RU2808688C2 - Method and device for image prediction - Google Patents

Abstract

FIELD: video encoding.

SUBSTANCE: potentially suitable affine block is determined by an image encoder or an image decoder. Moreover, a potentially suitable affine block is an affine-coded block that is spatially adjacent to the current image block. Motion vectors (MV) of at least two control points of the current image block are obtained in accordance with the motion vectors of at least two control points of a potentially suitable affine block. The motion vector of each pixel point or each pixel matrix of the current image block is obtained in accordance with the motion vectors of at least two control points of the current image block. Based on the motion vector of each pixel point or each pixel matrix of the current image block, the predicted values of the current image block are obtained.

EFFECT: increase of efficiency of image encoding.

13 cl, 8 dwg

Description

FIELD OF TECHNOLOGY TO WHICH THE INVENTION RELATES

[0001] The present invention relates to the field of video image processing and, in particular, to an image prediction method and apparatus.

BACKGROUND ART

[0002] With the rapid development of Internet technology and the increasing diversity in people's material and spiritual culture, the Internet requirements for video applications, especially for high-definition video applications, are increasing. However, high definition video has a very large amount of data. To transmit high-definition video over the Internet with limited bandwidth, the first problem that needs to be solved is the compression and encoding of high-definition video. Currently, there are two international organizations that are involved in developing video encoding standards, i.e., the Moving Picture Experts Group (MPEG) of the International Organization for Standardization (ISO)/International Electrotechnical Commission (IEC). International Electrotechnical Commission (IEC) and the Video Coding Experts Group (VCEG) of the International Telecommunication Union-Telecommunication standardization sector (ITU-T). MPEG, established in 1986, is responsible for developing related multimedia standards that are primarily applicable to storage, broadcasting television programs, streaming media over the Internet or wireless networks, and the like. ITU-T primarily develops video coding standards for the field of real-time video transmissions, such as radiotelephony, video conferencing, or other applications.

[0003] In recent decades, international video encoding standards have been successfully developed for various applications, mainly including: MPEG-1 standard for Video Compact Disc (VCD), MPEG-2 standard for Digital Video Disc ( digital video disc (abbreviated as “DVD”) and digital video broadcasting (abbreviated as “DVB”), the H.261 standard, the H.263 standard and the H.264 standard for video conferencing, the MPEG-4 standard, which allows encoding of objects in any form, and the latest High Efficiency Video Coding (HEVC) standard.

[0004] Compared with the latest video encoding standard, HEVC, for a sequence including rotation and zoom movements, motion-compensated prediction technology with a motion vector field based on linear changes can significantly improve encoding efficiency. In the existing motion compensation prediction technology based on affine transformation, after the affine transformation parameter of an image block is obtained, the motion vector of each point in the image block is required to be calculated, and motion compensation prediction is required to be performed according to the motion vector of the point to obtain the signal predictions with compensation for the movement of each point.

[0005] An affine transformation based motion model is a method for efficiently encoding rotation and scaling motions. The motion vector of each pixel in an image block to be processed can be obtained using the motion vectors of the four vertices of the block. For motions involving only rotation and scaling, an affine transformation model can be obtained by obtaining only the motion vectors of any two control points in the image block to be processed. The affine transformation model parameter is calculated using the motion vector of the reference point to obtain the motion vector of each pixel point in the current block. The rate-based distortion cost for the affine motion prediction mode is comparable to the rate-based distortion cost for the intra-frame prediction mode in the video coding standard. If the rate-based distortion cost of the affine motion prediction mode is less, the affine motion prediction mode is used to encode the image block to be processed. Such a block is referred to as an affine motion prediction block, and the corresponding prediction unit is referred to as an affine motion prediction unit. In addition, the motion information of the four vertices of the affine motion prediction block and the motion information of the center position are stored in a base unit with size 4x4. The center position motion vector value is the average of the motion vector values of the two control points. The motion information includes a motion vector, a reference frame index, a prediction direction, and the like.

[0006] To improve encoding efficiency, motion information of a block adjacent to an image block to be processed can be directly reused as reference point motion information in the image block to be processed without requiring additional motion search performance using the motion vector of the adjacent block as starting point for searching and encoding the motion vector difference, reducing the number of bits used to encode the motion vector difference. This method has a relatively high requirement for the accuracy of the motion vector information of the adjacent block reused for the current block, but the accuracy of the motion vector information of the adjacent block cannot be ensured in the prior art. Therefore, the coding efficiency cannot be improved.

BRIEF DESCRIPTION OF THE INVENTION

[0007] The present invention provides an image prediction method and apparatus for improving coding efficiency.

[0008] According to a first aspect, an image prediction method is provided, comprising the steps of:

obtaining a first fiducial unit for the image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit by using the same affine model;

obtaining motion information of motion compensation base units at two or more preset positions in the first reference unit; And

obtaining motion information of the motion compensation base unit of the image unit according to the motion information.

[0009] With reference to the first aspect, in a first possible implementation of the first aspect, obtaining a first reference unit for an image unit comprising the steps of:

checking, in a predetermined order, whether a predicted image for a prediction unit adjacent to the image unit is obtained using the affine model; And

if a predicted image is obtained for a prediction unit adjacent to the image unit by using said affine model, stop this check and use the prediction unit as the first reference unit.

[0010] With reference to the first aspect, in a second possible implementation of the first aspect, the preset position is the position of the corner point of the unit in the first reference unit.

[0011] With reference to any one of the first aspect, the first possible implementation of the first aspect or the second possible implementation of the first aspect, in the third possible implementation of the first aspect, said obtaining motion information of base motion compensation units at two or more preset positions in the first reference unit comprises obtaining motion information of motion compensation base units at three preset positions in the first reference unit, and,

Accordingly, said acquisition of motion information of an image unit motion compensation base unit according to the motion information comprises a step of obtaining a prediction direction, a reference frame index, and a motion vector of an image unit motion compensation base unit in accordance with the prediction directions, reference frame indices, and motion vectors. base motion compensation units at these three preset positions in the first reference unit.

[0012] With reference to the third possible implementation of the first aspect, in the fourth possible implementation of the first aspect, said obtaining the prediction direction, reference frame index and motion vector of the base unit of image unit motion compensation in accordance with the prediction directions, reference frame indices and motion vectors of the base units compensation of movement in three preset positions in the first reference unit contains the stages in which:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at the three preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at the three preset positions in the first reference unit; And

obtaining a motion vector of the image unit motion compensation base unit by interpolation in accordance with the motion vectors of the motion compensation base units at the three preset positions in the first reference unit.

[0013] With reference to the fourth possible implementation of the first aspect, in the fifth possible implementation of the first aspect, the said acquisition of the motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at three preset positions in the first reference unit is implemented in the following way:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , , And , , are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at said three preset positions in the first reference unit, and , , And , , are the corresponding horizontal and vertical motion vectors.

[0014] With reference to any one of the first aspect, the first possible implementation of the first aspect or the second possible implementation of the first aspect, in the sixth possible implementation of the first aspect, said obtaining motion information of base motion compensation units at two or more preset positions in the first reference unit comprises obtaining motion information of motion compensation base units at two preset positions in the first reference unit; And

Accordingly, said acquisition of motion information of an image unit motion compensation base unit according to the motion information comprises a step of obtaining a prediction direction, a reference frame index, and a motion vector of an image unit motion compensation base unit in accordance with the prediction directions, reference frame indices, and motion vectors. base motion compensation units at these two preset positions in the first reference unit.

[0015] With reference to the sixth possible implementation of the first aspect, in the seventh possible implementation of the first aspect, said obtaining the prediction direction, reference frame index and motion vector of the base unit of image unit motion compensation in accordance with the prediction directions, reference frame indices and motion vectors of the base units compensation of movement in two preset positions in the first reference unit contains the stages in which:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at the two preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at the two preset positions in the first reference unit; And

obtaining a motion vector of the image unit motion compensation base unit by interpolation in accordance with the motion vectors of the motion compensation base units at the two preset positions in the first reference unit.

[0016] With reference to the seventh possible implementation of the first aspect, in the eighth possible implementation of the first aspect, the said acquisition of the motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at two preset positions in the first reference unit is implemented in the following way:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , And , are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at said two preset positions in the first reference unit, and , And , are the corresponding horizontal and vertical motion vectors.

[0017] According to a second aspect, there is provided an image prediction device including:

a first acquisition module configured to obtain a unit for the image unit, wherein corresponding predicted images are obtained for the image unit and the first reference unit by using the same affine model;

a second acquisition module configured to receive motion information of motion compensation reference units at two or more preset positions in the first reference unit; And

a third acquisition module configured to obtain motion information of the base unit of motion compensation of the image unit in accordance with the position information and said motion information.

[0018] With reference to the second aspect, in a first possible implementation of the second aspect, the first acquisition module is configured:

check, in a predetermined order, whether a predicted image is obtained for a prediction unit adjacent to the image unit using the affine model; And

if a predicted image is obtained for a prediction unit adjacent to the image unit using the affine model, stop this check and use the prediction unit as the first reference unit.

[0019] With reference to the second aspect, in a second possible implementation of the second aspect, the preset position is the position of the corner point of the unit in the first reference unit.

[0020] With reference to any one of the second aspect, the first possible implementation of the second aspect, or the second possible implementation of the second aspect, in the third possible implementation of the second aspect, the third acquisition module is configured to receive motion information of base motion compensation units at three preset positions in the first reference unit, and

accordingly, said acquisition of motion information of the image unit motion compensation base unit according to the motion information comprises obtaining a prediction direction, a reference frame index, and a motion vector of the image unit motion compensation base unit according to the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at these three preset positions in the first reference unit.

[0021] With reference to the third possible implementation of the second aspect, in the fourth possible implementation of the second aspect, the third acquisition module is further configured:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at the three preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at the three preset positions in the first reference unit; And

obtain a motion vector of the image unit motion compensation base unit by interpolation in accordance with motion vectors of the motion compensation base units at said three preset positions in the first reference unit.

[0022] With reference to the fourth possible implementation of the second aspect, in the fifth possible implementation of the second aspect, the said acquisition of the motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at three preset positions in the first reference unit is implemented in the following way:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , , And , , are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at said three preset positions in the first reference unit, and , , And , , are the corresponding horizontal and vertical motion vectors.

[0023] With reference to any one of the second aspect, the first possible implementation of the second aspect or the second possible implementation of the second aspect, in the sixth possible implementation of the second aspect, the third acquisition module is further configured to receive motion information of the basic motion compensation units at two preset positions in the first reference unit; And

accordingly, said acquisition of motion information of the image unit motion compensation base unit according to the motion information comprises obtaining a prediction direction, a reference frame index, and a motion vector of the image unit motion compensation base unit according to the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at these two preset positions in the first reference unit.

[0024] With reference to the sixth possible implementation of the second aspect, in the seventh possible implementation of the second aspect, the third acquisition module is further configured:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at said two preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at the two preset positions in the first reference unit; And

obtain a motion vector of the image unit motion compensation base unit by interpolation in accordance with the motion vectors of the motion compensation base units at said two preset positions in the first reference unit.

[0025] With reference to the seventh possible implementation of the second aspect, in the eighth possible implementation of the second aspect, the said acquisition of the motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at two preset positions in the first reference unit is implemented in the following way:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , And , are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at said two preset positions in the first reference unit, and , And , are the corresponding horizontal and vertical motion vectors.

[0026] Based on the above-described technical solutions, according to the image prediction method and apparatus in embodiments of the present invention, the motion information of the first reference unit using the same affine prediction model is reused, and therefore a more accurate motion vector of the current image unit is obtained without encoding and decoding the difference of motion vectors and without additional search for a high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the complexity of encoding and decoding, thereby increasing the efficiency of encoding and decoding.

BRIEF DESCRIPTION OF THE DRAWINGS

[0027] To more clearly describe the technical solutions in the embodiments of the present invention, the following is a brief description of the accompanying drawings necessary to describe the embodiments or prior art. It will be appreciated that the accompanying drawings in the following description show only some embodiments of the present invention, and one skilled in the art will be able to derive other drawings from the accompanying drawings without creative effort.

[0028] FIG. 1 is a flowchart of an image prediction method according to an embodiment of the present invention;

[0029] FIG. 2 is a schematic representation of potentially suitable prediction units (candidate prediction units) in accordance with an embodiment of the present invention;

[0030] FIG. 3 is a schematic diagram of an affine image block and reference points in accordance with an embodiment of the present invention;

[0031] FIG. 4 is another flowchart of an image prediction method according to an embodiment of the present invention;

[0032] FIG. 5 is a block diagram of an image prediction apparatus according to an embodiment of the present invention;

[0033] FIG. 6 is another block diagram of an image prediction apparatus according to an embodiment of the present invention;

[0034] FIG. 7 is another block diagram of an image prediction apparatus according to an embodiment of the present invention; And

[0035] FIG. 8 is another block diagram of an image prediction apparatus according to an embodiment of the present invention.

DESCRIPTION OF IMPLEMENTATION OPTIONS

[0036] The following describes in completeness and clarity the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. It will be appreciated that the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments derived by one skilled in the art from the embodiments of the present invention without creative efforts will fall within the scope of protection of the present invention.

[0037] It should be understood that, in embodiments of the present invention, an affine rotation and scale model based on two control points in each prediction unit (four parameters) is used, or an affine rotation and scale model based on three control points in each prediction unit (six parameters), or a forward affine model based on four control points in each prediction unit (eight parameters) or the like may be used, and no restrictions are placed on it.

[0038] It should be understood that the basic motion compensation unit is the minimum unit for performing motion compensation, and the unit has uniform motion information. In embodiments of the present invention, a pixel dot is used as a basic motion compensation unit, or a pixel matrix or the like. can be used as a basic motion compensation unit, and no restriction is placed on it.

[0039] It should be understood that in a 4-parameter affine model, from the motion information of any two motion compensation base units, the motion information of all motion compensation units in the motion prediction affine unit in which the two motion compensation base units are localized can be interpolated. Preferably, more accurate interpolated motion information can be obtained by selecting motion compensation base units at any two corner points of the unit in the affine motion prediction unit. The position of the corner point of the unit is the point at which two adjacent sides of the affine motion prediction unit, that is, the first reference unit, converge. For example, if the affine motion prediction unit is a triangle, the corner points of the unit are the three vertices of the triangle; if the affine motion prediction unit is a quadrilateral, the corner points of the unit are the four vertices of the quadrilateral, etc. In embodiments, for ease of description, the basic motion compensation units from which the motion information of all motion compensation units in the affine motion prediction unit are interpolated are referred to as reference points.

[0040] An embodiment of the present invention provides an image prediction method that is implemented by a decoder. The method, in particular, includes the following steps:

obtaining a first fiducial unit for the image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit using the same affine model;

obtaining motion information of motion compensation base units at two or more predetermined positions in the first reference unit; And

obtaining motion information of the motion compensation base unit of the image unit according to the motion information.

[0041] According to the image prediction method and apparatus in this embodiment of the present invention, the motion information of the first reference unit using the same affine motion prediction model is reused, and therefore a more accurate motion vector of the current image unit is obtained without encoding and decoding differences between motion vectors and without additional search for a high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the complexity of encoding and decoding, thereby increasing the efficiency of encoding and decoding.

[0042] Preferably, the preset position is a position of a corner point of a unit in the first reference unit, and the position of the corner point is a point at which two adjacent sides of the first reference unit converge.

[0043] Preferably, for example, obtaining motion information of motion compensation base units at two or more preset positions in the first reference unit includes:

obtaining motion information of motion compensation base units at three preset positions in the first reference unit; and accordingly, obtaining motion information of the base image unit motion compensation unit in accordance with the motion information includes: obtaining the prediction direction, reference frame index and motion vector of the base image unit motion compensation unit in accordance with the prediction directions, reference frame indices and motion vectors of the base units motion compensation units at three preset positions in the first reference unit.

[0044] Obtaining the prediction direction, reference frame index, and motion vector of the motion compensation base unit of an image unit according to the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at three preset positions in the first reference unit may include:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit; assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit; and obtaining a motion vector of the motion compensation base unit of the image unit by interpolation according to the motion vectors of the motion compensation base units at three preset positions in the first reference unit.

[0045] Obtaining the motion vector of the image unit motion compensation base unit by interpolation in accordance with the motion vectors of the motion compensation base units at three preset positions in the first reference unit is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at three preset positions in the first reference unit, and , And And , And are the corresponding horizontal and vertical motion vectors. [0046] The above formulas that are used to obtain the motion vector of the image unit motion compensation base unit by interpolation are conventional formulas that are used to perform interpolation according to the motion information of the motion compensation base units at any selected preset positions in the first reference unit . In a particular embodiment, the formulas can be simplified for different application scenarios in accordance with the characteristics of the selected preset provisions. For example, in a specific embodiment of the present invention, the upper left, lower left and upper right corner points of the unit in the unit are selected as preset positions, and the formulas are simplified to the following form:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0047] Alternatively, for example, the method includes:

obtaining motion information of motion compensation base units at two preset positions in the first reference unit; and accordingly, obtaining motion information of the base image unit motion compensation unit in accordance with the motion information includes: obtaining the prediction direction, reference frame index and motion vector of the base image unit motion compensation unit in accordance with the prediction directions, reference frame indices and motion vectors of the base units motion compensation units at two preset positions in the first reference unit.

[0048] Obtaining the motion vector of the image unit motion compensation base unit by interpolation in accordance with the motion vectors of the motion compensation base units at two preset positions in the first reference unit is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, And And And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at two preset positions in the first reference unit, and And And And are the corresponding horizontal and vertical motion vectors.

[0049] The above formulas that are used to obtain the motion vector of the image unit motion compensation base unit by interpolation are conventional formulas that are used to perform interpolation according to the motion information of the motion compensation base units at any selected preset positions in the first reference unit . In a particular embodiment, the formulas can be simplified for different application scenarios in accordance with the characteristics of the selected preset provisions. For example, in a specific embodiment of the present invention, the upper left and upper right corner points of the unit in the unit are selected as preset positions, and the formulas are simplified to the following form:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the affine motion prediction unit, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0050] Moreover, the present invention further provides an image prediction method applicable to an encoding process. The image prediction method applicable to the encoding process is the same as the above-described image prediction method applicable to the decoding process. Therefore, the same result of improving encoding efficiency can be obtained.

[0051] The following describes specific implementations of the present invention using specific examples.

[0052] FIG. 1 is a flowchart of an image prediction method according to an embodiment of the present invention. The method shown in FIG. 1 may be performed by a decoding device such as a decoder.

[0053] More specifically, as shown in FIG. 1, method 1000 includes the following steps.

[0054] S1100. Reading the bitstream and parsing the bitstream to obtain a syntax element, wherein the syntax element is used to determine whether an image block to be processed reuses motion information of an affine model of an adjacent block. When it is determined that the image block to be processed reuses the motion information of the affine model of the adjacent block, the process proceeds to step S1200.

[0055] S1200. Definition of an affine motion prediction unit that is used as a reference for the image block to be processed.

[0056] The first specific method of implementing step 1210 includes the following steps:

[0057] S1211. Determining a set of candidate prediction units for the image block to be processed.

[0058] As shown in FIG. 2, in this embodiment of the present invention, prediction units in which 4x4 blocks at five positions A, B, C, D and E that are spatially adjacent to the image block to be processed are respectively localized and selected to form a set. It should be understood that, alternatively, some of the positions A, B, C, D and E may be selected, or prediction units in which other blocks adjacent in the time domain to the image block to be processed may be selected to form a set , and no restrictions are placed on it.

[0059]S1212. Checking, in a predetermined order, whether a prediction unit in a set is an affine motion prediction unit.

[0060] A predetermined order is an order that is predetermined in the protocol by the encoding and decoding parties and is kept consistent between the encoding and decoding parties. In this embodiment of the present invention, the preset order is A, B, C, D and E. It should be understood that, alternatively, the preset order may be B, D, A, E and C or other order, and no restrictions do not overlap with this.

[0061] S1213. If the prediction unit is found to be an affine motion prediction unit, the check is stopped and the prediction unit is used as the affine motion prediction unit referenced by the image block to be processed.

[0062] A second specific method of implementing step 1220 includes the following steps.

[0063] S1221. Determining a set of candidate prediction units for the image block to be processed.

[0064] In addition to the same set determination method as in S1211, the method may further include: removing from the set a non-affine motion prediction unit in the set. For example, the set determined using the determination method according to S1211 includes prediction units in which A, B, C, D and E are located; and the prediction units in which C and D are localized are non-affine motion prediction units. Therefore, C and D are removed from the set, and a particular set of candidate prediction units includes prediction units in which A, B, and E are located.

[0065] Alternatively, the method may include: limiting the capacity of the set and removing from the set a candidate prediction unit for which the test order ranks outside the allowed number of candidate prediction units. For example, the number of candidate prediction units in a set is set to 2 at most, and the set that corresponds to the test order and which is determined using the determination method of S1211 in accordance with the test order of S1212 is A, D, C, E, and B. Therefore, C, E and B are removed from the set, and the particular set of candidate prediction units includes prediction units in which A and D are located. It should be understood that the capacity of the set may be set to 3 or 4, or may not be limited if the capacity is within a range that is greater than 0 but less than or equal to the maximum number of candidate prediction units determined using the determination method of S1211.

[0066] Alternatively, the above two set determination methods can be used in combination.

[0067] S1222. Reading the bitstream and parsing the bitstream to obtain a syntax element, wherein the syntax element is used to indicate index information of an affine motion prediction unit that is in a set of candidate prediction units, and which is used as a reference for an image block to be processed.

[0068] S1223. Determining, in accordance with the index information, an affine motion prediction unit that is used as a reference for the image block to be processed.

[0069] For example, if the set of candidate prediction units includes two candidate prediction units A and B, the index of A is 0, and the index of B is 1. The bit stream is read and parsed to obtain index information. If the index information is 0, the affine motion prediction unit referred to by the image block to be processed is the prediction unit in which A is localized; if the index information is 1, the affine motion prediction unit referred to by the image block to be processed is the prediction unit in which B is localized.

[0070] S1300. Obtaining position information and motion information of control points in an affine motion prediction unit.

[0071] The first specific method of implementing step 1310 includes: selecting any three reference points from the four reference points in the motion prediction affine unit, and obtaining position information, motion vector information, prediction directions, and reference frame indices of the three reference points. Since all pixel points in the same affine motion prediction unit have the same prediction direction and the same reference frame index, the prediction directions and reference frame indices of all reference points are also the same.

[0072] A second specific method of implementing step 1320 includes: selecting any two control points, which may be referred to as the first and second control points, from the four control points in the motion prediction affine unit, and obtaining position information, motion vector information, prediction directions, and indexes of reference frames of two control points; and then selecting any control point, which may be referred to as a third control point, from the remaining two control points and obtaining position information of the third control point. The motion vector information of the third reference point can be obtained in accordance with the first and second reference points as follows:

Where And are, respectively, the horizontal and vertical coordinates of the third control point, And are the corresponding horizontal and vertical motion vectors, And And And are, respectively, the horizontal and vertical coordinates of the first and second control points, and And And And are the corresponding horizontal and vertical motion vectors.

[0073] Since all pixel points in the same affine motion prediction unit have the same prediction direction and the same reference frame index, the prediction directions and reference frame indices of all reference points are also the same. The third reference point has the same prediction direction and reference frame index as those of the first and second reference points.

[0074] According to either of the previous two specific implementations, after step S1300, motion information and position information of at least three reference points are obtained.

[0075] As shown in FIG. 3, in this embodiment of the present invention, the selected control points are represented by three vertices: the upper left (x ₂ , y ₂ ), upper right (x ₃ , y ₃ ) and lower left (x ₄ , y ₄ ) vertices of the affine prediction unit motion in which the adjacent block A of the image block to be processed is localized. In this embodiment, x ₂ is equal to x ₄ , and y ₂ is equal to y ₃ .

[0076] It should be understood that the reference point is selected randomly and no restrictions are placed on it.

[0077] S1400. Obtaining motion information of a pixel point in an image block in accordance with the position information and motion information of the above three selected reference points.

[0078] The first specific method of implementing step 1410 includes the following steps:

[0079]S1411. Since the prediction directions and reference frame indices of all reference points are the same, the prediction direction and reference frame index of any one of the reference points are assigned to a pixel point in the image block to be processed. Alternatively, the bit stream may be parsed to obtain reference frame information of an image block to be processed, where the reference frame information includes a prediction direction and a reference frame index.

[0080] S1412. Obtaining, in accordance with the position information and motion vectors obtained in S1300, any three reference points, the motion vector of a pixel point in the image block to be processed. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of any three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0081] A second specific method of implementing step 1420 includes the following steps.

[0082] S1421. Since the prediction directions and reference frame indices of all reference points are the same, the prediction direction and reference frame index of any one of the reference points are assigned to a pixel point in the image block to be processed. Alternatively, the bitstream may be parsed to obtain reference frame information of an image block to be processed, wherein the reference frame information includes a prediction direction and a reference frame index.

[0083] S1422. As shown in FIG. 3, obtaining, according to the position information and motion vectors obtained in S1300, any three reference points, the motion vectors of any two reference points in the image block to be processed. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0084]S1423. Obtaining, in accordance with the motion vectors obtained in S1422, the two reference points, the motion vector of the pixel point in the image block to be processed. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0085] A third specific method of implementing step 1430 includes the following steps.

[0086] S1431. Since the prediction directions and reference frame indexes of all reference points are the same, the prediction direction and reference frame index of any one of the reference points are assigned to a pixel point in the image block to be processed. Alternatively, the bitstream may be parsed to obtain reference frame information of an image block to be processed, wherein the reference frame information includes a prediction direction and a reference frame index.

[0087] S1432. As shown in FIG. 3, obtaining, in accordance with the position information and motion vectors obtained in S1400, any three reference points, the motion vectors of any two reference points in the image block to be processed. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0088] S1433. Parses the bitstream to obtain the remainder of the motion vectors of the two reference points in the image block to be processed and adds the remainder to the motion vectors obtained in S1432 of the two reference points to obtain the updated motion vectors of the two reference points in the image block to be processed. Due to the addition of the remainder, the updated motion vectors are more accurate.

[0089] S1434. Obtaining, in accordance with the motion vectors obtained in S1433, the two reference points, the motion vector of the pixel point in the image block to be processed. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0090] S1500. Performing motion compensation in accordance with the prediction direction, reference frame index, and motion vector that are obtained to find a predicted value of a pixel point in an image block to be processed, wherein the predicted values of all pixel points form a predicted image of the image to be processed.

[0091] It should be understood that, in this embodiment of the present invention, the terms “first”, “second”, “third” and “fourth” are intended only to distinguish pixel points, but should not be construed as any limitation on the scope of protection of the present invention. For example, the first control point may also be referred to as the second control point, and the second control point may be the first control point, etc.

[0092] It should also be understood that the sequence numbers of the above processes do not indicate execution sequences. The execution sequences of the processes should be determined in accordance with the functions and internal logic of the processes and should not be construed as any limitation of the processes for implementing this embodiment of the present invention.

[0093] It should also be understood that a method for obtaining a predicted image using a 6-parameter affine transform model or an 8-parameter affine transform model is substantially the same as a method for obtaining a predicted image using a 4-parameter affine transform model. For brevity, details are not described here.

[0094] More specifically, the decoding side reuses the motion information of an affine motion prediction unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additionally searching for a high-precision motion vector . In this way, the prediction accuracy is improved while maintaining the complexity of encoding and decoding, thereby increasing the efficiency of encoding and decoding.

[0095] To better demonstrate the useful results of the present invention, a performance comparison is made between this embodiment of the present invention and a prediction method using a two-checkpoint based affine transformation model according to the prior art with respect to a Random Access encoding configuration. and low delay encoding configurations. A negative value indicates the percentage of improvement in the encoding performance, and a positive value indicates the percentage of deterioration in the encoding performance.

Table 1. Affine sequence

[0096] It can be seen that, under test conditions corresponding to low bit rate (low bit rate) random access, high bit rate random access, low bit rate low latency and high bit rate low latency, the present invention correspondingly reduces the bit rate by 2.7%, 2.1%, 5.7% and 4.3% when the complexity remains essentially the same, thereby increasing coding efficiency.

[0097] The image prediction method implemented on the decoding side according to this embodiment of the present invention has been described in detail above with reference to FIGS. 1 - fig. 3, and the image prediction method implemented on the encoding side according to an embodiment of the present invention will now be described in detail with reference to FIG. 2 - fig. 4. It should be noted that the corresponding operations on the encoding side are essentially consistent with those on the decoding side. To avoid repetition, the details are not described here.

[0098] FIG. 4 is another flowchart of an image prediction method 2000 according to an embodiment of the present invention. The method shown in FIG. 4 may be performed by an encoding device such as an encoder.

[0100] More specifically, as shown in FIG. 4, method 2000 includes the following steps.

[0101] S2100. Definition of an affine motion prediction unit that is used as a reference for the image block to be processed.

[0102] The first specific method of implementing step 2110 includes the following steps.

[0103] S2111. Determining a set of candidate prediction units for the image block to be processed.

[0104] As shown in FIG. 2, in this embodiment of the present invention, prediction units in which 4x4 blocks are respectively located at five positions A, B, C, D and E that are spatially adjacent to the image block to be processed are selected to form a set. It should be understood that, alternatively, some of the positions A, B, C, D and E may be selected to form the set, or prediction units may be selected in which other blocks are located adjacent in the time domain to the image block to be processed, and no restrictions are placed on this.

[0105] S2112. Checks, in a predetermined order, whether a prediction unit in a set is an affine motion prediction unit.

[0106] A predetermined order is an order that is predetermined in the protocol by the encoding and decoding parties and is kept consistent between the encoding and decoding parties. In this embodiment of the present invention, the preset order is A, B, C, D and E. It should be understood that, alternatively, the preset order may be B, D, A, E and C or other order, and no restrictions do not overlap with this.

[0107] S2113. If the prediction unit is found to be an affine motion prediction unit, the check is stopped and the prediction unit is used as the affine motion prediction unit referenced by the image block to be processed.

[0108] S2114. If none of the candidate prediction units is an affine motion prediction unit, then a syntax element is encoded, the syntax element is used to indicate that the image block to be processed does not reuse the affine prediction information of the adjacent block, and execution of the subsequent step is stopped; otherwise, the next step continues.

[0109] A second specific implementation of step 2120 includes the following steps:

[0110] S2121. Determining a set of candidate prediction units for the image block to be processed.

[0111] In addition to the same set determination method as in S2111, the method may further include: removing from the block set a non-affine motion prediction unit in the set. For example, the set determined using the determination method according to S1211 includes prediction units in which A, B, C, D and E are located; and the prediction units in which C and D are localized are not affine motion prediction units. Therefore, C and D are removed from the set, and the determined set of candidate prediction units includes prediction units in which A, B, and E are located.

[0112] Alternatively, the method may include: limiting the capacity of the set and removing from the set a candidate prediction unit for which the test order ranks outside the allowed number of candidate prediction units. For example, the number of candidate prediction units in a set is set to 2 at most, and the set that is determined using the determination method of S2112 according to the checking order of S2112 is A, D, C, E, and B. Therefore, C, E, and B are removed from the set, and the particular set of candidate prediction units includes the prediction units in which A and D are located. It will be appreciated that the capacity of the set may be set to 3 or 4, or may not be limited if the capacity is within the range , which is greater than 0 but less than or equal to the maximum number of candidate prediction units determined using the determination method of S1211.

[0113] Alternatively, the above two set determination methods can be used in combination.

[0114] S2122. If none of the candidate prediction units is an affine motion prediction unit, then a syntax element is encoded, the syntax element is used to indicate that the image block to be processed does not reuse the affine prediction information of the adjacent block, and execution of the subsequent step is stopped; if at least one of the candidate prediction units is an affine motion prediction unit, then the subsequent step continues.

[0115] S2123. Assigning an index value to each candidate prediction unit in the set of candidate prediction units. The candidate prediction unit indicated by each index value is predetermined in the protocol by the encoding and decoding sides and is kept consistent between the encoding and decoding sides.

[0116] S2124. Iterates through all the candidate prediction units in the set of candidate prediction units and performs processing in accordance with subsequent steps S2200 to S2400 to obtain a predicted image of the image block to be processed and performs encoding for the image block to be processed; and comparing the encoding gains obtained by the selected candidate prediction vectors, and selecting the candidate prediction unit with the highest encoding gain as the affine motion prediction unit referred to by the image block to be processed.

[0117] For example, if a set of candidate prediction units includes two candidate prediction units A and B, the index of A is 0 and the index of B is 1. If the gain that is obtained using A as the affine motion prediction unit to which refers to the image block to be processed, to perform a subsequent processing step, obtain a predicted image of the image block to be processed, and perform encoding for the image block to be processed is the largest, then A is the affine motion prediction unit referred to by the image block to be processed; otherwise, B is the affine motion prediction unit referred to by the image block to be processed.

[0118] It should be understood that to improve the encoding speed, step S2124 can be replaced with a fast algorithm. For example, some, but not all, of the candidate prediction units in a set of candidate prediction units may be selected for crawling. Alternatively, the entire encoding process of the image block to be processed is not performed, and after obtaining the predicted image, according to the accuracy of the predicted images, a prediction unit that can be used to obtain the most accurate predicted image is used as an affine motion prediction unit referred to image block to be processed.

[0119] S2200. Obtaining position information and motion information of control points in an affine motion prediction unit. This step is consistent with S1300 and the details are not re-described.

[0120] S2300. Obtaining motion information of a pixel point in an image block in accordance with the position information and motion information of the above three selected reference points. This step is consistent with S1400 and the details are not re-described.

[0121] S2400. Performing motion compensation in accordance with the prediction direction, reference frame index, and motion vector that are obtained to find a predicted value of a pixel point in an image block to be processed, wherein the predicted values of all pixel points form a predicted image of the image to be processed. This step is consistent with S1500 and the details are not re-described.

[0122] S2500. Comparing the encoding characteristics that are performed using the predicted image obtained using steps S2100 to S2400 with the encoding characteristics for encoding that is performed using the predicted image obtained using other methods, and selecting the method in which the largest encoding gain is obtained, as a method for encoding a predicted image.

[0123] It should be understood that to improve encoding speed, S2500 can be replaced by the fast algorithm described in S2124, and the details are not described again.

[0124] S2600. If a method for obtaining a predicted image using steps S2100 to S2400 is used in S2500, a syntax element is encoded, wherein the syntax element is used to indicate that an image block to be processed obtains a predicted image using a method for obtaining a predicted image using steps S2100 to S2400, and another syntax element is encoded, the syntax element being used to indicate index information of the prediction unit selected in S2124.

[0125] It should be understood that, the image prediction method in this embodiment of the present invention, as a prediction technology, can also be technically combined with another prediction technology according to the prior art, for example with a technology using multiple coordinate systems for prediction, and none no restrictions are placed on this.

[0126] More specifically, the encoding side reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additionally searching for a high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0127] The following describes in detail a decoding apparatus in accordance with an embodiment of the present invention with reference to FIGS. 5. As shown in FIG. 5, the decoding device 10 includes:

a first acquisition module 11 configured to obtain an affine image unit motion prediction unit;

a second acquisition module 12 configured to obtain position information and motion information of the affine motion prediction unit; And

a third acquisition module 13 configured to acquire motion information of an image unit in accordance with the position information and the motion information.

[0128] More specifically, the encoding side reuses the motion information of the first reference unit using the same affine motion prediction model and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additionally searching for a high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0129] In this embodiment of the present invention, optionally, the bitstream is read and parsed to obtain a syntax element, wherein the syntax element is used to determine whether an image block to be processed reuses motion affine model information of an adjacent block. When it is determined that the image block to be processed reuses the motion information of the affine model of the adjacent block, the process continues.

[0130] In this embodiment of the present invention, optionally, the first acquisition module 11 is specifically configured to determine an affine motion prediction unit that is used as a reference for an image block to be processed, including: determining a set of candidate prediction units for an image block to be processed, checking, in a predetermined order, whether the prediction unit in the set is an affine motion prediction unit, and if it is found that the prediction unit is an affine motion prediction unit, stopping the check and using the prediction unit as the affine motion prediction unit, referenced by the image block to be processed; or determining a set of candidate prediction units for an image block to be processed, reading the bit stream, and parsing the bit stream to obtain a syntax element, wherein the syntax element is used to indicate index information of an affine motion prediction unit that is in the set of candidate prediction units and that is used as a reference for the image block to be processed, and determining, in accordance with the index information, an affine motion prediction unit that is used as a reference for the image block to be processed.

[0131] In this embodiment of the present invention, optionally, the second acquisition module 12 is specifically configured to obtain position information and motion information of the reference points in the affine motion prediction unit includes: selecting any three reference points from the four reference points in an affine motion prediction unit and obtaining position information, motion vector information, prediction directions, and reference frame indices of the three reference points; or selecting any two reference points, which may be referred to as the first and second reference points, from the four reference points in the affine motion prediction unit and obtaining position information, motion vector information, prediction directions and reference frame indices of the two reference points and then selecting any reference point , which may be referred to as a third control point, from the remaining two control points and obtaining position information of the third control point. The motion vector information of the third reference point can be obtained in accordance with the first and second reference points as follows:

Where And are, respectively, the horizontal and vertical coordinates of the third control point, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of the first and second control points, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0132] Since all pixel points in the same affine motion prediction unit have the same prediction direction and the same reference frame index, the prediction directions and reference frame indices of all reference points are also the same. The third reference point has the same prediction direction and reference frame index as those of the first and second reference points.

[0133] In this embodiment of the present invention, optionally, the third acquisition unit 13 is specifically configured to obtain motion information of a pixel point in an image block according to the position information and motion information of the above three selected reference points, including: assignment the prediction direction and index of the reference frame of any one of the control points of the pixel point in the image block to be processed; or parsing the bit stream to obtain reference frame information of an image block to be processed, the reference frame information including a prediction direction and a reference frame index.

[0134] In this embodiment of the present invention, optionally, the third acquiring unit 13 is specifically configured to obtain motion information of a pixel point in an image block according to the position information and motion information of the above three selected reference points, including: acquiring , according to the received position information and motion vectors of any three control points, the motion vector of a pixel point in the image block to be processed. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of any three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0135] Alternatively, the motion vectors of any two reference points in the image block to be processed are obtained in accordance with the acquired position information and the motion vectors of any three reference points. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, And are, respectively, the horizontal and vertical coordinates of the control point in the image block to be processed, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0136] The motion vector of a pixel point in the image block to be processed is obtained in accordance with the acquired motion vectors of the two reference points. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0137] Alternatively, motion vectors of any two reference points in the image block to be processed are obtained in accordance with the acquired position information and motion vectors of any three reference points. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0138] The bit stream is parsed to obtain the remainder of the motion vector of the two reference points in the image block to be processed, and the remainder is added to the resulting motion vector of the two reference points to obtain an updated motion vector of the two reference points in the image block to be processed.

[0139] The motion vector of a pixel point in the image block to be processed is obtained in accordance with the obtained motion vectors of the two reference points. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0140] In this embodiment of the present invention, optionally, the third acquisition unit 13 is specifically configured to perform motion compensation in accordance with the prediction direction, reference frame index, and motion vector that are acquired to find the predicted value of a pixel point in an image block to be processed. The predicted values of all pixel points form a predicted image of the image to be processed.

[0141] It should be understood that the encoding device 10 according to this embodiment of the present invention can accordingly execute the image prediction method 1000 in the embodiment of the present invention, and the above and other operations and/or functions of the modules of the encoding device 10 are separately used to implement the corresponding the procedure of the method according to Fig. 1. For brevity, the details are not described here.

[0142] Therefore, the decoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0143] The following describes in detail an encoding apparatus according to an embodiment of the present invention with reference to FIG. 6. As shown in FIG. 6, the encoding device 20 includes:

a first acquisition unit 21 configured to obtain an affine image unit motion prediction unit;

a second acquisition module 22 configured to obtain position information and motion information of the affine motion prediction unit; And

a third acquisition module 23 configured to acquire motion information of an image unit in accordance with the position information and the motion information.

[0144] More specifically, the encoding side reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additionally searching for a high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0145] In this embodiment of the present invention, optionally, the first acquisition unit 21 is specifically configured to determine an affine motion prediction unit that is used as a reference for an image block to be processed, including: determining a set of candidate prediction units for an image block to be processed, checking, in a predetermined manner, that the prediction unit in the set is an affine motion prediction unit, and if it is found that the prediction unit is an affine motion prediction unit, stopping the check and using the prediction unit as the affine motion prediction unit, referenced by the image block to be processed, or if none of the candidate prediction units is an affine motion prediction unit, encoding a syntax element, wherein the syntax element is used to indicate that the image block to be processed does not reuse affine prediction information of an adjacent block, and stopping the execution of the subsequent stage; otherwise, continue with the next step; or determining a set of candidate prediction units for the image block to be processed, if none of the candidate prediction units is an affine motion prediction unit, encoding a syntax element, wherein the syntax element is used to indicate that the image block to be processed does not reuse affine information predicting the adjacent block, and stopping the execution of the subsequent stage; otherwise, continuing to execute the subsequent step and assigning an index value to each candidate prediction unit in the set of candidate prediction units. The candidate prediction unit represented by each index value is predefined in the protocol by the encoding and decoding parties and is kept consistent between the encoding and decoding parties. It traverses all the candidate prediction units in the set of candidate prediction units to obtain a predicted image of the image block to be processed and performs encoding for the image block to be processed. The encoding gains provided by the selected candidate prediction vectors are compared, and the candidate prediction unit with the largest gain is selected as the affine motion prediction unit referred to by the image block to be processed.

[0146] In this embodiment of the present invention, optionally, the second acquisition unit 22 is specifically configured to obtain position information and motion information of reference points in the motion prediction affine unit. The function of the module is the same as that of the second acquisition module 12, and the details are not described again.

[0147] In this embodiment of the present invention, optionally, the third acquisition unit 23 is specifically configured to acquire motion information of a pixel point in an image block in accordance with the position information and motion information of the above three selected reference points. The function of the module is the same as that of the third acquisition module 13, and the details are not described again.

[0148] In this embodiment of the present invention, optionally, the third acquisition unit 23 is specifically configured to perform motion compensation in accordance with the prediction direction, reference frame index and motion vector that are obtained to find the predicted value of a pixel point in the image block to be processed . The predicted values of all pixel points form a predicted image of the image to be processed. The function of the module is the same as that of the third acquisition module 13, and the details are not described again.

[0149] In this embodiment of the present invention, optionally, the encoding characteristics for encoding that is performed using a predicted image acquired by this device are compared with the encoding characteristics for encoding that is performed using a predicted image obtained using other devices, and the device , by which the largest encoding gain is obtained is selected as a device for encoding the predicted image.

[0150] In this embodiment of the present invention, optionally, if a predicted image obtained using this device is used, a syntax element is encoded, wherein the syntax element is used to indicate that a method for obtaining a predicted image using this device is used, and another is encoded a syntax element, wherein the syntax element is used to indicate index information of the selected prediction unit.

[0151] It should be understood that the encoding device 20 according to this embodiment of the present invention may accordingly execute the image prediction method 2000 in the embodiment of the present invention, and the above and other operations and/or functions of the modules of the encoding device 20 are separately used to implement the corresponding the procedure of the method according to Fig. 4. For brevity, details are not described here.

[0152] Therefore, the encoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0153] As shown in FIG. 7, an embodiment of the present invention further provides a decoding apparatus 30 including a processor 31, a memory 32, and a bus system 33. The processor 31 and the memory 32 are connected using a bus system 33, the memory 32 is configured to store instructions, and the processor 31 is configured to execute instructions stored in the memory 32. The memory 32 of the decoding device 30 stores program code, and the processor 31 can call the program code , stored in the memory 32 to perform the following operations: obtaining a first fiducial unit for an image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit using the same affine model; obtaining motion information of motion compensation base units at two or more predetermined positions in the first reference unit; and obtaining motion information of the motion compensation base unit of the image unit according to the motion information.

[0154] Therefore, the decoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0155] It should be understood that in this embodiment of the present invention, the processor 31 may be a Central Processing Unit (CPU), and the processor 31 may also be another general purpose processor, digital signal processor (DSP), application specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable device, discrete gate or transistor logic device, discrete hardware component, etc. The general purpose processor may be a microprocessor, or the processor may be any suitable processor, or the like.

[0156] Memory 32 may include read-only memory and random access memory and may provide instructions and data to processor 31. The memory portion 32 may further include non-volatile random access memory. For example, memory 32 may further store device type information.

[0157] The bus system 33, in addition to the data bus, may further include a power bus, a control bus, a status signaling bus, and the like. However, for clarity of description, the various types of buses in the drawings are all labeled as 33 bus system.

[0158] During implementation, each step of the above methods may be implemented by a hardware integrated circuit in the processor 31 or an instruction in the form of software. The steps of the methods disclosed with reference to embodiments of the present invention may be directly performed using a hardware processor or may be performed using a combination of hardware and software modules in the processor. The software module may be located in an off-the-shelf, practical storage medium such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or a register. The storage medium is located in the memory 32, and the processor 31 reads information from the memory 32 and performs the steps in the above methods in combination with the hardware of the processor 31. To avoid repetition, the details are not described again here.

[0159] Optionally, in an embodiment, processor 31 is specifically configured to: read the bitstream and parse the bitstream to obtain a syntax element, wherein the syntax element is used to determine whether the image block to be processed reuses motion information of the affine model adjacent block. When it is determined that the image block to be processed reuses the motion information of the affine model of the adjacent block, the process continues.

[0160] Optionally, in an embodiment, the processor 31 is specifically configured to determine an affine motion prediction unit that is used as a reference for the image block to be processed includes: determining a set of candidate prediction units for the image block to be processed , checking, in a predetermined order, whether the prediction unit in the set is an affine motion prediction unit, and if it is found that the prediction unit is an affine motion prediction unit, stopping the check and using the prediction unit as the affine motion prediction unit referenced to be processed image block; or determining a set of candidate prediction units for an image block to be processed, reading the bit stream, and parsing the bit stream to obtain a syntax element, wherein the syntax element is used to indicate index information of an affine motion prediction unit that is in the set of candidate prediction units and which is used as a reference for the image block to be processed, and determining, in accordance with the index information, an affine motion prediction unit that is used as a reference for the image block to be processed.

[0161] Optionally, in an embodiment, the processor 31 is specifically configured to obtain position information and motion information of the control points in the affine motion prediction unit includes: selecting any three control points from the four control points in the affine motion prediction unit and obtaining position information, motion vector information, prediction directions and reference frame indices of the three reference points, or selecting any two reference points, which may be referred to as the first and second reference points, from the four reference points in the affine prediction unit and obtaining the position information, vector information motion, prediction directions and reference frame indices of the two reference points; and then selecting any reference point which may be referred to as a third reference point from the remaining two reference points and obtaining position information of the third reference point. The motion vector information of the third reference point can be obtained in accordance with the first and second reference points as follows:

Where And are, respectively, the horizontal and vertical coordinates of the third control point, equals , And are the corresponding horizontal and vertical motion vectors, And And And are, respectively, the horizontal and vertical coordinates of the first and second control points, And And And are the corresponding horizontal and vertical motion vectors.

[0162] Since all pixel points in the same motion prediction affine unit have the same prediction direction and the same reference frame index, the prediction directions and reference frame indices of all reference points are also the same. The third reference point has the same prediction direction and reference frame index as those of the first and second reference points.

[0163] Optionally, in an embodiment, the processor 31 is specifically configured to obtain motion information of a pixel point in an image block according to the position information and motion information of the above three selected reference points, including: assigning a prediction direction and a reference index a frame of any one of the control points of a pixel point in the image block to be processed; or parsing the bit stream to obtain reference frame information of an image block to be processed, the reference frame information including a prediction direction and a reference frame index.

[0164] Optionally, in an embodiment, the processor 31 is specifically configured to obtain motion information of a pixel point in an image block in accordance with the position information and motion information of the above three selected reference points, including: obtaining, in accordance with the received position information and motion vectors of any three control points, the motion vector of a pixel point in the image block to be processed. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of any three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0165] Alternatively, the motion vectors of any two reference points in the image block to be processed are obtained in accordance with the acquired position information and the motion vectors of any three reference points. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0166] The motion vector of a pixel point in the image block to be processed is obtained in accordance with the obtained motion vectors of the two reference points. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0167] Alternatively, the motion vectors of any two reference points in the image block to be processed are obtained in accordance with the acquired position information and the motion vectors of any three reference points. This is implemented as follows:

,

,

Where And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, equals , And And And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the three control points in the affine motion prediction unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

[0168] The bit stream is parsed to obtain the remainder of the motion vector of the two reference points in the image block to be processed, and the remainder is added to the obtained motion vectors of the two reference points to obtain the updated motion vectors of the two reference points in the image block to be processed.

[0169] The motion vector of a pixel point in the image block to be processed is obtained in accordance with the obtained motion vectors of the two reference points. This is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of a pixel point in the image block to be processed, And are the corresponding horizontal and vertical motion vectors, and And And And are, respectively, the horizontal and vertical coordinates of two control points in the image block to be processed, where equals , And And And And are the corresponding horizontal and vertical motion vectors.

[0170] Optionally, in an embodiment, the processor 31 is specifically configured to: perform motion compensation in accordance with the prediction direction, reference frame index, and motion vector that are obtained, to find a pixel point in an image block to be processed. The predicted values of all pixel points form a predicted image of the image to be processed.

[0171] It should be understood that the decoding device 30 in accordance with this embodiment of the present invention may correspond to the decoding device 10 in the embodiment of the present invention and may correspond to the corresponding main entity that performs the method 1000 in the embodiment of the present invention. The above and other operations and/or functions of the modules of the decoding device 30 are separately used to implement the corresponding procedure of the method in FIG. 1. For brevity, the details are not described here.

[0172] Therefore, the decoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0173] As shown in FIG. 8, an embodiment of the present invention further provides an encoding device 40 including a processor 41, a memory 42, and a bus system 43. The processor 41 and the memory 42 are connected using a bus system 43, the memory 42 is configured to store instructions, and the processor 41 is configured to execute instructions stored in the memory 42. The memory 42 of the encoder 40 stores program code, and the processor 41 can call the program code , stored in the memory 42 to perform the following operations: obtaining a first fiducial unit for an image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit using the same affine model; obtaining motion information of motion compensation base units at two or more predetermined positions in the first reference unit; and obtaining motion information of the motion compensation base unit of the image unit according to the motion information.

[0174] Therefore, the encoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0175] It should be understood that in this embodiment of the present invention, the processor 41 may be a Central Processing Unit (CPU), and the processor 41 may also be another general purpose processor, digital signal processor (DSP), application specific integrated circuit (ASIC), field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general purpose processor may be a microprocessor, or the processor may be any suitable processor, or the like.

[0176] Memory 42 may include read-only memory and random access memory and may provide instructions and data to processor 41. The memory portion 42 may further include non-volatile random access memory. For example, memory 42 may further store device type information.

[0177] The bus system 43, in addition to the data bus, may also include a power bus, a control bus, a status signaling bus, and the like. However, for clarity of description, the various types of buses in the drawing are all labeled as bus system 43.

[0178] During implementation, each step of the above methods may be implemented by a hardware integrated circuit in the processor 41 or an instruction in the form of software. The steps of the methods disclosed with reference to embodiments of the present invention may be directly performed using a hardware processor or may be performed using a combination of hardware and software modules in the processor. The software module may be located in an off-the-shelf, practical storage medium such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or a register. The storage medium is located in the memory 42 and the processor 41 reads information from the memory 42 and performs the steps in the above methods in combination with the hardware of the processor 41. To avoid repetition, the details are not described again here.

[0179] Optionally, in this embodiment, the processor 41 is specifically configured to determine an affine motion prediction unit that is used as a reference for the image block to be processed includes: determining a set of candidate prediction units for the block to be processed image, checking, in a predetermined order, that the prediction unit in the set is an affine motion prediction unit, and if it is found that the prediction unit is an affine motion prediction unit, stopping the check and using the prediction unit as the affine motion prediction unit referred to by the subject processing the image block, or if none of the candidate prediction units is an affine motion prediction unit, encoding a syntax element, wherein the syntax element is used to indicate that the image block to be processed does not reuse the affine prediction information of the adjacent block, and stopping execution of the subsequent step ; otherwise, continue with the next step; or determining a set of candidate prediction units for the image block to be processed, if none of the candidate prediction units is an affine motion prediction unit, encoding a syntax element, wherein the syntax element is used to indicate that the image block to be processed does not reuse affine prediction information adjacent block, and stopping the execution of the subsequent stage; otherwise, continuing to execute the subsequent step and assigning an index value to each candidate prediction unit in the set of candidate prediction units. The candidate prediction unit represented by each index value is predefined in the protocol by the encoding and decoding parties and is kept consistent between the encoding and decoding parties. It traverses all the candidate prediction units in the set of candidate prediction units to obtain a predicted image of the image block to be processed and performs encoding for the image block to be processed. The encoding gains provided by the selected candidate prediction vectors are compared, and the candidate prediction unit with the largest gain is selected as the affine motion prediction unit referred to by the image block to be processed.

[0180] Optionally, in an embodiment, processor 41 is specifically configured to obtain position information and motion information of reference points in the motion prediction affine unit. The function of the module is the same as that of the processor 31, and the details are not described again.

[0181] Optionally, in an embodiment, the processor 41 is specifically configured to obtain motion information of a pixel point in an image block in accordance with the position information and motion information of the three selected reference points. The function of the module is the same as that of the processor 31, and the details are not described again.

[0182] Optionally, in an embodiment, the processor 41 is specifically configured to perform motion compensation in accordance with the prediction direction, reference frame index, and motion vector that are obtained to find a prediction value of a pixel point in an image block to be processed. The predicted values of all pixel points form a predicted image of the image to be processed. The function of the module is the same as that of the processor 31, and the details are not described again.

[0183] Optionally, in an embodiment, the processor 41 is specifically configured to: compare the characteristics of encoding that is performed using a predicted image obtained by this device with the characteristics of encoding that is performed using a predicted image obtained using another device, and select the device , by which the largest encoding gain is obtained, as a device for encoding the predicted image.

[0184] Optionally, in an embodiment, the processor 41 is specifically configured to: if a predicted image obtained using this device is used, encode a syntax element, wherein the syntax element is used to indicate that a method for obtaining a predicted image using this device is used, and encode another syntax element, the syntax element being used to indicate index information of the selected prediction unit.

[0185] It should be understood that the encoding device 40 in accordance with this embodiment of the present invention may correspond to the encoding device 20 in the embodiment of the present invention and may correspond to the corresponding main entity that performs the method 2000 in the embodiment of the present invention. The above and other operations and/or functions of the modules of the encoding device 40 are separately used to implement the corresponding procedure of the method in FIG. 4. For brevity, details are not described here.

[0186] Therefore, the encoding device in this embodiment of the present invention reuses the motion information of the first reference unit using the same affine motion prediction model, and therefore obtains a more accurate motion vector of the current image unit without encoding and decoding the motion vector difference and without additional search high-precision motion vector. In this way, the prediction accuracy is improved while maintaining the encoding and decoding complexity, thereby improving the encoding and decoding performance.

[0187] It should be understood that “embodiment” or “embodiment” referred to throughout the specification means that particular features, structures, or characteristics related to the embodiment are included in at least one embodiment of the present invention. Therefore, “in an embodiment” or “in an embodiment” appearing throughout the specification do not refer to the same embodiment. Moreover, these particular features, structures, or characteristics may be combined in one or more embodiments using any suitable method.

[0188] It should be understood that the sequence numbers of the above processes do not indicate the sequence of execution in the various embodiments of the present invention. Process execution sequences should be determined in accordance with the functions and internal logic of the processes and should not be construed as any limitation on the processes for implementing embodiments of the present invention.

[0189] Additionally, the terms "system" and "network" may be used interchangeably throughout this specification. It should be clear that the term "and/or" in this specification describes only an association relationship for describing associated objects and represents that three relationships can exist. For example, A and/or B can represent the following three cases: only A exists, both A and B exist, and only B exists. Additionally, the symbol ʺ/ʺ in this specification typically denotes an ʺorʺ relationship between associated objects.

[0190] It should be understood that in embodiments of this application, ʺB, corresponding to Aʺ indicates that B is associated with A, and B can be determined in accordance with A. However, it should also be understood that determining A in accordance with B does not mean that B is defined in accordance with A only; that is, B can also be determined according to A and/or other information.

[0191] One skilled in the art will appreciate that, in combination with the examples described in the embodiments disclosed in this specification, the modules and algorithms of the steps may be implemented by electronic hardware, computer software, or a combination thereof. In order to clearly describe the interchangeability between hardware and software, the compositions and steps of each example are described above according to functions. Whether functions are performed by hardware or software depends on the specific application and design limitations of the technical solutions. Various methods may be used by one skilled in the art to implement the described functions for each particular application, but such implementation should not be assumed to be beyond the scope of the present invention.

[0192] It will be understood by one skilled in the art that, for purposes of convenience and brevity of description, with respect to the detailed operating process of the above-mentioned system, apparatus and module, reference may be made to the corresponding process in the above-mentioned method embodiments, and the details will not be described again herein.

[0193] In some embodiments provided in this application, it will be appreciated that the disclosed system, apparatus, and method may be implemented in other ways. For example, the described embodiment of the device is merely an example. For example, modularization is simply a division into logical functions, and a different division may be used in an actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not performed. In addition, the mutual connections or direct connections or communication connections shown or described may be implemented using certain interfaces. Indirect connections or communication connections between devices or modules may be implemented in electronic, mechanical or other form.

[0194] Modules described as separate parts may or may not be physically separate, and parts depicted as modules may or may not be physical modules, may be located in a single location, or may be distributed across multiple network modules. Some or all of the modules may be selected according to actual requirements to achieve the objectives of the solutions in the embodiments.

[0195] In addition, the functional modules in embodiments of the present invention may be integrated into a single processing module, or each of the modules may exist separately physically, or two or more modules may be integrated into a single module.

[0196] When the integrated module is implemented in the form of a software function module and is sold or used as an independent product, the integrated module may be stored in a computer-readable storage medium. Based on such understanding, the technical solutions according to the present invention, substantially or in part contributing to the prior art, or some of the technical solutions can be implemented in the form of a software product. The software product is stored on a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to perform all or some of the steps of the methods described in embodiments of the present invention. The above-mentioned storage medium includes: any medium that can store program code, such as a USB flash drive, removable hard disk, Read-Only Memory (ROM), Random Access Memory RAM), magnetic disk or optical disk.

[0197] The above descriptions are merely specific implementations of the present invention, but are not intended to limit the scope of protection of the present invention. Any change or substitution easily accomplished by one skilled in the art within the technical scope disclosed in the present invention should be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should correspond to the scope of protection defined by the claims.

Clause 1. An image prediction method comprising:

obtaining a first fiducial unit for the image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit using the same affine model;

obtaining motion information of motion compensation base units at two or more predetermined positions in the first reference unit; And

obtaining motion information of the motion compensation base unit of the image unit according to the motion information.

Provision 2. The method according to Proposition 1, in which the acquisition of the first reference unit for the image unit contains:

checking, in a predetermined order, whether a predicted image for a prediction unit adjacent to the image unit is obtained using the affine model; And

if so, stop testing and use the prediction unit as the first reference unit.

Provision 3. The method according to Proposition 1, in which the preset position is the position of the corner point of the unit in the first reference unit.

Provision 4. The method according to any of Provisions 1-3, wherein obtaining motion information of basic motion compensation units at two or more preset positions in the first reference unit comprises:

obtaining motion information of motion compensation base units at three preset positions in the first reference unit; And

Accordingly, obtaining motion information of the motion compensation base unit of the image unit according to the motion information comprises:

obtaining a prediction direction, a reference frame index, and a motion vector of the motion compensation base unit of the image unit in accordance with the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at three preset positions in the first reference unit.

Provision 5. The method according to Proposition 4, wherein obtaining the prediction direction, reference frame index and motion vector of the motion compensation base unit of an image unit according to the prediction directions, reference frame indices and motion vectors of the motion compensation base units at three preset positions in the first reference unit contains:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit; And

obtaining a motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at three preset positions in the first reference unit.

Provision 6. The method according to Proposition 5, wherein obtaining the motion vector of the motion compensation base unit of an image unit by interpolation according to the motion vectors of the motion compensation base units at three preset positions in the first reference unit is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at three preset positions in the first reference unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

Provision 7. The method according to any of Provisions 1-3, wherein obtaining motion information of basic motion compensation units at two or more predetermined positions in the first reference unit comprises:

obtaining motion information of motion compensation base units at two predetermined positions in the first reference unit; And

Accordingly, obtaining motion information of the motion compensation base unit of the image unit according to the motion information comprises:

obtaining a prediction direction, a reference frame index, and a motion vector of the motion compensation base unit of the image unit in accordance with the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at two preset positions in the first reference unit.

Provision 8. The method according to Proposition 7, wherein obtaining the prediction direction, reference frame index and motion vector of the motion compensation base unit of an image unit according to the prediction directions, reference frame indices and motion vectors of the motion compensation base units at two preset positions in the first reference unit contains:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at two preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at two preset positions in the first reference unit; And

obtaining a motion vector of the motion compensation base unit of the image unit by interpolation in accordance with the motion vectors of the motion compensation base units at two preset positions in the first reference unit.

Provision 9. The method according to Proposition 8, wherein obtaining the motion vector of the motion compensation base unit of an image unit by interpolation according to the motion vectors of the motion compensation base units at two preset positions in the first reference unit is implemented as follows:

Where And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, And And And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at two preset positions in the first reference unit, and And And And are the corresponding horizontal and vertical motion vectors.

Clause 10. An image prediction device comprising:

a first acquisition module configured to obtain a first fiducial unit for an image unit, wherein corresponding predicted images are obtained for the image unit and the first fiducial unit using the same affine model;

a second acquisition module configured to receive motion information of motion compensation reference units at two or more preset positions in the first reference unit; And

a third acquiring module configured to obtain motion information of the base unit of motion compensation of the image unit in accordance with the position information and the motion information.

Statement 11. A device according to Statement 10, in which the first receiving module is configured to:

check, in a predetermined order, whether a predicted image is obtained for a prediction unit adjacent to the image unit using the affine model; And

if so, stop testing and use the prediction unit as the first reference unit.

Position 12. A device according to Position 10 in which the preset position is the position of the corner point of the unit in the first reference unit.

Statement 13. A device according to any of Statements 10-12, in which the third receiving module is configured to:

obtain motion information of motion compensation base units at three preset positions in the first reference unit; And

Accordingly, obtaining motion information of the motion compensation base unit of the image unit according to the motion information comprises:

obtaining a prediction direction, a reference frame index, and a motion vector of the motion compensation base unit of the image unit in accordance with the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at three preset positions in the first reference unit.

Statement 14. A device according to Statement 13, in which the third receiving module is further configured to:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at three preset positions in the first reference unit; And

obtain a motion vector of the image unit motion compensation base unit by interpolation according to the motion vectors of the motion compensation base units at three preset positions in the first reference unit.

Provision 15. The apparatus according to Proposition 14, wherein obtaining the motion vector of the motion compensation base unit of an image unit by interpolation in accordance with the motion vectors of the motion compensation base units at three preset positions in the first reference unit is implemented as follows:

where in And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, , And And , And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at three preset positions in the first reference unit, and , And And , And are the corresponding horizontal and vertical motion vectors.

Statement 16. A device according to any of Statements 10-12, in which the third receiving module is configured to:

obtain motion information of motion compensation base units at two preset positions in the first reference unit; And

Accordingly, obtaining motion information of the motion compensation base unit of the image unit according to the motion information comprises:

obtaining a prediction direction, a reference frame index, and a motion vector of the motion compensation base unit of the image unit in accordance with the prediction directions, reference frame indices, and motion vectors of the motion compensation base units at two preset positions in the first reference unit.

Statement 17. A device according to Statement 16, in which the third receiving module is further configured to:

assigning the same prediction direction to the motion compensation base unit of the image unit and the motion compensation base units at two preset positions in the first reference unit;

assigning the same reference frame index to the motion compensation base unit of the image unit and the motion compensation base units at two preset positions in the first reference unit; And

obtain a motion vector of the image unit motion compensation base unit by interpolation according to the motion vectors of the motion compensation base units at two preset positions in the first reference unit.

Provision 18. The apparatus according to Proposition 17, wherein obtaining the motion vector of the motion compensation base unit of an image unit by interpolation according to the motion vectors of the motion compensation base units at two preset positions in the first reference unit is implemented as follows:

where in And are, respectively, the horizontal and vertical coordinates of the basic unit of motion compensation of the image unit, And are the corresponding horizontal and vertical motion vectors, And And And are, respectively, the horizontal and vertical coordinates of the basic motion compensation units at two preset positions in the first reference unit, and And And And are the corresponding horizontal and vertical motion vectors.

Claims (23)

1. A method for intra-prediction of a current block of an image, for use in image encoding or image decoding, the method comprising the steps of: determining, by an image encoder or an image decoder, a potential candidate affine block, wherein the potential candidate affine block is an affine-coded block that is spatially adjacent to the current image block; obtaining motion vectors (MVs) of at least two control points of the current image block in accordance with the motion vectors of at least two control points of a potentially suitable affine block; obtaining a motion vector of each pixel point or each pixel matrix of the current image block in accordance with the motion vectors of at least two control points of the current image block; And Based on the motion vector of each pixel point or each pixel matrix of the current image block, predicted values of the current image block are obtained. 2. The method of claim 1, wherein said obtaining, based on the motion vector of each pixel point or each pixel matrix of the current image block, predicted values of the current image block comprises the steps of: obtaining predicted values of each pixel point or each pixel matrix of the current image block by motion compensation in accordance with the reference frame information of the current image block and the motion vector of each pixel point or each pixel matrix of the current image block; And obtaining predicted values of the current image block based on the predicted values of each pixel point or each pixel matrix of the current image block. 3. The method of claim 2, wherein the reference frame information of the current image block is the same as the reference frame information of the potentially suitable affine block. 4. The method according to any one of claims 1 to 3, wherein the MVs of said at least two reference points of the current image block correspond to a potentially suitable affine block or index included in the bitstream. 5. The method according to any one of claims 1 to 4, wherein said obtaining motion vectors (MV) of at least two control points of the current image block in accordance with the motion vectors of at least two control points of a potentially suitable affine block comprises the steps of: of which: obtaining motion vectors (MVs) of three control points of the current image block according to motion vectors of three control points of a potentially suitable affine block when the six-parameter affine model is used for the current image block; or obtain motion vectors (MVs) of two control points of the current image block according to the motion vectors of two control points of a potentially suitable affine block when the four-parameter affine model is used for the current image block. 6. The method according to any one of claims 1 to 4, wherein said obtaining motion vectors (MV) of at least two control points of the current image block in accordance with the motion vectors of at least two control points of a potentially suitable affine block comprises the step of: wherein motion vectors (MVs) of at least two control points of the current image block are obtained in accordance with the motion vectors of at least two control points of a potentially suitable affine block and at least one of the width and height of a potential suitable affine block. 7. The method of claim 6, wherein said at least one of the width and height of the candidate affine block is associated with coordinate positions of said at least two control points of the candidate affine block. 8. The method according to any one of claims 1 to 7, wherein the candidate affine block is a rectangular block, and the control points of the potential affine block comprise two or more of: a top left corner point, a top right corner point, a bottom left corner point point and the lower right corner point of a potentially matching affine block. 9. The method according to any one of claims 1 to 8, wherein the current image block is a rectangular block, and the control points of the current image block comprise two or more of: a top left corner point, a top right corner point, a bottom left corner point, and the bottom right corner point of the current image block. 10. The method according to any one of claims 1 to 9, wherein said obtaining the motion vector of each pixel point or each pixel matrix of the current image block in accordance with the motion vectors of at least two control points of the current image block comprises the step of obtaining the vector movement of each pixel point of the current image block in accordance with the movement vectors of two control points of the current image block: Where And represent, respectively, the horizontal and vertical coordinates of a pixel point in the current image block, And represent the corresponding horizontal and vertical motion vectors of a pixel point in an image block, And And And represent, respectively, the horizontal and vertical coordinates of two control points in the current image block, and And And And represent the corresponding horizontal and vertical motion vectors of two control points in the current image block. 11. An encoder characterized in that it contains processing circuits for performing the method according to any one of claims 1 to 10. 12. A decoder, characterized in that it contains processing circuits for performing the method according to any one of claims 1 to 10. 13. A non-transitory computer-readable storage medium on which computer instructions are stored which, when executed by one or more processors, cause one or more processors to carry out the method of any one of claims 1 to 10.

Images