designated as c_v；Wherein,

For diagonally opposed from upper right to lower-left, by predictive mode (k3, p3), according to corresponding motion vector mv_ Ibc_tk3_p3, intra prediction directionGenerate the prediction block of current block with classified types tk3, pass through The Coding cost that complete rdo calculates is designated as c_t；Wherein,

For diagonally opposed from left to bottom right, by predictive mode (k4, p4), according to corresponding motion vector mv_ Ibc_dk4_p4, intra prediction directionGenerate the prediction block of current block with classified types dk4, pass through The Coding cost that complete rdo calculates is designated as c_d；

Compare the size of c_dip, c_ibc, c_h, c_v, c_t, c_d, corresponding for minimum code cost predictive mode is made For the predictive mode that present encoding block is final.

Further, if described multiple classified types include, upper left is 7 kinds of classified types of l type and upper right is 7 kinds of l type Classified types, then, when all sub-blocks are carried out with intra block duplication ibc prediction, determine the optimal motion vector of each sub-block, and remember Record each sub-block to be included by the forecast error that optimal motion vector prediction obtains:

Upper left is designated as l1～l7 for 7 kinds of classified types of l type, upper right is that 7 kinds of classified types of l type are designated as r1～r7；

After each in hunting zone is completed, obtain the minimum of each sub-block in every kind of division with reference to block search Forecast error and corresponding optimal motion vector, carry out following labelling:

Upper left is that the optimal motion vector mv of two sub-blocks in upper left and bottom right of the 1st kind of classified types l1 in l type remembers respectively For mv_ibc_l1_1, mv_ibc_l1_2, corresponding forecast error is designated as sad_ibc_l1_1 and sad_ibc_l1_2；Class successively Push away, the optimal motion vector mv of two sub-blocks in the 7th kind of classified types l7 upper left and bottom right is designated as mv_ibc_l7_1 and mv_ respectively Ibc_l7_2, corresponding forecast error is designated as sad_ibc_l7_1 and sad_ibc_l7_2；

Upper right is that the optimal motion vector mv of two sub-blocks of upper right and lower-left of the 1st kind of classified types r1 in l type remembers respectively For mv_ibc_r1_1, mv_ibc_r1_2, corresponding forecast error is designated as sad_ibc_r1_1 and sad_ibc_r1_2；Class successively Push away, the optimal motion vector mv of two sub-blocks of the 7th kind of classified types r7 upper right and lower-left is designated as mv_ibc_r7_1 and mv_ respectively Ibc_r7_2, corresponding forecast error is designated as sad_ibc_r7_1 and sad_ibc_r7_2.

Further, described optimum prediction direction dip prediction being carried out to all sub-blocks, obtaining each sub-block, and record Each sub-block is included by the forecast error that optimum prediction direction prediction obtains:

Dip prediction is and determines optimum prediction direction in 35 kinds of infra-frame predictions of hevc, in the predictive value in every kind of direction After generation, calculate the forecast error of each sub-block between prediction block and present encoding block；When 35 prediction direction have been traveled through After one-tenth, obtain the forecast error of minimum of each sub-block and corresponding intra prediction direction in every kind of division, carried out as follows Labelling:

Upper left be l type in the 1st kind of classified types l1 two sub-blocks in upper left and bottom right optimum prediction direction remember respectively For mode_dip_l1_1 and mode_dip_l1_2, corresponding forecast error is designated as sad_dip_l1_1 and sad_dip_l1_2； The like, the optimum prediction direction of two sub-blocks in the 7th kind of classified types l7 upper left and bottom right is designated as mode_dip_l7_1 respectively With mode_dip_l7_2, corresponding forecast error is designated as sad_dip_l7_1 and sad_dip_l7_2；

Upper right be l type in the 1st kind of classified types r1 two sub-blocks of upper right and lower-left optimum prediction direction remember respectively For mode_dip_r1_1 and mode_dip_r1_2, corresponding forecast error is designated as sad_dip_r1_1 and sad_dip_r1_2； The like, the optimum prediction direction of two sub-blocks of the 7th kind of classified types r7 upper right and lower-left is designated as mode_dip_r7_1 respectively With mode_dip_r7_2, corresponding forecast error is designated as sad_dip_r7_1 and sad_dip_r7_2.

Further, described determine optimal dividing and corresponding in all classified types by quick rate-distortion optimization rdo Prediction combination include:

It is in the respective 7 kinds of classified types of l type, to determine the minimum division of overall cost and pre- in upper left for l type and upper right Survey combination, its process is as follows:

In 7 kinds of classified types for l type l1～l7 for the upper left, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_l\_i\_1=sad\_ibc\_li\_1+sad\_dip\_li\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_li\_1,mode\_dip\_li\_2)\\ c\_l\_i\_2=sad\_ibc\_li\_2+sad\_dip\_li\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_li\_2,mode\_dip\_li\_1)\end{array};$

In above formula, i=1,2 ..., 7；Bits function is that mv and dip direction is carried out expend needed for entropy code with bit number One estimation；Find out c_l_i_1, in c_l_i_2, minimum combination, is designated as c_l_k1_p1, corresponding predictive mode be designated as (k1, P1), k1 is a number in 1 to 7, and indicating classified types is lk1；P1 is a number in 1 to 2, indicates prediction compound mode, 1 refers to that the l type part of upper left carries out ibc prediction, and the square portion of bottom right carries out dip prediction, otherwise 2；

In 7 kinds of classified types for l type r1～r7 for the upper right, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_r\_i\_1=sad\_ibc\_ri\_1+sad\_dip\_ri\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ri\_1,mode\_dip\_ri\_2)\\ c\_r\_i\_2=sad\_ibc\_ri\_2+sad\_dip\_ri\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ri\_2,mode\_dip\_ri\_1)\end{array};$

Find out c_r_i_1, in c_r_i_2, minimum combination, is designated as c_r_k2_p2, corresponding predictive mode be designated as (k2, P2), k2 is a number in 1 to 7, and indicating division position is rk2；P2 is a number in 1 to 2, indicates prediction compound mode, 1 refers to that the l type part of upper right carries out ibc prediction, and the square portion of lower-left carries out dip prediction, otherwise 2.

Further, described compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, to current Encoding block carries out the Coding cost of dip prediction, and the Coding cost of optimal dividing and corresponding prediction combination, selects coding generation The minimum predictive mode of valency includes as the final predictive mode of present encoding block:

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For upper left for l type classified types, by predictive mode (k1, p1), according to corresponding motion vector mv_ibc_ Lk1_p1, intra prediction directionGenerate the prediction block of current block with classified types lk1, through complete The Coding cost that rdo calculates is designated as c_l；Wherein,

For upper right for l type classified types, by predictive mode (k2, p2), according to corresponding motion vector mv_ibc_ Rk2_p2, intra prediction directionGenerate the prediction block of current block with classified types rk2, through complete The Coding cost that rdo calculates is designated as c_r；Wherein,

Compare the size of c_dip, c_ibc, c_l, c_r, corresponding for minimum code cost predictive mode is compiled as current The final predictive mode of code block.

As seen from the above technical solution provided by the invention, by flexible block partitioning scheme, encoding block is divided into two So that same piece of predictive content can comprise local message and non-local information simultaneously, then rdo determines often individual sub-block The prediction of individual block Least-cost.Devise quick rdo and the combination of complete rdo simultaneously, on the premise of not affecting coding efficiency, Reduce encoder complexity.General effect is substantially not increase decoding end complexity, and obtains higher compression efficiency.

Brief description

In order to be illustrated more clearly that the technical scheme of the embodiment of the present invention, below will be to required use in embodiment description Accompanying drawing be briefly described it should be apparent that, drawings in the following description are only some embodiments of the present invention, for this For the those of ordinary skill in field, on the premise of not paying creative work, other can also be obtained according to these accompanying drawings Accompanying drawing.

Fig. 1 is the video intra-frame encoding method that a kind of bonding position prediction provided in an embodiment of the present invention and block replicate prediction Flow chart；

Fig. 2 is v1 in vertical direction provided in an embodiment of the present invention, the division schematic diagram of v2 ..., v7；

Fig. 3 is h1 in horizontal direction provided in an embodiment of the present invention, the division schematic diagram of h2 ..., h7；

Fig. 4 is diagonally opposed middle t1, the t2 from upper right to lower-left provided in an embodiment of the present invention ..., the division of t7 is illustrated Figure；

Fig. 5 is middle d1, d2 from left to bottom right diagonally opposed provided in an embodiment of the present invention ..., the division of d7 is shown It is intended to；

Fig. 6 is the schematic diagram of ibc prediction provided in an embodiment of the present invention；

Fig. 7 is l1, l2 in the classified types that upper left provided in an embodiment of the present invention is l type ..., the division schematic diagram of l7；

Fig. 8 is r1, r2 in the classified types that upper right provided in an embodiment of the present invention is l type ..., the division schematic diagram of r7；

Fig. 9 is that the coding efficiency after the scheme according to embodiment one provided in an embodiment of the present invention is tested is illustrated Figure.

Specific embodiment

With reference to the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Ground description is it is clear that described embodiment is only a part of embodiment of the present invention, rather than whole embodiments.Based on this Inventive embodiment, the every other enforcement that those of ordinary skill in the art are obtained under the premise of not making creative work Example, broadly falls into protection scope of the present invention.

Fig. 1 is the video intra-frame encoding method that a kind of bonding position prediction provided in an embodiment of the present invention and block replicate prediction Flow chart, as shown in figure 1, it mainly comprises the steps:

1) respectively present encoding block is divided using the predefined segmentation template comprising multiple classified types, often Present encoding block is all divided into two sub-blocks by a kind of classified types.

2) all sub-blocks are carried out with intra block duplication (ibc) prediction: first determine the hunting zone of ibc prediction, recycling is searched Each reference block in the range of rope carries out ibc prediction to all sub-blocks, determines the optimal motion vector of each sub-block, and records every The forecast error that individual sub-block is obtained by optimal motion vector prediction.

3) all sub-blocks are carried out with dip prediction (frame in direction prediction), obtain the optimum prediction direction of each sub-block, and remember Record the forecast error that each sub-block is obtained by optimum prediction direction prediction.

4) after obtaining forecast error under ibc prediction with dip prediction for all sub-blocks, by quick rate-distortion optimization (rdo) optimal dividing and corresponding prediction combination are determined in all classified types.

5) compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, present encoding block is carried out The Coding cost that dip predicts, and the Coding cost of optimal dividing and corresponding prediction combination, select the pre- of Coding cost minimum Survey pattern is as the final predictive mode of present encoding block.

Above-mentioned steps 2) with step 3) do not differentiate between successively execution, additionally, in execution step 2) with step 3) when also can be simultaneously Whole encoding block is carried out with corresponding ibc prediction and dip prediction.

In the embodiment of the present invention, because the quantity of classified types is related to the species of prediction, that is, classified types are more flexible, in advance The species surveyed is more, but encoder complexity is also higher simultaneously, so considering coding efficiency and complexity in the present embodiment, Propose the dividing mode of two kinds of forms: the first is to cover horizontal direction, vertical direction, the diagonal side from upper right to lower-left To, and from diagonally opposed totally 28 kinds of classified types left to bottom right, wherein, all there are 7 kinds of classified types, water in each direction Square be referred to as rectangular partition to the classified types with vertical direction, from upper right to lower-left diagonally opposed with from left to bottom right Diagonally opposed classified types be referred to as triangle division；Second is to include 7 kinds of classified types that upper left is l type and upper right For 7 kinds of classified types of l type, both dividing mode are all that present encoding block is divided into two sub-blocks: a sub-block is just Square block, another sub-block is the l type block of two pieces of adjacent rectangle block compositions.

In order to make it easy to understand, the above-mentioned step of execution to be respectively directed to above two dividing mode with reference to two embodiments Rapid 1)～5) process elaborate.

Embodiment one

A kind of bonding position prediction that the present embodiment provides and block replicate the step of the video intra-frame encoding method of prediction such as Under:

The first step, determine classified types.

In the present embodiment, described multiple classified types are to cover horizontal direction, vertical direction, right from upper right to lower-left Angular direction, and from diagonally opposed totally 28 kinds of classified types left to bottom right.

7 kinds of divisions of horizontal direction are designated as h1～h7,7 kinds of divisions of vertical direction are designated as v1～v7；From upper right to a left side Under diagonally opposed 7 kind division be designated as t1～t7, the diagonally opposed 7 kind division from left to bottom right is designated as d1～d7.

As shown in Fig. 2 being v1, the division schematic diagram of v2 ..., v7.Represent present encoding block in one of v1 square, Square is divided into two rectangles in left and right, and two rectangular elevation are the same, and the width of left side rectangle accounts for whole square width 1/8；The width of the right rectangle accounts for the 7/8 of entirely square width；In v2, a square is divided into two rectangles in left and right, Two rectangular elevation are the same, and the width of left side rectangle accounts for the 2/8 of entirely square width；The width of the right rectangle accounts for whole pros The 6/8 of shape width；In the same manner, in v7, the width of left side rectangle accounts for the 7/8 of entirely square width；The width of the right rectangle accounts for whole The 1/8 of individual square width.H1, h2 ..., h7 and v1, v2 ..., v7 correspond to, be the former be in level, the latter is vertical Direction.H1, h2 ..., h7 divides as shown in Figure 3.

The diagonally opposed division from upper right to lower-left in triangle division is described below again, t1, t2, t3 and t4 divide one Individual square is divided into the remaining part of isosceles right triangle and the lower right corner in the upper left corner；T5, t6 and t7 divide pros Shape is divided into the remaining part of isosceles right angle trigonometry type and the upper left corner in the lower right corner.As shown in figure 4, during t1 divides, isosceles right angle three Angular bottom side length is the 1/4 of the square length of side；During t2 divides, the bottom side length of isosceles right triangle is the 2/ of the square length of side 4；During t4 divides in the same manner, the bottom side length of isosceles right triangle is the 4/4 of the square length of side, that is, be equal to the square length of side.t5、 T6, t7 divide similar with the division of t1 to t4.D1, d2 ..., the division of d7 and t1, t2 ..., t7 divides with regard to vertical direction pair Claim, as shown in Figure 5.

Second step, according to the segmentation template of above-mentioned classified types, present encoding block is divided after, need to all sons Block carries out intra block duplication (ibc) prediction.

It needs to be determined that hunting zone before search, hunting zone can be less than or equal to the reconstruction portion in present frame Point.If hunting zone set bigger, it is higher to search plain complexity, is possible to find more preferable predictive value to current block simultaneously. Hunting zone begins to after determining search for,

The used full search of the present embodiment, that is, for each reference block in hunting zone, calculate its with every The forecast error of one sub-block, for example (sad, absolute error and).Because needing below by rdo (rate distortion Optimization, rate-distortion optimization) determine optimal prediction modes, so calculate here reference block and present encoding block it Between, sad of different shapes.

Fig. 6 represents the schematic diagram of a certain step during ibc, and (i.e. present encoding block is located the rectangular area shown in dotted line Rectangular area) it is still uncoded part, its periphery is the part of reconstruction in present frame, and present encoding block is in search model Search in enclosing reaches to one of block, and arrow represents that (block vector, block vector represent reference block and current to bv The coordinate difference of block, similar to motion vector when inter frame motion estimation, i.e. motion vector, represents bv with mv below, Being write as mv is a kind of customary expression).From above, present encoding block has 28 kinds of dividing mode, every kind of dividing mode Present encoding block is divided into two sub-blocks, this two sub-blocks, sad be required for calculate, so needing altogether to calculate 56 divided block Sad add whole encoding block sad, that is, 57 kinds.Because two sub-block sad sums that the every kind of division in 28 kinds obtains are equal to The sad of whole block, so actual 29 sad of calculation that amount to (are left 28 sad and can deduct corresponding fritter according to the sad of bulk Sad obtain).

After each in hunting zone is completed, obtain the minimum of each sub-block in every kind of division with reference to block search Forecast error and corresponding optimal motion vector, carry out following labelling:

In horizontal direction, the optimal motion vector mv of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mv_ibc_ respectively H1_1 and mv_ibc_h1_2, corresponding forecast error is designated as sad_ibc_h1_1 and sad_ibc_h1_2；The like, the 7th kind The optimal motion vector mv of upper and lower two sub-blocks of classified types h7 is designated as mv_ibc_h7_1 and mv_ibc_h7_2, accordingly respectively Forecast error is designated as sad_ibc_h7_1 and sad_ibc_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_ respectively V1_1 and mv_ibc_v1_2, corresponding forecast error is designated as sad_ibc_v1_1 and sad_ibc_v1_2；The like, the 7th kind Classified types v7 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_v7_1 and mv_ibc_v7_2 respectively, accordingly Forecast error is designated as sad_ibc_v7_1 and sad_ibc_v7_2；

The optimal motion arrow of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right in diagonally opposed from upper right to lower-left Amount mv is designated as mv_ibc_t1_1 and mv_ibc_t1_2 respectively, and corresponding forecast error is designated as sad_ibc_t1_1 and sad_ibc_ t1_2；The like, the optimal motion vector mv of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as mv_ respectively Ibc_t7_1 and mv_ibc_t7_2, corresponding forecast error is designated as sad_ibc_t7_1 and sad_ibc_t7_2；

The optimal motion arrow of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left from left to bottom right diagonally opposed Amount mv is designated as mv_ibc_d1_1 and mv_ibc_d1_2 respectively, and corresponding forecast error is designated as sad_ibc_d1_1 and sad_ibc_ d1_2；The like, the optimal motion vector mv of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as mv_ respectively Ibc_d7_1 and mv_ibc_d7_2, corresponding forecast error is designated as sad_ibc_d7_1 and sad_ibc_d7_2.

3rd step, dip prediction is carried out to all sub-blocks, obtain the optimum prediction direction of each sub-block, and record each son The forecast error that block is obtained by optimum prediction direction prediction.

Dip prediction is and determines optimum prediction direction in 35 kinds of infra-frame predictions of hevc, because before not carrying out rdo Not can determine that the optimum dividing mode of current coded unit, so we need to find for each part in every kind of division Optimum prediction direction, records the sad under each part optimum prediction direction simultaneously.After predictive value in every kind of direction generates, Calculate the forecast error (i.e. sad) of each sub-block between prediction block and present encoding block；From above, need altogether to calculate 56 Divide sub-block sad add whole encoding block sad, that is, 57 kinds.

After 35 prediction direction traversals are completed, obtain the forecast error of the minimum of each sub-block in every kind of division With corresponding intra prediction direction, carry out following labelling:

In horizontal direction, the optimum prediction direction of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mode_dip_ respectively H1_1 and mode_dip_h1_2, corresponding forecast error is designated as sad_dip_h1_1 and sad_dip_h1_2；The like, the 7th The optimum prediction direction planting upper and lower two sub-blocks of classified types h7 is designated as mode_dip_h7_1 and mode_dip_h7_2, phase respectively The forecast error answered is designated as sad_dip_h7_1 and sad_dip_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimum prediction direction of two sub-blocks be designated as mode_dip_ respectively V1_1 and mode_dip_v1_2, corresponding forecast error is designated as sad_dip_v1_1 and sad_dip_v1_2；The like, the 7th Kind of classified types h7 about the optimum prediction direction of two sub-blocks be designated as mode_dip_v7_1 and mode_dip_v7_2, phase respectively The forecast error answered is designated as sad_dip_v7_1 and sad_dip_v7_2；

The optimum prediction side of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right in diagonally opposed from upper right to lower-left To being designated as mode_dip_t1_1 and mode_dip_t1_2 respectively, corresponding forecast error is designated as sad_dip_t1_1 and sad_ dip_t1_2；The like, the optimum prediction direction of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as respectively Mode_dip_t7_1 and mode_dip_t7_2, corresponding forecast error is designated as sad_dip_t7_1 and sad_dip_t7_2；

The optimum prediction side of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left from left to bottom right diagonally opposed To being designated as mode_dip_d1_1 and mode_dip_d1_2 respectively, corresponding forecast error is designated as sad_dip_d1_1 and sad_ dip_d1_2；The like, the optimum prediction direction of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as respectively Mode_dip_d7_1 and mode_dip_d7_2, corresponding forecast error is designated as sad_dip_d7_1 and sad_dip_d7_2.

4th step, determine optimal dividing by quick rate-distortion optimization rdo in all classified types and corresponding predict Combination.

In the horizontal direction, vertical direction, diagonally opposed from upper right to lower-left and diagonally opposed each from left to bottom right From 7 kinds of classified types in determine the minimum division of overall cost and prediction combination, its process is as follows:

In the horizontal direction in 7 kinds of classified types of h1～h7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_h\_i\_1=sad\_ibc\_hi\_1+sad\_dip\_hi\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_hi\_1,mode\_dip\_hi\_2)\\ c\_h\_i\_2=sad\_ibc\_hi\_2+sad\_dip\_hi\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_hi\_2,mode\_dip\_hi\_1)\end{array};$

In above formula, i=1,2 ..., 7；λ is Lagrange multiplier, to be determined by the parameter that encoder is specified, bits Function is the estimation that mv and dip direction is carried out expend needed for entropy code with bit number；Find out c_h_i_1, in c_h_i_2 Little combination, is designated as c_h_k1_p1, and corresponding predictive mode is designated as (k1, p1), and k1 is a number in 1 to 7, indicates division Type is hk1；P1 is a number in 1 to 2, indicates prediction compound mode, and 1 refers to that left-hand component carries out ibc prediction, the right portion Divide and carry out dip prediction, otherwise 2；

In 7 kinds of classified types of vertical direction v1～v7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_v\_i\_1=sad\_ibc\_vi\_1+sad\_dip\_vi\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_vi\_1,mode\_dip\_vi\_2)\\ c\_v\_i\_2=sad\_ibc\_vi\_2+sad\_dip\_vi\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_vi\_2,mode\_dip\_vi\_1)\end{array};$

Find out c_v_i_1 from above formula, in c_v_i_2, minimum combination, is designated as c_v_k2_p2, corresponding predictive mode It is designated as (k2, p2), k2 is a number in 1 to 7, indicating classified types is vk2；P2 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper part carries out ibc prediction, and lower part carries out dip prediction, otherwise 2；

In 7 kinds of divisions of the diagonally opposed t1～t7 of upper right to lower-left, have 14 kinds of combinations, its corresponding cost is respectively For:

$\begin{array}{c}c\_t\_i\_1=sad\_ibc\_ti\_1+sad\_dip\_ti\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ti\_1,mode\_dip\_ti\_2)\\ c\_t\_i\_2=sad\_ibc\_ti\_2+sad\_dip\_ti\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ti\_2,mode\_dip\_ti\_1)\end{array};$

Find out c_t_i_1 from above formula, in c_t_i_2, minimum combination, is designated as c_t_k3_p3, corresponding predictive mode It is designated as (k3, p3), k3 is a number in 1 to 7, indicating classified types is tk3；P3 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper left carries out ibc prediction, and lower right-most portion carries out dip prediction, otherwise 2；

In 7 kinds of divisions of diagonally opposed d1～d7 left to bottom right, have 14 kinds of combinations, its corresponding cost is respectively For:

$\begin{array}{c}c\_d\_i\_1=sad\_ibc\_di\_1+sad\_dip\_di\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_di\_1,mode\_dip\_di\_2)\\ c\_d\_i\_2=sad\_ibc\_di\_2+sad\_dip\_di\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_di\_2,mode\_dip\_di\_1)\end{array}$

Find out c_d_i_1 from above formula, in c_d_i_2, minimum combination, is designated as c_d_k4_p4, corresponding predictive mode It is designated as (k4, p4), k4 is a number in 1 to 7, indicating classified types is dk4；P4 is a number in 1 to 2, indicates prediction Compound mode, 1 refers to that upper right portion carries out ibc prediction, and bottom left section carries out dip prediction, otherwise 2.

5th step, compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, to present encoding block Carry out the Coding cost of dip prediction, and the Coding cost of optimal dividing and corresponding prediction combination, select Coding cost minimum Predictive mode as the final predictive mode of present encoding block.

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For horizontal direction, by predictive mode (k1, p1), pre- according to corresponding motion vector mv_ibc_hk1_p1, frame in Survey directionGenerate the prediction block of current block, the volume calculating through complete rdo with classified types hk1 Code cost is designated as c_h；Wherein,

For vertical direction, by predictive mode (k2, p2), pre- according to corresponding motion vector mv_ibc_vk2_p2, frame in Survey directionGenerate the prediction block of current block, the volume calculating through complete rdo with classified types vk2 Code cost is designated as c_v；Wherein,

For diagonally opposed from upper right to lower-left, by predictive mode (k3, p3), according to corresponding motion vector mv_ Ibc_tk3_p3, intra prediction directionGenerate the prediction block of current block with classified types tk3, pass through The Coding cost that complete rdo calculates is designated as c_t；Wherein,

For diagonally opposed from left to bottom right, by predictive mode (k4, p4), according to corresponding motion vector mv_ Ibc_dk4_p4, intra prediction directionGenerate the prediction block of current block with classified types dk4, pass through The Coding cost that complete rdo calculates is designated as c_d；

Compare the size of c_dip, c_ibc, c_h, c_v, c_t, c_d, corresponding for minimum code cost predictive mode is made For the predictive mode that present encoding block is final.

Embodiment two

A kind of bonding position prediction that the present embodiment provides and block replicate the step of the video intra-frame encoding method of prediction such as Under:

The first step, determine classified types.

In the present embodiment, multiple classified types include 7 kinds of classified types that upper left is l type and 7 kinds of divisions that upper right is l type Type.

Upper left is designated as l1～l7 for 7 kinds of classified types of l type, upper right is that 7 kinds of classified types of l type are designated as r1～r7；

As shown in fig. 7, being l1, the division schematic diagram of l2 ..., l7.In l1, square is divided into the l type of upper left With the square of bottom right, the width of the l type of upper left accounts for the 1/8 of entirely square width；The foursquare width in bottom right just accounts for entirely The 7/8 of square width；In l2, a square is divided into the l type of upper left and the square of bottom right, the width of the l type of upper left Account for the 2/8 of entirely square width；The foursquare width in bottom right accounts for the 6/8 of entirely square width；In the same manner, in l7, one just Square be divided into the l type of upper left and the square of bottom right, the width of the l type of upper left accounts for the 7/8 of entirely square width；Bottom right Foursquare width accounts for the 1/8 of entirely square width.R1, r2 ..., the division of r7 and l1, l2 ..., l7 divides with regard to vertical Direction is symmetrical, as shown in Figure 8.

Second step, according to the segmentation template of above-mentioned classified types, present encoding block is divided after, need to all sons Block carries out intra block duplication (ibc) prediction.

The Computing Principle of this step is consistent with the second step in embodiment one, therefore repeats no more, and the difference of the two mainly exists Inconsistent in the labelling of each sub-block.

In the present embodiment, after each in hunting zone is completed with reference to block search, obtain every in every kind of division The forecast error of the minimum of individual sub-block and corresponding optimal motion vector, carry out following labelling:

Upper left is that the optimal motion vector mv of two sub-blocks in upper left and bottom right of the 1st kind of classified types l1 in l type remembers respectively For mv_ibc_l1_1, mv_ibc_l1_2, corresponding forecast error is designated as sad_ibc_l1_1 and sad_ibc_l1_2；Class successively Push away, the optimal motion vector mv of two sub-blocks in the 7th kind of classified types l7 upper left and bottom right is designated as mv_ibc_l7_1 and mv_ respectively Ibc_l7_2, corresponding forecast error is designated as sad_ibc_l7_1 and sad_ibc_l7_2；

Upper right is that the optimal motion vector mv of two sub-blocks of upper right and lower-left of the 1st kind of classified types r1 in l type remembers respectively For mv_ibc_r1_1, mv_ibc_r1_2, corresponding forecast error is designated as sad_ibc_r1_1 and sad_ibc_r1_2；Class successively Push away, the optimal motion vector mv of two sub-blocks of the 7th kind of classified types r7 upper right and lower-left is designated as mv_ibc_r7_1 and mv_ respectively Ibc_r7_2, corresponding forecast error is designated as sad_ibc_r7_1 and sad_ibc_r7_2.

3rd step, dip prediction is carried out to all sub-blocks, obtain the optimum prediction direction of each sub-block, and record each son The forecast error that block is obtained by optimum prediction direction prediction.

The Computing Principle of this step is consistent with the second step in embodiment one, therefore repeats no more, and the difference of the two mainly exists Inconsistent in the labelling of each sub-block.

In the present embodiment, after 35 prediction direction traversal is completed, obtain in every kind of division each sub-block Little forecast error and corresponding intra prediction direction, carry out following labelling:

Upper left be l type in the 1st kind of classified types l1 two sub-blocks in upper left and bottom right optimum prediction direction remember respectively For mode_dip_l1_1 and mode_dip_l1_2, corresponding forecast error is designated as sad_dip_l1_1 and sad_dip_l1_2； The like, the optimum prediction direction of two sub-blocks in the 7th kind of classified types l7 upper left and bottom right is designated as mode_dip_l7_1 respectively With mode_dip_l7_2, corresponding forecast error is designated as sad_dip_l7_1 and sad_dip_l7_2；

Upper right be l type in the 1st kind of classified types r1 two sub-blocks of upper right and lower-left optimum prediction direction remember respectively For mode_dip_r1_1 and mode_dip_r1_2, corresponding forecast error is designated as sad_dip_r1_1 and sad_dip_r1_2； The like, the optimum prediction direction of two sub-blocks of the 7th kind of classified types r7 upper right and lower-left is designated as mode_dip_r7_1 respectively With mode_dip_r7_2, corresponding forecast error is designated as sad_dip_r7_1 and sad_dip_r7_2.

4th step, determine optimal dividing by quick rate-distortion optimization rdo in all classified types and corresponding predict Combination.

The Computing Principle of this step is consistent with the second step in embodiment one, therefore repeats no more, and the difference of the two mainly exists Inconsistent in the labelling of each sub-block.

In the present embodiment, in upper left be l type and upper right is to determine overall cost in the respective 7 kinds of classified types of l type Little division and prediction combination, its process is as follows:

In 7 kinds of classified types for l type l1～l7 for the upper left, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_l\_i\_1=sad\_ibc\_li\_1+sad\_dip\_li\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_li\_1,mode\_dip\_li\_2)\\ c\_l\_i\_2=sad\_ibc\_li\_2+sad\_dip\_li\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_li\_2,mode\_dip\_li\_1)\end{array};$

In above formula, i=1,2 ..., 7；Bits function is that mv and dip direction is carried out expend needed for entropy code with bit number One estimation；Find out c_l_i_1, in c_l_i_2, minimum combination, is designated as c_l_k1_p1, corresponding predictive mode be designated as (k1, P1), k1 is a number in 1 to 7, and indicating classified types is lk1；P1 is a number in 1 to 2, indicates prediction compound mode, 1 refers to that the l type part of upper left carries out ibc prediction, and the square portion of bottom right carries out dip prediction, otherwise 2；

In 7 kinds of classified types for l type r1～r7 for the upper right, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_r\_i\_1=sad\_ibc\_ri\_1+sad\_dip\_ri\_2+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ri\_1,mode\_dip\_ri\_2)\\ c\_r\_i\_2=sad\_ibc\_ri\_2+sad\_dip\_ri\_1+\\ \sqrt{\mathrm{\λ}}\×bits(mv\_ibc\_ri\_2,mode\_dip\_ri\_1)\end{array};$

Find out c_r_i_1, in c_r_i_2, minimum combination, is designated as c_r_k2_p2, corresponding predictive mode be designated as (k2, P2), k2 is a number in 1 to 7, and indicating division position is rk2；P2 is a number in 1 to 2, indicates prediction compound mode, 1 refers to that the l type part of upper right carries out ibc prediction, and the square portion of lower-left carries out dip prediction, otherwise 2.

5th step, compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, to present encoding block Carry out the Coding cost of dip prediction, and the Coding cost of optimal dividing and corresponding prediction combination, select Coding cost minimum Predictive mode as the final predictive mode of present encoding block.

This step process is consistent with the 5th step in embodiment one, and difference is also only that the labelling of some parameters is inconsistent； The process of this step is specific as follows:

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For upper left for l type classified types, by predictive mode (k1, p1), according to corresponding motion vector mv_ibc_ Lk1_p1, intra prediction directionGenerate the prediction block of current block with classified types lk1, through complete The Coding cost that rdo calculates is designated as c_l；Wherein,

For upper right for l type classified types, by predictive mode (k2, p2), according to corresponding motion vector mv_ibc_ Rk2_p2, intra prediction directionGenerate the prediction block of current block with classified types rk2, through complete The Coding cost that rdo calculates is designated as c_r；Wherein,

Compare the size of c_dip, c_ibc, c_l, c_r, corresponding for minimum code cost predictive mode is compiled as current The final predictive mode of code block.

In above-mentioned two embodiment of the present invention, encoding block is divided into by two sub-blocks by flexible block partitioning scheme so that Same piece of predictive content can comprise local message and non-local information simultaneously, and then rdo determines each block cost Little prediction.Devise quick rdo and the combination of complete rdo simultaneously, on the premise of not affecting coding efficiency, reduce coding Complexity.General effect is substantially not increase decoding end complexity, and obtains higher compression efficiency.

In order to further illustrate, the effect of above-described embodiment, also it is tested based on above-described embodiment one.Test condition Including: frame in configuration (all intra configuration, ai), quantization step (quantization step, qp) is arranged For { 22,27,32,37 }, hunting zone is | mv_x|+|mv_y| < 128, the frame number of test is 5 frames, based on software be hm12.0, Cycle testss are logical sequencings row (not including class f, because it is scc sequence) of hevc.Contrast experiment has two groups: first group It is the performance comparison that hm12.0 and hm12.0 itself is integrated with after ibc technology；Second group be hm12.0 be integrated with ibc technology with Hm12.0 is integrated with the performance comparison after our technology.This two groups of experimental results are shown in accompanying drawing 9, and the first row of accompanying drawing 9 is test Sequence names, secondary series is first group of experimental result, and the 3rd row are second group of experimental results, and last two rows are coding side and decoding The change of end complexity.It can be seen that the code check that scheme proposed by the invention can obtain 2.1% with respect to hm is saved, relatively Save in the code check that the hm being integrated with ibc can obtain 0.9%, and decoding complex degree is basically unchanged.

Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment can To be realized by software it is also possible to realize in the way of by software plus necessary general hardware platform.Based on such understanding, The technical scheme of above-described embodiment can be embodied in the form of software product, this software product can be stored in one non-easily In the property lost storage medium (can be cd-rom, u disk, portable hard drive etc.), including some instructions so that a computer equipment (can be personal computer, server, or network equipment etc.) executes the method described in each embodiment of the present invention.

The above, the only present invention preferably specific embodiment, but protection scope of the present invention is not limited thereto, Any those familiar with the art in the technical scope of present disclosure, the change or replacement that can readily occur in, All should be included within the scope of the present invention.Therefore, protection scope of the present invention should be with the protection model of claims Enclose and be defined.

Claims (10)

1. a kind of bonding position prediction and block replicate the video intra-frame encoding method of prediction it is characterised in that including:

Respectively present encoding block is divided using the predefined segmentation template comprising multiple classified types, each is drawn Present encoding block is all divided into two sub-blocks by classifying type；

All sub-blocks are carried out with intra block and replicates ibc prediction: first determine the hunting zone of ibc prediction, recycle in hunting zone Each reference block ibc prediction is carried out to all sub-blocks, determine the optimal motion vector of each sub-block, and record each sub-block by The forecast error that optimal motion vector prediction obtains；

Dip prediction is carried out to all sub-blocks, obtains the optimum prediction direction of each sub-block, and record each sub-block by optimum prediction The forecast error that direction prediction obtains；

After obtaining forecast error under ibc prediction with dip prediction for all sub-blocks, by quick rate-distortion optimization rdo in institute Have in classified types and determine optimal dividing and corresponding prediction combination；

Compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, dip prediction is carried out to present encoding block Coding cost, and the Coding cost of optimal dividing and corresponding prediction combination, select the minimum predictive mode of Coding cost As the predictive mode that present encoding block is final.

2. a kind of bonding position prediction according to claim 1 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described multiple classified types include:

Cover horizontal direction, vertical direction, diagonally opposed from upper right to lower-left, and diagonally opposed from left to bottom right Totally 28 kinds of classified types, wherein, all there are 7 kinds of classified types in each direction, the classified types system of horizontal direction and vertical direction Referred to as rectangular partition, from upper right to lower-left diagonally opposed be referred to as three from diagonally opposed classified types left to bottom right Angle divides；

Or, including 7 kinds of classified types that upper left is l type and upper right is 7 kinds of classified types of l type, both dividing mode It is all that present encoding block is divided into two sub-blocks: a sub-block is square block, and another sub-block is two pieces of adjacent rectangle blocks The l type block of composition.

3. a kind of bonding position prediction according to claim 2 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, if described multiple classified types are to cover horizontal direction, vertical direction, diagonally opposed from upper right to lower-left, with And from diagonally opposed totally 28 kinds of classified types left to bottom right, then when all sub-blocks are carried out with intra block duplication ibc prediction, Determine the optimal motion vector of each sub-block, and record the forecast error bag that each sub-block is obtained by optimal motion vector prediction Include:

7 kinds of divisions of horizontal direction are designated as h1～h7,7 kinds of divisions of vertical direction are designated as v1～v7；From upper right to lower-left 7 kinds of diagonally opposed divisions are designated as t1～t7, and the diagonally opposed 7 kind division from left to bottom right is designated as d1～d7；

When in hunting zone each with reference to block search complete after, obtained every kind of divide in each sub-block minimum pre- Survey error and corresponding optimal motion vector, carry out following labelling:

In horizontal direction, the optimal motion vector mv of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mv_ibc_h1_1 respectively With mv_ibc_h1_2, corresponding forecast error is designated as sad_ibc_h1_1 and sad_ibc_h1_2；The like, the 7th kind of division The optimal motion vector mv of upper and lower two sub-blocks of type h7 is designated as mv_ibc_h7_1 and mv_ibc_h7_2 respectively, corresponding prediction Error is designated as sad_ibc_h7_1 and sad_ibc_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_v1_1 respectively With mv_ibc_v1_2, corresponding forecast error is designated as sad_ibc_v1_1 and sad_ibc_v1_2；The like, the 7th kind of division Type v7 about the optimal motion vector mv of two sub-blocks be designated as mv_ibc_v7_1 and mv_ibc_v7_2 respectively, corresponding predict Error is designated as sad_ibc_v7_1 and sad_ibc_v7_2；

The optimal motion vector mv of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right in diagonally opposed from upper right to lower-left It is designated as mv_ibc_t1_1 and mv_ibc_t1_2 respectively, corresponding forecast error is designated as sad_ibc_t1_1 and sad_ibc_t1_ 2；The like, the optimal motion vector mv of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as mv_ibc_ respectively T7_1 and mv_ibc_t7_2, corresponding forecast error is designated as sad_ibc_t7_1 and sad_ibc_t7_2；

The optimal motion vector mv of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left from left to bottom right diagonally opposed It is designated as mv_ibc_d1_1 and mv_ibc_d1_2 respectively, corresponding forecast error is designated as sad_ibc_d1_1 and sad_ibc_d1_ 2；The like, the optimal motion vector mv of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as mv_ibc_ respectively D7_1 and mv_ibc_d7_2, corresponding forecast error is designated as sad_ibc_d7_1 and sad_ibc_d7_2.

4. a kind of bonding position prediction according to claim 3 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described optimum prediction direction dip prediction being carried out to all sub-blocks, obtaining each sub-block, and record each sub-block by The forecast error that optimum prediction direction prediction obtains includes:

Dip prediction is and determines optimum prediction direction in 35 kinds of infra-frame predictions of hevc, and the predictive value in every kind of direction generates Afterwards, calculate the forecast error of each sub-block between prediction block and present encoding block；Complete it when 35 prediction direction are traveled through Afterwards, obtain the forecast error of minimum of each sub-block and corresponding intra prediction direction in every kind of division, carry out following labelling:

In horizontal direction, the optimum prediction direction of upper and lower two sub-blocks of the 1st kind of classified types h1 is designated as mode_dip_h1_1 respectively With mode_dip_h1_2, corresponding forecast error is designated as sad_dip_h1_1 and sad_dip_h1_2；The like, draw for the 7th kind The optimum prediction direction of upper and lower two sub-blocks of classifying type h7 is designated as mode_dip_h7_1 and mode_dip_h7_2, accordingly respectively Forecast error is designated as sad_dip_h7_1 and sad_dip_h7_2；

In vertical direction the 1st kind of classified types v1 about the optimum prediction direction of two sub-blocks be designated as mode_dip_v1_1 respectively With mode_dip_v1_2, corresponding forecast error is designated as sad_dip_v1_1 and sad_dip_v1_2；The like, draw for the 7th kind Classifying type h7 about the optimum prediction direction of two sub-blocks be designated as mode_dip_v7_1 and mode_dip_v7_2 respectively, accordingly Forecast error is designated as sad_dip_v7_1 and sad_dip_v7_2；

In diagonally opposed from upper right to lower-left, the optimum prediction direction of two sub-blocks in the 1st kind of classified types t1 upper left and bottom right is divided It is not designated as mode_dip_t1_1 and mode_dip_t1_2, corresponding forecast error is designated as sad_dip_t1_1 and sad_dip_ t1_2；The like, the optimum prediction direction of two sub-blocks in the 7th kind of classified types t7 upper left and bottom right is designated as mode_ respectively Dip_t7_1 and mode_dip_t7_2, corresponding forecast error is designated as sad_dip_t7_1 and sad_dip_t7_2；

From left to bottom right diagonally opposed, the optimum prediction direction of two sub-blocks of the 1st kind of classified types d1 upper right and lower-left is divided It is not designated as mode_dip_d1_1 and mode_dip_d1_2, corresponding forecast error is designated as sad_dip_d1_1 and sad_dip_ d1_2；The like, the optimum prediction direction of two sub-blocks of the 7th kind of classified types t7 upper right and lower-left is designated as mode_ respectively Dip_d7_1 and mode_dip_d7_2, corresponding forecast error is designated as sad_dip_d7_1 and sad_dip_d7_2.

5. a kind of bonding position prediction according to claim 4 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described determine optimal dividing and corresponding prediction combination in all classified types by quick rate-distortion optimization rdo Including:

In the horizontal direction, vertical direction, diagonally opposed from upper right to lower-left and from diagonally opposed respective 7 left to bottom right Plant and in classified types, determine the minimum division of overall cost and prediction combination, its process is as follows:

In the horizontal direction in 7 kinds of classified types of h1～h7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_h\_i\_1=sad\_ibc\_hi\_1+sad\_dip\_hi\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_hi\_1,mode\_dip\_hi\_2\right)\end{array}$

$\begin{array}{c}c\_h\_i\_2=sad\_ibc\_hi\_2+sad\_dip\_hi\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_hi\_2,mode\_dip\_hi\_1\right)\end{array};$

In above formula, i=1,2 ..., 7；λ is Lagrange multiplier, to be determined by the parameter that encoder is specified, bits function It is the estimation that mv and dip direction is carried out expend needed for entropy code with bit number；Find out c_h_i_1, minimum in c_h_i_2 Combination, is designated as c_h_k1_p1, corresponding predictive mode is designated as (k1, p1), k1 is a number in 1 to 7, indicates classified types For hk1；P1 is a number in 1 to 2, indicates prediction compound mode, and 1 refers to that left-hand component carries out ibc prediction, and right-hand component enters Row dip predicts, otherwise 2；

In 7 kinds of classified types of vertical direction v1～v7, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_v\_i\_1=sad\_ibc\_vi\_1+sad\_dip\_vi\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_vi\_1,mode\_dip\_vi\_2\right)\end{array}$

$\begin{array}{c}c\_v\_i\_2=sad\_ibc\_vi\_2+sad\_dip\_vi\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_vi\_2,mode\_dip\_vi\_1\right)\end{array};$

Find out c_v_i_1 from above formula, in c_v_i_2, minimum combination, is designated as c_v_k2_p2, corresponding predictive mode is designated as (k2, p2), k2 is a number in 1 to 7, and indicating classified types is vk2；P2 is a number in 1 to 2, indicates prediction combination Mode, 1 refers to that upper part carries out ibc prediction, and lower part carries out dip prediction, otherwise 2；

In 7 kinds of divisions of the diagonally opposed t1～t7 of upper right to lower-left, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_t\_i\_1=sad\_ibc\_ti\_1+sad\_dip\_ti\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_ti\_1,mode\_dip\_ti\_2\right)\end{array}$

$\begin{array}{c}c\_t\_i\_2=sad\_ibc\_ti\_2+sad\_dip\_ti\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_ti\_2,mode\_dip\_ti\_1\right)\end{array};$

Find out c_t_i_1 from above formula, in c_t_i_2, minimum combination, is designated as c_t_k3_p3, corresponding predictive mode is designated as (k3, p3), k3 is a number in 1 to 7, and indicating classified types is tk3；P3 is a number in 1 to 2, indicates prediction combination Mode, 1 refers to that upper left carries out ibc prediction, and lower right-most portion carries out dip prediction, otherwise 2；

In 7 kinds of divisions of diagonally opposed d1～d7 left to bottom right, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_d\_i\_1=sad\_ibc\_di\_1+sad\_dip\_di\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_di\_1,mode\_dip\_di\_2\right)\end{array}$

$\begin{array}{c}c\_d\_i\_2=sad\_ibc\_di\_2+sad\_dip\_di\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_di\_2,mode\_dip\_di\_1\right)\end{array}$

Find out c_d_i_1 from above formula, in c_d_i_2, minimum combination, is designated as c_d_k4_p4, corresponding predictive mode is designated as (k4, p4), k4 is a number in 1 to 7, and indicating classified types is dk4；P4 is a number in 1 to 2, indicates prediction combination Mode, 1 refers to that upper right portion carries out ibc prediction, and bottom left section carries out dip prediction, otherwise 2.

6. a kind of bonding position prediction according to claim 5 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, present encoding block is carried out The Coding cost that dip predicts, and the Coding cost of optimal dividing and corresponding prediction combination, select the pre- of Coding cost minimum Survey pattern includes as the final predictive mode of present encoding block:

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For horizontal direction, by predictive mode (k1, p1), according to corresponding motion vector mv_ibc_hk1_p1, infra-frame prediction side ToGenerate the prediction block of current block, the coding generation calculated through complete rdo with classified types hk1 Valency is designated as c_h；Wherein,

For vertical direction, by predictive mode (k2, p2), according to corresponding motion vector mv_ibc_vk2_p2, infra-frame prediction side ToGenerate the prediction block of current block, the coding generation calculated through complete rdo with classified types vk2 Valency is designated as c_v；Wherein,

For diagonally opposed from upper right to lower-left, by predictive mode (k3, p3), according to corresponding motion vector mv_ibc_ Tk3_p3, intra prediction directionGenerate the prediction block of current block with classified types tk3, through complete The Coding cost that rdo calculates is designated as c_t；Wherein,

For diagonally opposed from left to bottom right, by predictive mode (k4, p4), according to corresponding motion vector mv_ibc_ Dk4_p4, intra prediction directionGenerate the prediction block of current block with classified types dk4, through complete The Coding cost that rdo calculates is designated as c_d；

The size of relatively c_dip, c_ibc, c_h, c_v, c_t, c_d, using corresponding for minimum code cost predictive mode as working as The final predictive mode of front encoding block.

7. a kind of bonding position prediction according to claim 2 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, if described multiple classified types include, and upper left is 7 kinds of classified types of l type and upper right is 7 kinds of classified types of l type, When then all sub-blocks are carried out with intra block duplication ibc prediction, determine the optimal motion vector of each sub-block, and record each sub-block Included by the forecast error that optimal motion vector prediction obtains:

Upper left is designated as l1～l7 for 7 kinds of classified types of l type, upper right is that 7 kinds of classified types of l type are designated as r1～r7；

When in hunting zone each with reference to block search complete after, obtained every kind of divide in each sub-block minimum pre- Survey error and corresponding optimal motion vector, carry out following labelling:

Upper left is that the optimal motion vector mv of two sub-blocks in upper left and bottom right of the 1st kind of classified types l1 in l type is designated as mv_ respectively Ibc_l1_1, mv_ibc_l1_2, corresponding forecast error is designated as sad_ibc_l1_1 and sad_ibc_l1_2；The like, the The optimal motion vector mv of two sub-blocks in 7 kinds of classified types l7 upper lefts and bottom right is designated as mv_ibc_l7_1 and mv_ibc_ respectively L7_2, corresponding forecast error is designated as sad_ibc_l7_1 and sad_ibc_l7_2；

Upper right is that the optimal motion vector mv of two sub-blocks of upper right and lower-left of the 1st kind of classified types r1 in l type is designated as mv_ respectively Ibc_r1_1, mv_ibc_r1_2, corresponding forecast error is designated as sad_ibc_r1_1 and sad_ibc_r1_2；The like, the The optimal motion vector mv of two sub-blocks of 7 kinds of classified types r7 upper rights and lower-lefts is designated as mv_ibc_r7_1 and mv_ibc_ respectively R7_2, corresponding forecast error is designated as sad_ibc_r7_1 and sad_ibc_r7_2.

8. a kind of bonding position prediction according to claim 7 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described optimum prediction direction dip prediction being carried out to all sub-blocks, obtaining each sub-block, and record each sub-block by The forecast error that optimum prediction direction prediction obtains includes:

Dip prediction is and determines optimum prediction direction in 35 kinds of infra-frame predictions of hevc, and the predictive value in every kind of direction generates Afterwards, calculate the forecast error of each sub-block between prediction block and present encoding block；Complete it when 35 prediction direction are traveled through Afterwards, obtain the forecast error of minimum of each sub-block and corresponding intra prediction direction in every kind of division, carry out following labelling:

Upper left be l type in the 1st kind of classified types l1 two sub-blocks in upper left and bottom right optimum prediction direction be designated as respectively Mode_dip_l1_1 and mode_dip_l1_2, corresponding forecast error is designated as sad_dip_l1_1 and sad_dip_l1_2；According to Secondary analogize, the optimum prediction direction of two sub-blocks in the 7th kind of classified types l7 upper left and bottom right be designated as respectively mode_dip_l7_1 with Mode_dip_l7_2, corresponding forecast error is designated as sad_dip_l7_1 and sad_dip_l7_2；

Upper right be l type in the 1st kind of classified types r1 two sub-blocks of upper right and lower-left optimum prediction direction be designated as respectively Mode_dip_r1_1 and mode_dip_r1_2, corresponding forecast error is designated as sad_dip_r1_1 and sad_dip_r1_2；According to Secondary analogize, the optimum prediction direction of two sub-blocks of the 7th kind of classified types r7 upper right and lower-left be designated as respectively mode_dip_r7_1 with Mode_dip_r7_2, corresponding forecast error is designated as sad_dip_r7_1 and sad_dip_r7_2.

9. a kind of bonding position prediction according to claim 8 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described determine optimal dividing and corresponding prediction combination in all classified types by quick rate-distortion optimization rdo Including:

In upper left be l type and upper right is to determine the minimum division of overall cost and prediction group in the respective 7 kinds of classified types of l type Close, its process is as follows:

In 7 kinds of classified types for l type l1～l7 for the upper left, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_l\_i\_1=sad\_ibc\_li\_1+sad\_dip\_li\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_li\_1,mode\_dip\_li\_2\right)\end{array}$

$\begin{array}{c}c\_l\_i\_2=sad\_ibc\_li\_2+sad\_dip\_li\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_li\_2,mode\_dip\_li\_1\right)\end{array};$

In above formula, i=1,2 ..., 7；Bits function is mv and dip direction is carried out expend needed for entropy code with bit number one Estimate；Find out c_l_i_1, in c_l_i_2, minimum combination, is designated as c_l_k1_p1, corresponding predictive mode is designated as (k1, p1), K1 is a number in 1 to 7, and indicating classified types is lk1；P1 is a number in 1 to 2, indicates prediction compound mode, and 1 is The l type part referring to upper left carries out ibc prediction, and the square portion of bottom right carries out dip prediction, otherwise 2；

In 7 kinds of classified types for l type r1～r7 for the upper right, have 14 kinds of combinations, its corresponding cost is respectively as follows:

$\begin{array}{c}c\_r\_i\_1=sad\_ibc\_ri\_1+sad\_dip\_ri\_2+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_ri\_1,mode\_dip\_ri\_2\right)\end{array}$

$\begin{array}{c}c\_r\_i\_2=sad\_ibc\_ri\_2+sad\_dip\_ri\_1+\\ \sqrt{\mathrm{\&lambda;}}\&times;bits\left(mv\_ibc\_ri\_2,mode\_dip\_ri\_1\right)\end{array};$

Find out c_r_i_1, in c_r_i_2, minimum combination, is designated as c_r_k2_p2, corresponding predictive mode is designated as (k2, p2), K2 is a number in 1 to 7, and indicating division position is rk2；P2 is a number in 1 to 2, indicates prediction compound mode, and 1 is The l type part referring to upper right carries out ibc prediction, and the square portion of lower-left carries out dip prediction, otherwise 2.

10. a kind of bonding position prediction according to claim 9 and block replicate the video intra-frame encoding method predicted, it is special Levy and be, described compared by complete rdo, present encoding block is carried out with the Coding cost of ibc prediction, present encoding block is carried out The Coding cost that dip predicts, and the Coding cost of optimal dividing and corresponding prediction combination, select the pre- of Coding cost minimum Survey pattern includes as the final predictive mode of present encoding block:

Dip prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_dip；

Ibc prediction is carried out to present encoding block, the Coding cost calculating through complete rdo is designated as c_ibc；

For upper left for l type classified types, by predictive mode (k1, p1), according to corresponding motion vector mv_ibc_lk1_ P1, intra prediction directionGenerate the prediction block of current block with classified types lk1, calculate through complete rdo Coding cost out is designated as c_l；Wherein,

For upper right for l type classified types, by predictive mode (k2, p2), according to corresponding motion vector mv_ibc_rk2_ P2, intra prediction directionGenerate the prediction block of current block with classified types rk2, through complete rdo The Coding cost calculating is designated as c_r；Wherein,

Compare the size of c_dip, c_ibc, c_l, c_r, using corresponding for minimum code cost predictive mode as present encoding block Final predictive mode.