Inter-prediction method, electronic device, storage medium and product

Abstract

Embodiments of the present application provide an inter-prediction method, an electronic device, a storage medium, and a product. On the basis of a first upper left illumination compensation model, a first upper illumination compensation model and a first left illumination compensation model, adding at least one of a second upper left illumination compensation model, a second upper illumination compensation model, and a second left illumination compensation model, so as to obtain a set of models comprising three or more illumination compensation models.

Description

Inter prediction method, electronic device, storage medium and product

[0001] Cross-reference to related applications

[0002] The present application is based on and claims priority to Chinese patent application No. 202411135159.4, filed on August 16, 2024, the entire contents of which are hereby incorporated by reference into the present application. TECHNICAL FIELD

[0003] Embodiments of the present application relate to the field of video coding technology, and in particular to an inter prediction method, an electronic device, a storage medium and a product. BACKGROUND

[0004] The inter prediction value correction technology is an encoding tool in Audio Video coding Standard 4 (AVS4) to compensate for the illumination difference between the encoded frame and the reference frame. Since there is an illumination difference in the time domain of the video image, the inter prediction value correction can compensate for this difference to improve the encoding efficiency.

[0005] In related technologies, the inter prediction value correction of AVS4 uses illumination compensation technology to eliminate the illumination angle change caused by illumination change and object movement, the shadow change caused by the relative movement of the front and rear objects, and the brightness change introduced by post-production. However, the existing illumination compensation model has only three models, and each model is set to have a relatively fixed position of the pixel point pair for deriving the illumination compensation model coefficients, and the encoding efficiency is low. How to improve the encoding and decoding efficiency is a problem that needs to be discussed and solved at present. SUMMARY

[0006] Embodiments of the present application provide an inter prediction method, an electronic device, a storage medium and a product.

[0007] In a first aspect, an embodiment of the present application provides an inter-frame prediction method applied to a decoding end, the method comprising: obtaining a code stream and parsing the code stream to obtain illumination compensation index, motion information and residual information of a current decoding unit; performing motion compensation on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit; correcting the prediction block of the current decoding unit according to target illumination compensation information corresponding to the illumination compensation index to obtain a corrected current decoding unit prediction block; and obtaining a reconstructed block of the current decoding unit according to the corrected current decoding unit prediction block and the residual information; wherein the target illumination compensation information is one of illumination compensation models; and the illumination compensation models comprise a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, and at least one of the following: a second upper-left illumination compensation model; a second upper illumination compensation model; and a second left illumination compensation model.

[0008] In a second aspect, an embodiment of the present application provides an inter-frame prediction method applied to an encoding end, the method comprising: performing motion compensation on a current encoding unit according to motion information to obtain a prediction block of the current encoding unit; correcting the prediction block of the current encoding unit based on rate-distortion optimization according to an illumination compensation model; determining target illumination compensation information according to a rate-distortion optimization cost; and writing an illumination compensation index corresponding to the target illumination compensation information into a code stream; wherein the target illumination compensation information is one of illumination compensation models; and the illumination compensation models comprise a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, and at least one of the following: a second upper-left illumination compensation model; a second upper illumination compensation model; and a second left illumination compensation model.

[0009] In a third aspect, an embodiment of the present application provides an inter-frame prediction method applied to a decoding end, the method comprising: obtaining a code stream and parsing the code stream to obtain illumination compensation index, motion information and residual information of a current decoding unit; performing motion compensation on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit; and obtaining a reconstructed block of the current decoding unit according to the prediction block of the current decoding unit and the residual information in a case where neither an upper neighboring block nor a left neighboring block of the current decoding unit is available.

[0010] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: at least one processor; at least one memory configured to store at least one program; and a frame inter prediction method according to the first aspect, the second aspect or the third aspect when the at least one program is executed by the at least one processor.

[0011] In a fifth aspect, an embodiment of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores computer executable instructions, and the computer executable instructions are configured to perform the inter prediction method according to the first aspect, the second aspect or the third aspect.

[0012] In a sixth aspect, an embodiment of the present application provides a program product, comprising a computer program or computer instructions, wherein the computer program or the computer instructions are stored in a computer readable storage medium, a processor of a computer device reads the computer program or the computer instructions from the computer readable storage medium, and the processor executes the computer program or the computer instructions, so that the computer device performs the inter prediction method according to the first aspect, the second aspect or the third aspect. BRIEF DESCRIPTION OF DRAWINGS

[0013] Fig. 1 is a schematic diagram of a video image system suitable for the inter prediction method according to an embodiment of the present application;

[0014] Fig. 2 is a flowchart of the inter prediction method applied to a decoding end according to an embodiment of the present application;

[0015] Figs. 3a to 3c are schematic diagrams of pixel pairs of a first upper-left illumination compensation model according to an example of the present application;

[0016] Figs. 4a to 4c are schematic diagrams of pixel pairs of a second upper-left illumination compensation model according to an example of the present application;

[0017] Fig. 5 is a schematic diagram of a current block and a neighboring block according to an example of the present application;

[0018] Fig. 6 is a schematic diagram of reconstructed pixels of a reference block according to an example of the present application;

[0019] Fig. 7 is a flowchart of the inter prediction method applied to a decoding end according to another embodiment of the present application;

[0020] Fig. 8 is a flowchart of the inter prediction method applied to an encoding end according to an embodiment of the present application;

[0021] Fig. 9 is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0023] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical sequence is shown in the flowchart, in some cases, the steps shown or described can be performed in a manner different from the module division in the device or the sequence in the flowchart. The terms "first", "second", and the like in the description and claims and the above drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0024] In the description of the embodiments of the present application, the words such as arrangement, installation, connection and the like should be understood broadly, and the person skilled in the art can determine the specific meaning of the above words in the embodiments of the present application in combination with the specific content of the technical solutions.

[0025] In the embodiments of the present application, the words "further", "exemplarily" or "optionally" are used to represent as an example, illustration or description, and should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. The use of the words "further", "exemplarily" or "optionally" aims to present the related concept in a specific manner.

[0026] In the related art, the inter prediction correction (IPC) technology is an encoding tool in the Audio Video coding Standard 4 (AVS4) to compensate for the illumination difference between the encoded frame and the reference frame. Since there is an illumination difference in the time domain of the video image, the inter prediction correction can compensate for this part of the difference to improve the encoding efficiency.

[0027] In the related art, the IPC of AVS4 adopts an illumination compensation technology to eliminate the illumination angle change caused by illumination change and object motion, the shadow change caused by the relative motion of the front and rear objects, and the brightness change introduced by post-production. However, the current inter prediction correction algorithm of AVS4 has the following problems:

[0028] (1) The existing illumination compensation model has only three models, and each model is set to have a relatively fixed position of the pixel point pair for deriving the illumination compensation model coefficient, and the encoding efficiency is low;

[0029] (2) The position of the four pixel point pairs corresponding to the illumination compensation model, which are set to derive the illumination compensation model coefficient, is relatively fixed, and the effect on a larger coding block is not good;

[0030] (3) The chroma component only applies IPC under certain conditions, which limits the improvement of the coding efficiency of the chroma component;

[0031] (4) When both the top and left neighboring blocks of the coding block are unavailable, the prediction pixel values of the coding block are directly corrected to a fixed value (for example, all the prediction pixel values of the coding block are corrected to 512 when the coding bit depth is 10) based on the coding bit depth, and the coding efficiency is low.

[0032] How to improve the coding efficiency is a problem to be discussed and solved at present.

[0033] Therefore, the embodiments of the present application provide an inter prediction method, an electronic device, a storage medium and a product. On the basis of the first up-left illumination compensation model, the first up illumination compensation model and the first left illumination compensation model, at least one of the following is newly added: a second up-left illumination compensation model, a second up illumination compensation model and a second left illumination compensation model, so as to obtain a model set containing three or more illumination compensation models.

[0034] In the process of video encoding, the motion compensation is first performed on the current coding unit according to the motion information to obtain a prediction block of the current coding unit. Then, the prediction block is corrected based on rate-distortion optimization according to each illumination compensation model in the model set to obtain a rate-distortion optimization cost corresponding to each illumination compensation model. Finally, the optimal illumination compensation model is determined from the illumination compensation models based on the rate-distortion optimization cost, that is, the target illumination compensation information is determined. The corresponding illumination compensation index is written into the code stream.

[0035] In the process of video decoding, the illumination compensation index, the motion information and the residual information of the current decoding unit are parsed from the obtained code stream. The motion compensation is performed on the current decoding unit through the motion information to obtain a prediction block of the current decoding unit. Then, the optimal illumination compensation model determined from the model set in the encoding stage can be determined through the target illumination compensation information corresponding to the illumination compensation index. The prediction block is corrected according to the illumination compensation model to obtain a corrected prediction block. Finally, the reconstructed block of the current decoding unit is obtained according to the corrected prediction block and the residual information, and the decoding is completed.

[0036] It can be seen that the embodiments of the present application expand the types of illumination compensation models by newly adding at least one illumination compensation model. In the process of coding and decoding, the possibility of selecting / choosing the illumination compensation model is increased. Thus, the illumination compensation model more suitable for the current coding unit or the current decoding unit is obtained, the effect of correcting the prediction block by using the illumination compensation model is improved, and finally the coding and decoding efficiency is improved.

[0037] The embodiments of the present application also re-determine the pixel point pair of the second up-left illumination compensation model according to the initial pixel point pair of different positions corresponding to the second up-left illumination compensation model and the pixel point pair corresponding to the first up-left illumination compensation model.

[0038] The embodiment of the present application further re-determines the pixel point pair of the second left illumination compensation model according to the pixel point pair corresponding to the first left illumination compensation model and the initial pixel point pair at different positions corresponding to the second left illumination compensation model.

[0039] The embodiment of the present application further re-determines the pixel point pair of the second left illumination compensation model according to the pixel point pair corresponding to the first left illumination compensation model and the initial pixel point pair at different positions corresponding to the second left illumination compensation model.

[0040] The improvement is arranged to derive the selection of the pixel point pair of the illumination compensation model coefficient, and the position of the pixel point pair is more flexible, and the derived illumination compensation model coefficient is more adaptive to the current encoding / decoding block, thereby improving the encoding and decoding efficiency.

[0041] The embodiment of the present application further removes the restriction of applying IPC to the chroma component, and does not only apply IPC to the chroma component under certain conditions. Instead, the chroma component is directly subjected to similar encoding and decoding processing through the inter-prediction method provided in the present application, thereby further improving the encoding and decoding efficiency of the chroma component.

[0042] The technical solution of the embodiment of the present application can be applied to various video coding standards that allow inter-prediction coding. For example, High Efficiency Video Coding (HEVC), Audio Video coding Standard (AVS), such as AVS4, Versatile Video Coding (VVC), such as H.266 / VVC standard, or next-generation video coding standard, and the present application is not limited thereto.

[0043] The embodiments of the present application will be further described below with reference to the accompanying drawings.

[0044] In embodiments of the present application, the video image device, the encoding device or the decoding device can be various types of devices that can perform inter prediction encoding technology, such as digital television, digital camera, digital broadcast system, wireless communication device, personal digital assistant (PDA), laptop computer, desktop computer, set-top box, display device, digital media player, video game console, digital recording device, mobile phone, terminal side device. The terminal side device can also be referred to as terminal, user equipment (UE), mobile station (MS), mobile terminal (MT), etc. The terminal side device can also be a car, smart car, wearable device, tablet (Pad), virtual reality (VR) terminal device, augmented reality (AR) terminal device, video image display or transmission device in smart home, etc. with video image coding function. Embodiments of the present application do not limit the specific technology and specific device form used by the encoding device or the decoding device.

[0045] In order to facilitate understanding of the technical solutions of the present application, the inter prediction method, electronic storage medium and computer program product are further described below in one application scenario of the technical solutions of the present application. FIG. 1 is a schematic diagram of a video image system suitable for an inter prediction method according to an embodiment of the present application. As shown in FIG. 1, the video image system 1000 includes a plurality of video image devices or coding devices that can transmit video image information to each other via, for example, a network 1100. For example, the video image system 1000 includes a first terminal device 1200 with video encoding function and a second terminal device 1300 with video decoding function interconnected via the network 1100.

[0046] The first terminal device 1200 can encode video image data to be transmitted to the second terminal device 1300 via the network 1100. It can be understood that the network 1100 can be a limited network, a wireless network, etc. Various network forms capable of transmitting video image data. The encoded video image data or video image information can be transmitted in the form of one or more encoded video bitstreams through the network 1100. The second terminal device 1300 can receive the encoded video image data from the network 1100, decode the encoded video image data to recover the video image, and display the video or graphics according to the recovered video image data.

[0047] In some application scenarios, the transmission of the video code stream can be bidirectional. That is, the first terminal device 1200 can also receive the encoded video image data, and decode the encoded video image data to recover the video image. The second terminal device 1300 can also encode the video image data to be transmitted to the first terminal device 1200 via the network 1100.

[0048] In some application scenarios, the first terminal device 1200 and the second terminal device 1300 can transmit the encoded video image data to a server. The server decodes the video image data and performs related data processing, encodes the processed video image data, and transmits the processed video image data to the first terminal device 1200 and the second terminal device 1300 or other terminal devices.

[0049] The embodiments of the present application can be applied to various application scenarios, such as video conferencing, video monitoring, video live broadcast, and other scenarios involving video coding. In the related art, since the existing illumination compensation model has only three models, and each model is set to have a relatively fixed position of the pixel point pair for deriving the illumination compensation model coefficient, the coding efficiency is low. Some embodiments of the present application provide an inter prediction method, an electronic device, a storage medium, and a product, which can improve the coding efficiency.

[0050] The inter prediction method is a method for correcting the inter prediction value of the prediction block of the current coding / decoding unit by using the illumination compensation model in the process of video coding and video decoding.

[0051] The inter prediction method of the embodiments of the present application can be executed on a server or a terminal.

[0052] FIG. 2 is a flowchart of an inter prediction method applied to a decoding end according to an embodiment of the present application. The inter prediction method is applied to a decoding end. As shown in FIG. 2, the inter prediction method includes at least but is not limited to steps 210 to 250.

[0053] In step 210, a code stream is obtained and parsed to obtain an illumination compensation index, motion information, and residual information of a current decoding unit.

[0054] In step 220, motion compensation is performed on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit.

[0055] Step 230, correcting the prediction block of the current decoding unit according to the target illumination compensation information corresponding to the illumination compensation index, to obtain a corrected prediction block of the current decoding unit; wherein the target illumination compensation information is one of the illumination compensation models; and the illumination compensation models include the first upper-left illumination compensation model, the first upper illumination compensation model, the first left illumination compensation model, and at least one of the following: the second upper-left illumination compensation model; the second upper illumination compensation model; and the second left illumination compensation model.

[0056] Step 240, obtaining a reconstructed block of the current decoding unit according to the corrected prediction block of the current decoding unit and the residual information.

[0057] In step 210, the bitstream refers to a data stream set for transmission after the video data is encoded by the video encoding end. The current decoding unit refers to a data unit currently decoded obtained by parsing the bitstream, which corresponds to the coding unit (CU) in the encoding stage. It should be noted that in the embodiments of the present application, the current coding unit and the current decoding unit can also be referred to as the current block.

[0058] The illumination compensation index refers to an index set to indicate the illumination compensation model determined in the encoding stage for the current decoding unit. The motion information, i.e., the motion vector (MV) information, refers to the motion vector (displacement of the reference unit to the current decoding unit) between the current coding unit and the corresponding reference unit. The reference unit refers to a unit best matched with the current coding unit determined from the coded image based on motion estimation, which can also be referred to as a reference block. The residual information refers to the pixel difference between the current coding unit and the prediction block obtained after motion compensation in the encoding stage.

[0059] In step 220, the prediction block of the current decoding unit refers to a data block obtained by predicting the current decoding unit using the motion information and the reference unit.

[0060] In step 230, the target illumination compensation information refers to the illumination compensation model corresponding to the illumination compensation index. The correction of the prediction block refers to the calculation of the compensation coefficient by the illumination compensation model, and the correction of each pixel value in the prediction block by using the compensation coefficient.

[0061] In the decoding stage, the illumination compensation model set to correct the prediction block of the current decoding unit is determined from the set of illumination compensation models by the encoding end in the encoding stage.

[0062] According to some embodiments, the set of illumination compensation models can exist in the following cases:

[0063] The first upper-left illumination compensation model, the first upper illumination compensation model, the first left illumination compensation model, the second upper-left illumination compensation model.

[0064] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper illumination compensation model.

[0065] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second left illumination compensation model.

[0066] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper-left illumination compensation model, a second upper illumination compensation model.

[0067] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper-left illumination compensation model, a second left illumination compensation model.

[0068] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper illumination compensation model, a second left illumination compensation model.

[0069] Or, the set of illumination compensation models includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper-left illumination compensation model, a second upper illumination compensation model, a second left illumination compensation model.

[0070] In step 240, the reconstructed block refers to a data block obtained by reconstructing the corrected prediction block after superimposing the residual information on the corrected prediction block.

[0071] The steps 210 to 240 described above analyze the illumination compensation index, the motion information and the residual information of the current decoding unit from the obtained code stream. The motion compensation is performed on the current decoding unit through the motion information to obtain the prediction block of the current decoding unit. The optimal illumination compensation model determined from the model set in the encoding stage can be determined through the target illumination compensation information corresponding to the illumination compensation index. The prediction block is corrected according to the illumination compensation model to obtain the corrected prediction block. Finally, the reconstructed block of the current decoding unit is obtained according to the corrected prediction block and the residual information, and the decoding is completed. By adding at least one illumination compensation model, the types of illumination compensation models are expanded. In the process of coding and decoding, the possibility of selecting / choosing the illumination compensation model is increased. Thus, the illumination compensation model more suitable for the current coding unit or the current decoding unit is obtained, the effect of correcting the prediction block by using the illumination compensation model is improved, and finally the coding and decoding efficiency is improved.

[0072] The above is a general description of steps 210 to 240, and the specific implementation process of each step is described in detail below.

[0073] In an embodiment, the first upper-left illumination compensation model and the second upper-left illumination compensation model are different; the first upper illumination compensation model and the second upper illumination compensation model are different; and the first left illumination compensation model and the second left illumination compensation model are different. Wherein, the illumination compensation model is obtained by using the reconstructed pixels of the neighboring units of the current coding unit and the reconstructed pixels of the reference units of the current coding unit to obtain four pixel pairs, and then using the four pixel pairs to calculate the scaling compensation coefficient and the offset compensation coefficient corresponding to the illumination compensation model. After the scaling compensation coefficient and the offset compensation coefficient are determined, the prediction block can be corrected based on the scaling compensation coefficient and the offset compensation coefficient. In this embodiment, different second upper-left illumination compensation models, different second upper illumination compensation models, and different second left illumination compensation models are added based on the first upper-left illumination compensation model, the first upper illumination compensation model, and the first left illumination compensation model. The selectability of the illumination compensation model is increased, and in the coding process, the illumination compensation model that is more suitable for the current coding / decoding unit can be selected to correct the prediction value, thereby improving the coding efficiency.

[0074] In an embodiment, the illumination compensation model at least meets one of the following conditions:

[0075] The pixel pair corresponding to the first upper-left illumination compensation model is different from the pixel pair corresponding to the second upper-left illumination compensation model;

[0076] The pixel pair corresponding to the first upper illumination compensation model is different from the pixel pair corresponding to the second upper illumination compensation model;

[0077] The pixel pair corresponding to the first left illumination compensation model is different from the pixel pair corresponding to the second left illumination compensation model;

[0078] Wherein, the pixel pair is obtained according to the reconstructed pixels of the current decoding unit and the reconstructed pixels of the reference units matched with the current decoding unit.

[0079] In this embodiment, the reconstructed pixels of the current decoding unit refer to the pixels of the decoded neighboring units corresponding to the current decoding unit. When selecting the pixel pairs, the pixels in the edge row (or column) of the neighboring unit close to the current decoding unit are selected; for example, when the pixels are taken from the upper neighboring unit of the current decoding unit, the pixels in the lower edge row of the neighboring unit are selected. The neighboring unit refers to the decoded unit adjacent to the upper or left side of the current decoding unit, which can also be referred to as a neighboring block. The reconstructed pixels of the reference unit refer to the pixels in the reference unit.

[0080] The pixel pairs are different in that the positions of the pixel pairs of the first upper-left (upper / left) illumination compensation model are different from the positions of the pixel pairs of the second upper-left (upper / left) illumination compensation model.

[0081] According to some embodiments, the first upper-left illumination compensation model and the second upper-left illumination compensation model are exemplified as follows:

[0082] FIGS. 3a-3c are schematic diagrams of pixel pairs of the first upper-left illumination compensation model according to an example of the present application. As shown in FIG. 3a, in the case where the neighboring blocks on the upper side and the left side of the current block are both available, the pixel pairs corresponding to the first upper-left illumination compensation model are selected from the upper side, in some embodiments, the 2nd and the last reconstructed pixels are selected from the pixels in the upper edge row of the reference block, and the 2nd and the last reconstructed pixels are selected from the pixels in the lower edge row of the neighboring block on the upper side of the current block (as the black dots in FIG. 3a), to form two pixel pairs. The other two pixel pairs are selected from the left side, in some embodiments, the 2nd and the last reconstructed pixels are selected from the pixels in the left edge column of the reference block, and the 2nd and the last reconstructed pixels are selected from the pixels in the right edge column of the neighboring block on the left side of the current block (as the black dots in FIG. 3a), to form two pixel pairs.

[0083] As shown in FIG. 3b, in the case where only the neighboring block on the upper side of the current block is available, the pixel pairs corresponding to the first upper-left illumination compensation model are all selected from the upper side, in some embodiments, four reconstructed pixels are selected from the pixels in the upper edge row of the reference block, and four reconstructed pixels are selected from the pixels in the lower edge row of the neighboring block on the upper side of the current block, to form four pixel pairs. The positions of the pixel selection are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the width (as the black dots in FIG. 3b).

[0084] As shown in FIG. 3c, in the case where only the neighboring block on the left side of the current block is available, the pixel pairs corresponding to the first upper-left illumination compensation model are all selected from the left side, in some embodiments, four reconstructed pixels are selected from the pixels in the left edge column of the reference block, and four reconstructed pixels are selected from the pixels in the right edge column of the neighboring block on the left side of the current block, to form four pixel pairs. The positions of the pixel selection are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the height (as the black dots in FIG. 3c).

[0085] FIGS. 4a-4c are schematic diagrams of pixel pairs of the second upper-left illumination compensation model according to an example of the present application. As shown in FIG. 4a, in the case where the neighboring blocks on the upper side and the left side of the current block are both available, the pixel pairs corresponding to the second upper-left illumination compensation model are selected from the upper side, in some embodiments, two reconstructed pixels are selected from the pixels in the upper edge row of the reference block, and two reconstructed pixels are selected from the pixels in the lower edge row of the neighboring block on the upper side of the current block, to form two pixel pairs. The positions of the pixel selection are: 3 / 8, 5 / 8 of the width (as the black dots in FIG. 4a).

[0086] The other two pixel pairs are selected from the left side. In some embodiments, two reconstructed pixels are selected from the pixels in the left edge row in the reference block, and two reconstructed pixels are selected from the pixels in the right edge column in the neighboring block on the left side of the current block (as the black dots in FIG. 4a), to form two pixel pairs. The positions of the selected pixels are 3 / 8 and 5 / 8 of the height (as the black dots in FIG. 4a).

[0087] As shown in FIG. 4b, in the case where only the neighboring block on the top side of the current block is available, the pixel pairs corresponding to the second top-left illumination compensation model are all selected from the top side. In some embodiments, four reconstructed pixels are selected from the pixels in the top edge row in the reference block, and four reconstructed pixels are selected from the pixels in the bottom edge row in the neighboring block on the top side of the current block, to form four pixel pairs. The positions of the selected pixels are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, and 3 / 4+1 / 8 of the width (as the black dots in FIG. 4b).

[0088] As shown in FIG. 4c, in the case where only the neighboring block on the left side of the current block is available, the pixel pairs corresponding to the second top-left illumination compensation model are all selected from the left side. In some embodiments, four reconstructed pixels are selected from the pixels in the left edge column in the reference block, and four reconstructed pixels are selected from the pixels in the right edge column in the neighboring block on the left side of the current block, to form four pixel pairs. The positions of the selected pixels are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, and 3 / 4+1 / 8 of the height (as the black dots in FIG. 4c).

[0089] It should be noted that the positions of the pixel pairs of the first top-left illumination compensation model and the second top-left illumination compensation model are not limited to the examples provided above. The positions of the pixel pairs can also be any positions in the corresponding selection regions, and the positions of the pixel pairs of the first top-left illumination compensation model and the second top-left illumination compensation model are different, which will not be described herein. The same applies to the first top illumination compensation model and the second top illumination compensation model, and the first left illumination compensation model and the second left illumination compensation model.

[0090] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0091] The reconstructed pixels of the current decoding unit corresponding to the first top-left illumination compensation model are the same as the reconstructed pixels of the current decoding unit corresponding to the second top-left illumination compensation model, and the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the first top-left illumination compensation model are the same as the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the second top-left illumination compensation model.

[0092] The reconstructed pixels of the current decoding unit corresponding to the first up-light compensation model are the same as the reconstructed pixels of the current decoding unit corresponding to the second up-light compensation model, and the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the first up-light compensation model are the same as the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the second up-light compensation model.

[0093] The reconstructed pixels of the current decoding unit corresponding to the first left-light compensation model are the same as the reconstructed pixels of the current decoding unit corresponding to the second left-light compensation model, and the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the first left-light compensation model are the same as the reconstructed pixels of the reference unit matched by the current decoding unit corresponding to the second left-light compensation model.

[0094] In the present embodiment, the reconstructed pixels of the current decoding unit refer to the pixels in the corresponding regions in the neighboring units above and / or left of the current decoding unit. According to some embodiments, FIG. 5 is a schematic diagram of a current block and neighboring blocks provided by an example of the present application. As shown in FIG. 5, in the up-left light compensation model, the reconstructed pixels of the current decoding unit (current block) include all the pixels in a row of the lower edge adjacent to the current block in the illustrated upper neighboring block (as shown by the circles in the upper neighboring block in the figure), and all the pixels in a column of the right edge adjacent to the current block in the illustrated left neighboring block (as shown by the circles in the left neighboring block in the figure).

[0095] In the up-light compensation model, the reconstructed pixels of the current decoding unit (current block) include all the pixels in a row of the lower edge adjacent to the current block in the illustrated upper neighboring block (as shown by the circles in the upper neighboring block in the figure).

[0096] In the left-light compensation model, the reconstructed pixels of the current decoding unit (current block) include all the pixels in a column of the right edge adjacent to the current block in the illustrated left neighboring block (as shown by the circles in the left neighboring block in the figure).

[0097] The reconstructed pixels of the reference unit refer to the pixels in the corresponding regions in the upper edge region and / or the left edge region in the reference unit. According to some embodiments, FIG. 6 is a schematic diagram of the reconstructed pixels of a reference block provided by an example of the present application. As shown in FIG. 6, in the up-left light compensation model, the reconstructed pixels of the reference unit (reference block) include all the pixels in a row of the upper edge in the reference block (as shown by the circles selected by the reconstructed pixel 1 frame in the figure), and all the pixels in a column of the left edge in the reference block (as shown by the circles selected by the reconstructed pixel 2 frame in the figure).

[0098] In the up-light compensation model, the reconstructed pixels of the reference unit (reference block) include all the pixels in a row of the upper edge in the reference block (as shown by the circles selected by the reconstructed pixel 1 frame in the figure).

[0099] In the left illumination compensation model, the reconstructed pixels of the reference unit (reference block) include all pixels of the left edge column in the reference block (as shown by the circle in which the reconstructed pixel 2 in the figure is framed).

[0100] In an embodiment, the illumination compensation model at least meets one of the following:

[0101] The pixel pairs corresponding to the illumination compensation model are obtained according to the width of the current decoding unit or the width of the reference unit matched by the current decoding unit.

[0102] The pixel pairs corresponding to the illumination compensation model are obtained according to the height of the current decoding unit or the height of the reference unit matched by the current decoding unit.

[0103] In the present embodiment, since the width and height of the current decoding unit are the same as the width and height of the reference unit matched by the current decoding unit, whether the positions of the pixel pairs are determined based on the width and height of the current decoding unit or based on the width and height of the reference unit matched by the current decoding unit are the same. Subsequently, for the convenience of description, the width of the current decoding unit and the height of the current decoding unit are used as examples.

[0104] According to some embodiments, for the upper left illumination compensation model, two pairs of pixel pairs are determined based on the width of the current decoding unit and a pre-set ratio or fraction, and another two pairs of column pixel pairs are determined based on the height of the current decoding unit and a pre-set ratio or fraction.

[0105] Taking Fig. 4a as an example, it is assumed that Fig. 4a is a schematic of pixel pairs of the second upper left illumination compensation model. It can be determined from Fig. 4a that the width of the current decoding unit (the current block in the figure) is eight pixel points, and from left to right, the eight pixel points correspond to positions 0 to 7 respectively. The height of the current decoding unit (the current block in the figure) is also eight pixel points, and from top to bottom, the eight pixel points correspond to positions 0 to 7 respectively.

[0106] It is assumed that the pre-set ratios are 3 / 8 and 5 / 8. At this time, the pixel points at positions 3 and 5 of the pixel points in the row above the current block in the figure (indicated by the black circles in the figure) are taken, that is, the pixel points at positions 3 and 5 of the pixel points in the row above the reference block in the figure (indicated by the black circles in the figure) are taken, to obtain two pairs of pixel pairs.

[0107] The pixel points at the positions of 3 / 8 and 5 / 8 of the height of the left side of the current block are taken, that is, the pixel points at the positions of 3 and 5 in the left column of the current block in the figure (indicated by black circles in the figure); the pixel points at the positions of 3 / 8 and 5 / 8 of the height of the left side of the reference block are taken, that is, the pixel points at the positions of 3 and 5 in the left column of the reference block in the figure (indicated by black circles in the figure), to obtain two pairs of pixel point pairs. Finally, four pairs of pixel point pairs corresponding to the second upper left illumination compensation model are obtained.

[0108] According to some embodiments, for the upper illumination compensation model, taking 4b as an example, it is assumed that FIG. 4b is a schematic of pixel point pairs of the second upper illumination compensation model. It can be determined from FIG. 4b that the width of the current decoding unit (the current block in the figure) is eight pixel points, and from left to right, the eight pixel points correspond to positions 0 to 7.

[0109] It is assumed that the preset ratios are 1 / 8, 3 / 8, 5 / 8 and 7 / 8. At this time, the pixel points at the positions of 1 / 8, 3 / 8, 5 / 8 and 7 / 8 of the width of the upper side of the current block are taken, that is, the pixel points at the positions of 1, 3, 5 and 7 in the upper row of the current block in the figure (indicated by black circles in the figure); the pixel points at the positions of 1 / 8, 3 / 8, 5 / 8 and 7 / 8 of the width of the upper side of the reference block are taken, that is, the pixel points at the positions of 1, 3, 5 and 7 in the upper row of the reference block in the figure (indicated by black circles in the figure), to obtain four pairs of pixel point pairs corresponding to the second upper illumination compensation model.

[0110] According to some embodiments, for the left illumination compensation model, taking 4c as an example, it is assumed that FIG. 4c is a schematic of pixel point pairs of the second left illumination compensation model. It can be determined from FIG. 4c that the height of the current decoding unit (the current block in the figure) is eight pixel points, and from top to bottom, the eight pixel points correspond to positions 0 to 7.

[0111] It is assumed that the preset ratios are 1 / 8, 3 / 8, 5 / 8 and 7 / 8. At this time, the pixel points at the positions of 1 / 8, 3 / 8, 5 / 8 and 7 / 8 of the height of the left side of the current block are taken, that is, the pixel points at the positions of 1, 3, 5 and 7 in the left column of the current block in the figure (indicated by black circles in the figure); the pixel points at the positions of 1 / 8, 3 / 8, 5 / 8 and 7 / 8 of the height of the left side of the reference block are taken, that is, the pixel points at the positions of 1, 3, 5 and 7 in the left column of the reference block in the figure (indicated by black circles in the figure), to obtain four pairs of pixel point pairs corresponding to the second left illumination compensation model.

[0112] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0113] The pixel point pairs corresponding to the second upper left illumination compensation model are obtained according to the position information of the pixel point pairs corresponding to the first upper left illumination compensation model.

[0114] The pixel pair corresponding to the second upper light compensation model is obtained according to the position information of the pixel pair corresponding to the first upper light compensation model.

[0115] The pixel pair corresponding to the second left light compensation model is obtained according to the position information of the pixel pair corresponding to the first left light compensation model.

[0116] In the embodiment, the position information of the pixel pair is height information or width information of the position of the pixel. Since the pixel pair is composed of the pixel of the neighboring block of the current block and the pixel corresponding to the position of the reference block matched by the current block, the height information or the width information corresponding to the two pixels in one pixel pair is the same.

[0117] In an embodiment, the pixel pair corresponding to the second upper left light compensation model includes at least one of the following: the row pixel pair corresponding to the second upper left light compensation model; and the column pixel pair corresponding to the second upper left light compensation model.

[0118] The pixel pair corresponding to the first upper left light compensation model includes at least one of the following: the row pixel pair corresponding to the first upper left light compensation model; and the column pixel pair corresponding to the first upper left light compensation model.

[0119] In some embodiments, when the pixel pair is a row pixel pair, the position information of the pixel pair is width information. When the pixel pair is a column pixel pair, the position information of the pixel pair is height information. According to some embodiments, the height information can be represented by a ratio of the height corresponding to the position of the pixel to the total height of the column. According to some embodiments, the width information can be represented by a ratio of the width corresponding to the position of the pixel to the total width of the row.

[0120] In the case where the neighboring units above and left of the current decoding unit are both available, the pixel pair corresponding to the second upper left light compensation model includes the row pixel pair and the column pixel pair. In the case where the neighboring unit above the current decoding unit is unavailable, the pixel pair corresponding to the second upper left light compensation model only includes the column pixel pair. In the case where the neighboring unit left of the current decoding unit is unavailable, the pixel pair corresponding to the second upper left light compensation model only includes the row pixel pair. The same applies to the first upper left light compensation unit, which is not described herein.

[0121] In an embodiment, the row pixel pair corresponding to the second upper left light compensation model is obtained by linear operation according to the position information of the row pixel pair corresponding to the first upper left light compensation model.

[0122] Or the column pixel pair corresponding to the second upper left light compensation model is obtained by linear operation according to the position information of the column pixel pair corresponding to the first upper left light compensation model.

[0123] In some embodiments, the position information of the row pixel point pair corresponding to the second upper left light compensation model is obtained by linear operation on the position information of the row pixel point pair corresponding to the first upper left light compensation model, and the specific row pixel point pair corresponding to the second upper left light compensation model is determined based on the position information. The column pixel point pair corresponding to the second upper left light compensation model is determined in the same way.

[0124] According to some embodiments, the linear operation formula is y=x+1 / 8, where y is the position information of the pixel point pair corresponding to the second upper left light compensation model, x is the position information of the pixel point pair corresponding to the first upper left light compensation model, and 1 / 8 is a preset coefficient, where the coefficient can be set based on requirements.

[0125] Taking the row pixel point as an example, it is assumed that the selected positions of the row pixel points corresponding to the first upper left light compensation model are 0 / 4, 1 / 4, 2 / 4 and 3 / 4 of the width, as shown in FIG. 3b. Based on the linear operation formula y=x+1 / 8, the selected positions of the row pixel points corresponding to the second upper left light compensation model are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8 and 3 / 4+1 / 8 of the width, as shown in FIG. 4b.

[0126] Taking the column pixel point as an example, it is assumed that the selected positions of the column pixel points corresponding to the first upper left light compensation model are 0 / 4, 1 / 4, 2 / 4 and 3 / 4 of the height, as shown in FIG. 3c. Based on the linear operation formula y=x+1 / 8, the selected positions of the column pixel points corresponding to the second upper left light compensation model are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8 and 3 / 4+1 / 8 of the height, as shown in FIG. 4c.

[0127] It should be noted that, in addition to the example y=x+1 / 8 described above, the linear operation formula can also be set as y=ax, y=ax+b, etc. according to requirements, where y is the position information of the pixel point pair corresponding to the second upper left light compensation model, x is the position information of the pixel point pair corresponding to the first upper left light compensation model, and a and b are set coefficients, which will not be described here.

[0128] In an embodiment, the row pixel point pair corresponding to the second upper left light compensation model is obtained by weighted operation on the row pixel point pair corresponding to the first upper left light compensation model and the initial row pixel point pair of the second upper left light compensation model;

[0129] or the column pixel point pair corresponding to the second upper left light compensation model is obtained by weighted operation on the column pixel point pair corresponding to the first upper left light compensation model and the initial column pixel point pair of the second upper left light compensation model;

[0130] The second upper-left illumination compensation model initial row pixel pair is obtained by linear operation according to position information of the first upper-left illumination compensation model corresponding row pixel pair.

[0131] The second upper-left illumination compensation model initial column pixel pair is obtained by linear operation according to position information of the first upper-left illumination compensation model corresponding column pixel pair.

[0132] In the embodiment, the pixel pair includes a reconstructed pixel of a neighboring block of the current block and a reconstructed pixel of a reference block corresponding to the illumination compensation model. Assuming that the pixel pair of the first upper-left illumination compensation model is column pixel pair 1, and the pixel pair corresponding to the second upper-left illumination compensation model is column pixel pair 2. The weighted operation between the column pixel pair 1 and the column pixel pair 2 is specifically: the weighted operation is performed on the reconstructed pixel of the neighboring block of the current block in the column pixel pair 1 and the reconstructed pixel of the neighboring block of the current block in the column pixel pair 2, and the weighted operation is performed on the reconstructed pixel of the reference block in the column pixel pair 1 and the reconstructed pixel of the reference block in the column pixel pair 2, to obtain two new pixels, which form a new pixel pair. The weights of the pixel pair of the first upper-left illumination compensation model and the pixel pair of the second upper-left illumination compensation model are both 1. It should be noted that the weight in the weighted operation process can be set arbitrarily according to requirements, and is not specifically limited here. It should be noted that the weighted average calculation of the column pixel pair and the weighted operation of the row pixel pair are both corresponding to the above process, and are not described separately here.

[0133] According to some embodiments, the weighted operation can be weighted average operation, or can be weighted operation.

[0134] According to some embodiments, the determination process of the row pixel pair and the column pixel pair corresponding to the second upper-left illumination compensation model is described taking FIG. 3a and FIG. 4a as examples.

[0135] In the example, it is assumed that FIG. 3a is a schematic of the pixel pair of the first upper-left illumination compensation model in the case that the neighboring blocks on the upper side and the left side of the current block are both available, and FIG. 4a is a schematic of the initial pixel pair of the second upper-left illumination compensation model in the case that the neighboring blocks on the upper side and the left side of the current block are both available. Correspondingly, the second upper-left illumination compensation model corresponds to two row pixel pairs and two column pixel pairs.

[0136] In FIG. 3a and FIG. 4a, for the positions of the 8 row pixels in the figure, the positions are 0 to 7 from left to right, and for the positions of the 8 column pixels in the figure, the positions are 0 to 7 from top to bottom.

[0137] The first pair of row pixel points of the second upper left illumination compensation model is obtained by performing a weighted average on the pixel at position 1 of the upper side of the current block in FIG. 3a and the pixel at position 3 of the upper side of the current block in FIG. 4a (black circles in the figure) and performing a weighted average on the pixel at position 1 of the upper side of the reference block in FIG. 3a and the pixel at position 3 of the upper side of the reference block in FIG. 4a (black circles in the figure).

[0138] The second pair of row pixel points of the second upper left illumination compensation model is obtained by performing a weighted average on the pixel at position 7 of the upper side of the current block in FIG. 3a and the pixel at position 5 of the upper side of the current block in FIG. 4a (black circles in the figure) and performing a weighted average on the pixel at position 7 of the upper side of the reference block in FIG. 3a and the pixel at position 5 of the upper side of the reference block in FIG. 4a (black circles in the figure).

[0139] The first pair of column pixel points of the second upper left illumination compensation model is obtained by performing a weighted average on the pixel at position 1 of the left side of the current block in FIG. 3a and the pixel at position 3 of the left side of the current block in FIG. 4a (black circles in the figure) and performing a weighted average on the pixel at position 1 of the left side of the reference block in FIG. 3a and the pixel at position 3 of the left side of the reference block in FIG. 4a (black circles in the figure).

[0140] The second pair of column pixel points of the second upper left illumination compensation model is obtained by performing a weighted average on the pixel at position 7 of the left side of the current block in FIG. 3a and the pixel at position 5 of the left side of the current block in FIG. 4a (black circles in the figure) and performing a weighted average on the pixel at position 7 of the left side of the reference block in FIG. 3a and the pixel at position 5 of the left side of the reference block in FIG. 4a (black circles in the figure).

[0141] In summary, the four pairs of new pixel points of the second upper left illumination compensation model are obtained. In the subsequent process of correcting the prediction block of the current decoding unit by using the second upper left illumination compensation model, the scaling compensation coefficient and the offset compensation coefficient are determined based on the four pairs of new pixel points of the second upper left illumination compensation model.

[0142] In the above embodiment, the pixel point pairs of the second upper left illumination compensation model are re-determined according to the pixel point pairs corresponding to the first upper left illumination compensation model and the initial pixel point pairs at different positions corresponding to the second upper left illumination compensation model. The selection of the pixel point pairs for deriving the illumination compensation model coefficients is improved, and the positions of the pixel point pairs are more flexible, which is more suitable for the current encoding / decoding block corresponding to the derived illumination compensation model coefficients, thereby improving the encoding / decoding efficiency.

[0143] In an embodiment, the pixel point pairs corresponding to the second upper illumination compensation model are obtained by linear operation according to the position information of the pixel point pairs corresponding to the first upper illumination compensation model.

[0144] or the second up-light compensation model is obtained by weighted operation according to the pixel point pair corresponding to the first up-light compensation model and the initial pixel point pair of the second up-light compensation model; wherein the initial pixel point pair of the second up-light compensation model is obtained by linear operation according to the position information of the row pixel point pair corresponding to the first up-light compensation model.

[0145] It should be noted that the specific process of obtaining the pixel point pair corresponding to the second up-light compensation model by linear operation corresponds to the specific implementation process of obtaining the row pixel point pair corresponding to the second up-left light compensation model by linear operation in the above embodiment, which will not be described here.

[0146] It should be noted that the specific process of obtaining the pixel point pair corresponding to the second up-light compensation model by weighted operation corresponds to the specific implementation process of obtaining the row pixel point pair corresponding to the second up-left light compensation model by weighted operation in the above embodiment.

[0147] According to some embodiments, taking Figures 3b and 4b as examples, the determination process of the pixel point pair corresponding to the second up-light compensation model is described.

[0148] In this example, it is assumed that Figure 3b is a schematic of the pixel point pair of the first up-light compensation model, and Figure 4b is a schematic of the initial pixel point pair of the second up-light compensation model.

[0149] The pixel points (black circles in the figure) at positions 0 and 1 on the top side of the current block in Figure 3b and the pixel points (black circles in the figure) at positions 0 and 1 on the top side of the reference block in Figure 4b are weighted and averaged to obtain the first pair of pixel points corresponding to the second up-light compensation model.

[0150] The pixel points (black circles in the figure) at positions 2 and 3 on the top side of the current block in Figure 3b and the pixel points (black circles in the figure) at positions 2 and 3 on the top side of the reference block in Figure 4b are weighted and averaged to obtain the second pair of pixel points corresponding to the second up-light compensation model.

[0151] The pixel points (black circles in the figure) at positions 4 and 5 on the top side of the current block in Figure 3b and the pixel points (black circles in the figure) at positions 4 and 5 on the top side of the reference block in Figure 4b are weighted and averaged to obtain the third pair of pixel points corresponding to the second up-light compensation model.

[0152] The fourth pair of pixel points corresponding to the second upper light compensation model is obtained by performing a weighted average on the current block upper side position 6 pixel in Fig. 3b and the current block upper side position 7 pixel in Fig. 4b (black circles in the figure), and performing a weighted average on the reference block upper side position 6 pixel in Fig. 3b and the reference block upper side position 7 pixel in Fig. 4b (black circles in the figure).

[0153] In summary, the four pairs of new pixel points of the second upper light compensation model are obtained. In the subsequent process of correcting the prediction block of the current decoding unit by using the second upper light compensation model, the scaling compensation coefficient and the offset compensation coefficient are determined based on the four pairs of new pixel points of the second upper light compensation model.

[0154] In an embodiment, the pixel point pair corresponding to the second left light compensation model is obtained by linear operation according to the position information of the pixel point pair corresponding to the first left light compensation model.

[0155] Or the pixel point pair corresponding to the second left light compensation model is obtained by weighted operation according to the pixel point pair corresponding to the first left light compensation model and the initial pixel point pair of the second left light compensation model; wherein the initial pixel point pair of the second left light compensation model is obtained by linear operation according to the position information of the pixel point pair corresponding to the first left light compensation model.

[0156] It should be noted that the specific process of obtaining the pixel point pair corresponding to the second left light compensation model by linear operation corresponds to the specific implementation process of obtaining the row pixel point pair corresponding to the second left light compensation model by linear operation in the above embodiment, which will not be described here.

[0157] It should be noted that the specific process of obtaining the pixel point pair corresponding to the second left light compensation model by weighted operation corresponds to the specific implementation process of obtaining the column pixel point pair corresponding to the second left light compensation model by weighted operation in the above embodiment.

[0158] According to some embodiments, the determination process of the pixel point pair corresponding to the second left light compensation model is described by taking Figs. 3c and 4c as examples.

[0159] In this example, it is assumed that Fig. 3c is a schematic of the pixel point pair of the first left light compensation model, and Fig. 4c is a schematic of the initial pixel point pair of the second left light compensation model.

[0160] The first pair of pixel points corresponding to the second left light compensation model is obtained by performing a weighted average on the current block left side position 0 pixel in Fig. 3c and the current block left side position 1 pixel in Fig. 4c (black circles in the figure), and performing a weighted average on the reference block left side position 0 pixel in Fig. 3c and the reference block left side position 1 pixel in Fig. 4c (black circles in the figure).

[0161] The second pair of pixel points corresponding to the second left illumination compensation model is obtained by performing a weighted average on the pixels (black circles in the figure) at position 2 on the left side of the current block in FIG. 3c and at position 3 on the left side of the current block in FIG. 4c, and performing a weighted average on the pixels (black circles in the figure) at position 2 on the left side of the reference block in FIG. 3c and at position 3 on the left side of the reference block in FIG. 4c.

[0162] The third pair of pixel points corresponding to the second left illumination compensation model is obtained by performing a weighted average on the pixels (black circles in the figure) at position 4 on the left side of the current block in FIG. 3c and at position 5 on the left side of the current block in FIG. 4c, and performing a weighted average on the pixels (black circles in the figure) at position 4 on the left side of the reference block in FIG. 3c and at position 5 on the left side of the reference block in FIG. 4c.

[0163] The fourth pair of pixel points corresponding to the second left illumination compensation model is obtained by performing a weighted average on the pixels (black circles in the figure) at position 6 on the left side of the current block in FIG. 3c and at position 7 on the left side of the current block in FIG. 4c, and performing a weighted average on the pixels (black circles in the figure) at position 6 on the left side of the reference block in FIG. 3c and at position 7 on the left side of the reference block in FIG. 4c.

[0164] In summary, the four pairs of new pixel points of the second left illumination compensation model are obtained. In the subsequent process of correcting the prediction block of the current decoding unit by using the second left illumination compensation model, the scaling compensation coefficient and the offset compensation coefficient are determined based on the four pairs of new pixel points of the second left illumination compensation model.

[0165] In an embodiment, step 210 comprises:

[0166] The code stream is obtained and parsed to obtain the inter prediction value correction identifier, the motion information and the residual information of the current decoding unit.

[0167] In the case where the inter prediction value correction identifier is enabled, the code stream is parsed to obtain the illumination compensation index.

[0168] In the embodiment, the inter prediction value correction identifier refers to an identifier set to indicate whether the current decoding unit needs to be corrected. The inter prediction value correction identifier is determined by the encoding end in the encoding stage through rate distortion optimization (RDO) decision. When the encoding end determines not to enable IPC through RDO decision, the inter prediction value correction identifier is set to be disabled, and is written into the code stream. When the encoding end determines to enable IPC through RDO decision, the inter prediction value correction identifier is set to be enabled, and the illumination compensation model set to be corrected is determined. The inter prediction value correction identifier and the illumination compensation index of the illumination compensation model are written into the code stream.

[0169] When the inter prediction value correction flag is turned on, it indicates that the corresponding current decoding unit needs to be corrected. Therefore, the illumination compensation index needs to be further parsed from the bitstream to determine the corresponding illumination compensation model through the illumination compensation index. Thus, the prediction block of the current decoding unit is corrected based on the illumination compensation model.

[0170] When the inter prediction value correction flag is turned off, it indicates that the corresponding current decoding unit does not need to be corrected. In this case, there is no need to parse the illumination compensation index from the bitstream. The decoder directly obtains the reconstructed block of the current decoding unit according to the prediction block and the residual information of the current decoding unit.

[0171] In an embodiment, the prediction block of the current decoding unit includes a chrominance component and a luminance component. Step 230 includes:

[0172] According to the target illumination compensation information corresponding to the illumination compensation index, the scaling compensation coefficient and the offset compensation coefficient are obtained;

[0173] According to the scaling compensation coefficient and the offset compensation coefficient, the chrominance component and the luminance component are corrected to obtain the corrected chrominance component and the corrected luminance component.

[0174] In the present embodiment, the current encoding / decoding unit includes a chrominance component and a luminance component, the chrominance component representing color information of the current encoding / decoding unit, and the luminance component representing luminance information of the current encoding / decoding unit. Correspondingly, the prediction block obtained by performing motion compensation on the current encoding / decoding unit also includes a predicted chrominance component and a predicted luminance component. The scaling compensation coefficient and the offset compensation coefficient are obtained based on the target illumination compensation information, i.e., the pixel point pair of the illumination compensation model.

[0175] The chrominance component and the luminance component are corrected based on the following formula according to the scaling compensation coefficient and the offset compensation coefficient: Pred'(x, y) = a * pred(x, y) + b

[0176] Wherein, pred(x, y) is the prediction value of the prediction block of the current encoding / decoding unit at (x, y), Pred'(x, y) is the corrected prediction value at (x, y), a is the scaling compensation coefficient, and b is the offset compensation coefficient.

[0177] In the existing IPC algorithm, the correction of the prediction value of the chroma component is only performed when the scaling compensation coefficient a of the corresponding illumination compensation model is 0, and the prediction pixel values of all positions of the prediction block of the corresponding current coding / decoding unit are corrected to a fixed value (offset compensation coefficient b), which causes greater prediction distortion. The above embodiments of the present application remove the restriction of applying IPC to the chroma component, and not only apply IPC to the chroma component when the scaling compensation coefficient a is 0. Instead, the scaling compensation coefficient and the offset compensation coefficient are directly used to correct the inter-frame prediction value of the chroma component similar to that of the luminance component, so as to further improve the coding efficiency of the chroma component.

[0178] FIG. 7 is a flowchart of an inter-frame prediction method applied to a decoding end according to another embodiment of the present application. The inter-frame prediction method is applied to the decoding end. As shown in FIG. 7, the inter-frame prediction method includes at least but is not limited to steps 710 to 730.

[0179] In step 710, a bitstream is obtained and parsed to obtain an illumination compensation index, motion information and residual information of a current decoding unit.

[0180] In step 720, motion compensation is performed on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit.

[0181] In step 730, when neither an upper neighboring block nor a left neighboring block of the current decoding unit is available, a reconstructed block of the current decoding unit is obtained according to the prediction block and the residual information of the current decoding unit.

[0182] In the present embodiment, the related implementation details and features of steps 710 and 720 correspond to those of steps 210 and 220 of the above embodiments, which are not repeated here.

[0183] In step 730, when neither an upper neighboring block nor a left neighboring block of the current decoding unit is available, it means that there is no decoding unit adjacent to the current decoding unit on the upper side and the left side of the current decoding unit. In the present embodiment, if neither an upper neighboring block nor a left neighboring block of the current decoding unit is available, the prediction block of the current decoding unit is not corrected, and the reconstructed block is directly obtained based on the prediction block and the residual information of the current decoding unit.

[0184] In the related art, if the upper and left neighboring blocks are both unavailable, the prediction pixel value is corrected to a fixed value related to the coding bit depth (for example, when the coding bit depth is 10, all prediction pixel values of the current coding / decoding block are corrected to 512), which causes greater prediction distortion and low coding efficiency. In the above steps 710 to 730 of the present application, when the upper and left neighboring blocks of the current decoding unit are both unavailable, the reconstructed block is directly obtained based on the prediction block and the residual information of the current decoding unit, so as to reduce the prediction distortion and improve the coding efficiency.

[0185] In an embodiment, in the process of decoding a video frame, for a decoding unit in which the upper and / or left neighboring blocks are available, the inter prediction method provided in the above steps 210 to 240 and the related embodiments is used for decoding. When decoding a decoding unit in which the upper and left neighboring blocks are both unavailable, the inter prediction method provided in the steps 710 to 730 is used for decoding.

[0186] FIG. 8 is a flowchart of an inter prediction method applied to the encoding end according to an embodiment of the present application. The inter prediction method is applied to the encoding end. As shown in FIG. 8, the inter prediction method includes at least but is not limited to the steps 810 to 840.

[0187] Step 810, performing motion compensation on the current coding unit according to the motion information to obtain a prediction block of the current coding unit;

[0188] Step 820, correcting the prediction block of the current coding unit based on rate-distortion optimization according to the illumination compensation model;

[0189] Step 830, determining target illumination compensation information according to the rate-distortion optimization cost;

[0190] Step 840, writing an illumination compensation index corresponding to the target illumination compensation information into a bitstream;

[0191] The target illumination compensation information is one of the illumination compensation models; the illumination compensation model includes a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, and at least one of the following:

[0192] a second upper-left illumination compensation model;

[0193] a second upper illumination compensation model;

[0194] a second left illumination compensation model.

[0195] In step 810, the current coding unit refers to the data block being currently coded. The current coding unit in each embodiment of the present application corresponds to the current decoding unit, and both can be referred to as the current block.

[0196] In step 820, the prediction block of the current coding unit is corrected based on rate-distortion optimization according to each different illumination compensation model, so that the optimization result of the prediction block by each illumination compensation model corresponds to a rate-distortion optimization cost. Taking an example in which the illumination compensation models include a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, a second upper-left illumination compensation model, a second upper illumination compensation model, and a second left illumination compensation model, the prediction block of the current coding unit is corrected based on rate-distortion optimization according to the six illumination compensation models respectively, and six rate-distortion costs are obtained.

[0197] In step 830, based on the rate-distortion optimization costs corresponding to the illumination compensation models obtained from step 820, the minimum rate-distortion optimization cost is determined, and the illumination compensation model corresponding to the minimum rate-distortion optimization cost is determined as the target illumination compensation information.

[0198] In step 840, the encoder also writes the residual information obtained by motion compensation on the current coding unit and the motion information into the bitstream.

[0199] The steps 810 to 840 described above expand the types of illumination compensation models by adding at least one new illumination compensation model based on the rate-distortion optimization cost. In the process of coding and decoding, the possibility of selecting the illumination compensation model is increased. The illumination compensation model matched with the current coding unit is determined from multiple illumination compensation models based on the rate-distortion optimization cost. Since the types of illumination compensation models are expanded, a more suitable illumination compensation model for the current coding unit or the current decoding unit can be obtained, the effect of correcting the prediction block by using the illumination compensation model is improved, and finally the coding and decoding efficiency is improved.

[0200] The above is a general description of steps 810 to 840, and the specific implementation process of steps 810 to 840 will be described in detail below. It should be noted that steps 810 to 840 are corresponding to the inter prediction method applied to the encoder provided in the above embodiments, and the inter prediction method applied to the encoder, therefore, the related implementation details and features, beneficial effects of steps 810 to 840 and the following embodiments of the encoder correspond to the embodiments of the inter prediction method applied to the decoder, which will not be repeated here.

[0201] In an embodiment, the encoder decides whether to enable the inter prediction value correction by RDO. If it is decided to enable the inter prediction value correction, the inter prediction value correction flag is set to be on, and the inter prediction value correction flag and the illumination compensation index of the selected illumination compensation model are written into the bitstream. If it is decided not to enable the inter prediction value correction, the inter prediction value correction flag is set to be off, and the inter prediction value correction flag is directly written into the bitstream without writing the illumination compensation index.

[0202] In an embodiment, the first upper-left illumination compensation model and the second upper-left illumination compensation model are different; the first upper illumination compensation model and the second upper illumination compensation model are different; and the first left illumination compensation model and the second left illumination compensation model are different.

[0203] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0204] The pixel pair corresponding to the first upper-left illumination compensation model is different from the pixel pair corresponding to the second upper-left illumination compensation model;

[0205] The pixel pair corresponding to the first upper illumination compensation model is different from the pixel pair corresponding to the second upper illumination compensation model;

[0206] The pixel pair corresponding to the first left illumination compensation model is different from the pixel pair corresponding to the second left illumination compensation model;

[0207] The pixel pair is obtained according to the reconstructed pixels of the current coding unit and the reconstructed pixels of the reference unit matched by the current coding unit.

[0208] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0209] The reconstructed pixels of the current coding unit corresponding to the first upper-left illumination compensation model are the same as the reconstructed pixels of the current coding unit corresponding to the second upper-left illumination compensation model, and the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the first upper-left illumination compensation model are the same as the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the second upper-left illumination compensation model;

[0210] The reconstructed pixels of the current coding unit corresponding to the first upper illumination compensation model are the same as the reconstructed pixels of the current coding unit corresponding to the second upper illumination compensation model, and the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the first upper illumination compensation model are the same as the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the second upper illumination compensation model;

[0211] The reconstructed pixels of the current coding unit corresponding to the first left illumination compensation model are the same as the reconstructed pixels of the current coding unit corresponding to the second left illumination compensation model, and the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the first left illumination compensation model are the same as the reconstructed pixels of the reference unit matched by the current coding unit corresponding to the second left illumination compensation model.

[0212] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0213] The pixel point pair corresponding to the illumination compensation model is obtained according to the width of the current coding unit or the width of the reference unit matched by the current decoding unit.

[0214] The pixel point pair corresponding to the illumination compensation model is obtained according to the height of the current coding unit or the height of the reference unit matched by the current decoding unit.

[0215] In an embodiment, the illumination compensation model at least satisfies one of the following:

[0216] The pixel point pair corresponding to the second upper left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first upper left illumination compensation model.

[0217] The pixel point pair corresponding to the second upper illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first upper illumination compensation model.

[0218] The pixel point pair corresponding to the second left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first left illumination compensation model.

[0219] In an embodiment, the pixel point pair corresponding to the second upper left illumination compensation model at least includes one of the following:

[0220] The row pixel point pair corresponding to the second upper left illumination compensation model;

[0221] The column pixel point pair corresponding to the second upper left illumination compensation model;

[0222] The pixel point pair corresponding to the first upper left illumination compensation model at least includes one of the following:

[0223] The row pixel point pair corresponding to the first upper left illumination compensation model;

[0224] The column pixel point pair corresponding to the first upper left illumination compensation model.

[0225] In an embodiment, the row pixel point pair corresponding to the second upper left illumination compensation model is obtained by linear operation according to the position information of the row pixel point pair corresponding to the first upper left illumination compensation model.

[0226] or the column pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the column pixel point pair corresponding to the first up-left illumination compensation model and the initial column pixel point pair of the second up-left illumination compensation model by weighted operation.

[0227] In an embodiment, the row pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the row pixel point pair corresponding to the first up-left illumination compensation model and the initial row pixel point pair of the second up-left illumination compensation model by weighted operation.

[0228] or the column pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the column pixel point pair corresponding to the first up-left illumination compensation model and the initial column pixel point pair of the second up-left illumination compensation model by weighted operation.

[0229] The initial row pixel point pair of the second up-left illumination compensation model is obtained according to the position information of the row pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0230] The initial column pixel point pair of the second up-left illumination compensation model is obtained according to the position information of the column pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0231] In an embodiment, the pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0232] or the pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the pixel point pair corresponding to the first up-left illumination compensation model and the initial pixel point pair of the second up-left illumination compensation model by weighted operation; wherein the initial pixel point pair of the second up-left illumination compensation model is obtained according to the position information of the row pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0233] In an embodiment, the pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0234] or the pixel point pair corresponding to the second up-left illumination compensation model is obtained according to the pixel point pair corresponding to the first up-left illumination compensation model and the initial pixel point pair of the second up-left illumination compensation model by weighted operation; wherein the initial pixel point pair of the second up-left illumination compensation model is obtained according to the position information of the row pixel point pair corresponding to the first up-left illumination compensation model by linear operation.

[0235] In order to further illustrate the method provided by the embodiments of the present application, the following examples are further described in detail.

[0236] Example 1:

[0237] Assuming that Fig. 3a to Fig. 3c are the available pixel pairs of three cases of the first upper left illumination compensation model IC_TL, and Fig. 4a to Fig. 4c are the available pixel pairs of three cases of the second upper left illumination compensation model IC_TL_NEW.

[0238] On the basis of the first upper left illumination compensation model IC_TL, the first upper illumination compensation model IC_T, and the first left illumination compensation model IC_L, the second upper left illumination compensation model IC_TL_NEW is added. The newly added second upper left illumination compensation model is different from the IC_TL model, and the difference lies in the different selection method of the 4 available pixel pairs set to calculate the scaling compensation coefficient a and the offset compensation coefficient b of the illumination compensation model.

[0239] The IC_TL model selects 2 pixel pairs from the upper side and 2 pixel pairs from the left side when both the upper side and the left side pixels are "available", as shown in Fig. 3a; when only the upper side pixels are available, 4 pixel pairs are selected from the upper side, and the selected positions are width: 0 / 4, 1 / 4, 2 / 4, 3 / 4, as shown in Fig. 3b; when only the left side pixels are available, 4 pixel pairs are selected from the left side, and the selected positions are height: 0 / 4, 1 / 4, 2 / 4, 3 / 4, as shown in Fig. 3c.

[0240] The newly added IC_TL_NEW model is another IC_TL model. For the IC_TL_NEW model, when both the upper side and the left side pixels are "available", 2 pixel pairs are selected from the upper side, and the selected positions are width: 3 / 8, 5 / 8, 2 pixel pairs are selected from the left side, and the selected positions are height: 3 / 8, 5 / 8, as shown in Fig. 4a; when only the upper side pixels are available, 4 pixel pairs are selected from the upper side, and the selected positions are width: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8, as shown in Fig. 4b; when only the left side pixels are available, 4 pixel pairs are selected from the left side, and the selected positions are height: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8, as shown in Fig. 4c. It should be noted that the selection positions of the 4 pixel pairs of the newly added IC_TL_NEW model are not limited to the examples listed, and as long as the selection method is different from the selection positions of the 4 pixel pairs of the IC_TL illumination compensation model, it is acceptable.

[0241] After adding a new illumination compensation model, the number of illumination compensation models changes from 3 to 4, therefore, the block-level syntax identification representation method needs to be adjusted, and the adjusted block-level syntax identification representation method is shown in Table 1.

[0242] Table 1, block-level syntax identification after adding IC_TL_NEW

[0243] Wherein, IC_flag is an inter prediction value correction flag, when IC_flag is 0, it indicates that the inter prediction value correction flag is closed, when IC_flag is 1, it indicates that the inter prediction value correction flag is opened. IC_index0 and IC_index1 jointly constitute an illumination compensation index.

[0244] At the encoding end, after RDO judgment, if it is determined that IPC is not enabled, IC_flag is directly written into the code stream. If it is determined that IPC is enabled, IC_flag and the illumination compensation index corresponding to the selected illumination compensation model are written into the code stream and sent to the decoding end. At the decoding end, IC_flag is first parsed from the received code stream. If IC_flag indicates that IPC is not enabled (IC_flag is 0), the prediction block is not corrected by IPC. If IC_flag indicates that IPC is enabled (IC_flag is 1), the illumination compensation index of the selected illumination compensation model also needs to be further parsed from the code stream, and the prediction block is corrected by the parsed illumination compensation model.

[0245] In addition, the existing IPC algorithm is a correction method with relatively large limitation for the correction of the prediction value of the chroma component. IPC is only applied when the scaling compensation coefficient a of the illumination compensation model is 0. At this time, the prediction pixel values of all positions of the current block are corrected to a fixed value b, which will cause greater prediction distortion and limit the further improvement of the coding efficiency of the chroma component to some extent. In the present example, in order to obtain higher coding efficiency, the restriction condition of applying IPC to the chroma component is released, so that the chroma component and the luminance component are corrected according to the inter prediction method provided in each embodiment of the present application.

[0246] Similar situations also occur when the upper and left adjacent blocks of the current block are both unavailable. In the related art, the prediction pixel values are directly corrected to a fixed value (such as 512) related to the coding bit depth, which will cause greater prediction distortion and low coding efficiency. In the present example, in order to obtain higher coding efficiency, when such a situation occurs, IPC is not applied to correct the inter prediction value, and the corresponding IC_flag is 0.

[0247] Example 2:

[0248] On the basis of example 1, the scaling compensation coefficient and the offset compensation coefficient derivation method of the IC_TL_NEW model are improved. The scaling compensation coefficient a and the offset compensation coefficient b are calculated from the four available pixel points of the illumination compensation model.

[0249] In the IC TL NEW model in the present example, when both the upper and left pixels are "available", four new pixel pairs are constructed by weighted average of the pixel pairs of IC TL shown in Fig. 3a and the pixel pairs of IC TL NEW shown in Fig. 4a. The four new pixel pairs are used for derivation of the scaling compensation coefficient and the offset compensation coefficient of the IC TL NEW model. Taking Figs. 3a and 4a as an example, the pixels (black circles in the figure) at positions 1 on the upper side of the current block and the reference block in Fig. 3a and at positions 3 on the upper side of the current block and the reference block in Fig. 4a are weighted and averaged respectively, the pixels (black circles in the figure) at positions 7 on the upper side of the current block and the reference block in Fig. 3a and at positions 5 on the upper side of the current block and the reference block in Fig. 4a are weighted and averaged respectively, the pixels (black circles in the figure) at positions 1 on the left side of the current block and the reference block in Fig. 3a and at positions 3 on the left side of the current block and the reference block in Fig. 4a are weighted and averaged respectively, and the pixels (black circles in the figure) at positions 7 on the left side of the current block and the reference block in Fig. 3a and at positions 5 on the left side of the current block and the reference block in Fig. 4a are weighted and averaged respectively, thereby constructing four new pixel pairs.

[0250] When only the upper pixels are "available", four new pixel pairs are constructed by weighted average of the pixel pairs of IC TL shown in Fig. 3b and the pixel pairs of IC TL NEW shown in Fig. 4b. The four new pixel pairs are used for derivation of the scaling compensation coefficient and the offset compensation coefficient of the IC TL NEW model. Taking Figs. 3b and 4b as an example, the pixels (black circles in the figure) at positions 0 of the current block and the reference block in Fig. 3b and at positions 1 of the current block and the reference block in Fig. 4b are weighted and averaged respectively; similarly, the pixels at positions 2 and 3, the pixels at positions 4 and 5, and the pixels at positions 6 and 7 are weighted and averaged, and the results of the weighted average are used to construct four new pixel pairs.

[0251] When only the left pixels are "available", four new pixel pairs are constructed by weighted average of the pixel pairs of IC TL shown in Fig. 3c and the pixel pairs of IC TL NEW shown in Fig. 4c. The four new pixel pairs are used for derivation of the scaling compensation coefficient and the offset compensation coefficient of the IC TL NEW model. Similarly, the pixel pair construction method when only the left pixels are "available" is similar to the pixel pair construction method when only the upper pixels are "available".

[0252] The pixel pair construction method of weighted average provided in the present example can compensate for the problem of large deviation in model coefficient calculation caused by uneven distribution of pixel pairs.

[0253] Example 3:

[0254] On the basis of example 1, the scaling compensation coefficient and the offset compensation coefficient derivation method of IC_T_NEW model and IC_L_NEW are improved. The scaling compensation coefficient a and the offset compensation coefficient b are calculated from the 4 available pixel pairs of the illumination compensation model.

[0255] In this example, 4 new pixel pairs are constructed by weighted average of the pixel pairs of IC_T shown in Fig. 3b and the pixel pairs of IC_T_NEW shown in Fig. 4b. The 4 new pixel pairs are used as the derivation of the scaling compensation coefficient and the offset compensation coefficient of IC_T_NEW model. Taking Fig. 3b and Fig. 4b as an example, the pixels of position 0 of the current block and the reference block in Fig. 3b and the pixels of position 1 of the current block and the reference block in Fig. 4b (black circles in the figure) are weighted and averaged respectively; similarly, the pixels of position 2 and position 3, the pixels of position 4 and position 5, and the pixels of position 6 and position 7 are weighted and averaged, and the results of the weighted average are used to construct 4 new pixel pairs.

[0256] 4 new pixel pairs are constructed by weighted average of the pixel pairs of IC_L shown in Fig. 3c and the pixel pairs of IC_L_NEW shown in Fig. 4c. The 4 new pixel pairs are used as the derivation of the scaling compensation coefficient and the offset compensation coefficient of IC_TL_NEW model. Similarly, the construction process of the 4 new pixel pairs of IC_L_NEW model is similar to that of IC_T_NEW model.

[0257] Example 4:

[0258] Assume that Fig. 3b is a schematic of the available pixel pairs of the first upper illumination compensation model IC_T, and Fig. 4b is a schematic of the available pixel pairs of the second upper illumination compensation model IC_T_NEW.

[0259] On the basis of the first upper-left illumination compensation model IC_TL, the first upper illumination compensation model IC_T, and the first left illumination compensation model IC_L, the second upper illumination compensation model IC_T_NEW is added. The newly added IC_T_NEW model is different from the IC_T model. The 4 pixel pairs of the IC_T model are all selected from the top side, and the selected positions are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the width, as shown in Fig. 3b.

[0260] The newly added IC_T_NEW model is another IC_T model. The 4 pixel pairs of the IC_T_NEW model are all selected from the top side, and the selected positions are: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8 of the width, as shown in Fig. 4b. It should be noted that the selection positions of the 4 pixel pairs of the newly added IC_T_NEW model are not limited to the examples listed, and any selection method that is different from the selection positions of the 4 pixel pairs of the IC_T model is acceptable.

[0261] After adding one illumination compensation model, the number of illumination compensation models is changed from 3 to 4, therefore, the block level syntax identification representation method needs to be adjusted, and the adjusted block level syntax identification representation method is shown in Table 2.

[0262] Table 2, block level syntax identification after adding IC_T_NEW

[0263] Wherein, IC_flag is the inter prediction value correction identification, when IC_flag is 0, it indicates that the inter prediction value correction identification is closed, when IC_flag is 1, it indicates that the inter prediction value correction identification is opened. IC_index0 and IC_index1 jointly constitute the illumination compensation index.

[0264] At the encoding end, after RDO judgment, if it is judged that IPC is not enabled, IC_flag is directly written into the code stream. If it is judged that IPC is enabled, IC_flag and the illumination compensation index corresponding to the selected illumination compensation model are written into the code stream and sent to the decoding end. At the decoding end, IC_flag is first parsed from the received code stream, if IC_flag indicates that IPC is not enabled (IC_flag is 0), the prediction block is not corrected by IPC. If IC_flag indicates that IPC is enabled (IC_flag is 1), the illumination compensation index of the selected illumination compensation model needs to be further parsed from the code stream, and the prediction block is corrected by the parsed illumination compensation model.

[0265] Example 5:

[0266] Assuming that Fig. 3c is a schematic diagram of available pixel point pairs of the first left illumination compensation model IC_L, and Fig. 4c is a schematic diagram of available pixel point pairs of the second left illumination compensation model IC_L_NEW.

[0267] On the basis of the first upper left illumination compensation model IC_TL, the first upper illumination compensation model IC_T and the first left illumination compensation model IC_L, the second left illumination compensation model IC_L_NEW is added. The added IC_L_NEW model is different from the IC_L model. The 4 pixel point pairs of the IC_L model are all selected from the left side, and the selected positions are height: 0 / 4, 1 / 4, 2 / 4, 3 / 4, as shown in Fig. 3c.

[0268] The added IC_L_NEW model is another IC_L model. The four pixel point pairs of the IC_L_NEW model 4 are selected from the left side, and the selected positions are heights of 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, and 3 / 4+1 / 8, as shown in FIG. 4c. It should be noted that the selected positions of the four pixel point pairs of the added IC_L_NEW model are not limited to the examples listed, and any selection method different from the selection method of the four pixel point pairs of the IC_L model is acceptable.

[0269] After adding the illumination compensation model, the number of illumination compensation models is changed from 3 to 4, and therefore, the block-level syntax identification representation method needs to be adjusted. The adjusted block-level syntax identification representation method is shown in Table 3.

[0270] Table 3: Block-level syntax identification after adding IC_L_NEW

[0271] IC_flag is an inter-prediction value correction identifier. When IC_flag is 0, the inter-prediction value correction identifier is closed. When IC_flag is 1, the inter-prediction value correction identifier is opened. IC_index0 and IC_index1 together form an illumination compensation index.

[0272] At the encoding end, after RDO judgment, if it is determined that IPC is not enabled, IC_flag is directly written into the code stream. If it is determined that IPC is enabled, IC_flag and the illumination compensation index corresponding to the selected illumination compensation model are written into the code stream and sent to the decoding end. At the decoding end, IC_flag is first parsed from the received code stream. If IC_flag indicates that IPC is not enabled (IC_flag is 0), the prediction block is not corrected by IPC. If IC_flag indicates that IPC is enabled (IC_flag is 1), the illumination compensation index of the selected illumination compensation model needs to be further parsed from the code stream, and the prediction block is corrected by the parsed illumination compensation model.

[0273] Example 6:

[0274] FIGS. 3a to 3c are schematic diagrams of available pixel points of the first upper-left illumination compensation model IC_TL, and FIG. 3b can also be used as a schematic diagram of available pixel points of the first upper illumination compensation model IC_T, and FIG. 3c can also be used as a schematic diagram of available pixel points of the first left illumination compensation model IC_L.

[0275] Figures 4a-4c are schematic diagrams of available pixel pairs for the second top-left illumination compensation model IC_TL_NEW, and 4b can also be a schematic diagram of available pixel pairs for the second top illumination compensation model IC_T_NEW, and 4c can also be a schematic diagram of available pixel pairs for the second left illumination compensation model IC_L_NEW.

[0276] The present example adds three illumination compensation models to the first top-left illumination compensation model IC_TL, the first top illumination compensation model IC_T, and the first left illumination compensation model IC_L: the second top-left illumination compensation model IC_TL_NEW, the second top illumination compensation model IC_T_NEW, and the second left illumination compensation model IC_L_NEW. The newly added IC_TL_NEW, IC_T_NEW, and IC_L_NEW are different from the IC_TL, IC_T, and IC_L models in that the selection method of the four available pixel pairs used to set the scaling compensation coefficient a and the offset compensation coefficient b of the corresponding illumination compensation model is different.

[0277] The IC_TL model selects two pixel pairs from the top side and two pixel pairs from the left side when both the top side and the left side pixels are "available", as shown in Figure 3a; selects four pixel pairs from the top side when only the top side pixels are available, and the selected positions are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the width, as shown in Figure 3b; and selects four pixel pairs from the left side when only the left side pixels are available, and the selected positions are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the height, as shown in Figure 3c.

[0278] The newly added IC_TL_NEW model is another IC_TL model. For the IC_TL_NEW model, two pixel pairs are selected from the top side when both the top side and the left side pixels are "available", and the selected positions are: 3 / 8, 5 / 8 of the width, and two pixel pairs are selected from the left side, and the selected positions are: 3 / 8, 5 / 8 of the height, as shown in Figure 4a; four pixel pairs are selected from the top side when only the top side pixels are available, and the selected positions are: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8 of the width, as shown in Figure 4b; and four pixel pairs are selected from the left side when only the left side pixels are available, and the selected positions are: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8 of the height, as shown in Figure 4c.

[0279] The IC_T model selects four pixel pairs from the top side, and the selected positions are: 0 / 4, 1 / 4, 2 / 4, 3 / 4 of the width, as shown in Figure 3b.

[0280] The added IC_T_NEW model is another IC_T model. The four pixel pairs of the IC_T_NEW model are all selected from the top side, and the selected positions are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, and 3 / 4+1 / 8 of the width, as shown in FIG. 4b.

[0281] The four pixel pairs of the IC_L model are all selected from the left side, and the selected positions are 0 / 4, 1 / 4, 2 / 4, and 3 / 4 of the height, as shown in FIG. 3c.

[0282] The added IC_L_NEW model is another IC_L model. The four pixel pairs of the IC_L_NEW model are all selected from the left side, and the selected positions are 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, and 3 / 4+1 / 8 of the height, as shown in FIG. 4c.

[0283] It should be noted that the selected positions of the four pixel pairs of the three added models are not limited to the examples listed, and any selection method different from the selection positions of the four pixel pairs of the IC_TL, IC_T, and IC_L models is acceptable.

[0284] After adding the three illumination compensation models, the number of illumination compensation models is changed from three to six, and therefore, the block-level syntax identification representation method needs to be adjusted. The adjusted block-level syntax identification representation method is shown in Table 4.

[0285] Table 4. Block-level syntax identification after adding IC_TL_NEW, IC_T_NEW, and IC_L_NEW

[0286] IC_flag is an inter-prediction value correction flag. When IC_flag is 0, the inter-prediction value correction flag is closed. When IC_flag is 1, the inter-prediction value correction flag is opened. IC_index0, IC_index1, and IC_index2 jointly form an illumination compensation index.

[0287] At the encoding end, if it is determined that IPC is not enabled after RDO judgment, IC_flag is directly written into the code stream. If it is determined that IPC is enabled, IC_flag and the illumination compensation index corresponding to the selected illumination compensation model are written into the code stream and sent to the decoding end. At the decoding end, IC_flag is first parsed from the received code stream. If IC_flag indicates that IPC is not enabled (IC_flag is 0), the prediction block is not corrected by IPC. If IC_flag indicates that IPC is enabled (IC_flag is 1), the illumination compensation index of the selected illumination compensation model needs to be further parsed from the code stream, and the prediction block is corrected by the parsed illumination compensation model.

[0288] Example 7:

[0289] Assuming that Figs. 3a-3c are a schematic of available pixel pairs for three cases of the first upper-left illumination compensation model IC TL, and Figs. 4a-4c are a schematic of available pixel pairs for three cases of the second upper-left illumination compensation model IC TL NEW.

[0290] On the basis of the first upper-left illumination compensation model IC TL, the first upper illumination compensation model IC T, and the first left illumination compensation model IC L, the second upper-left illumination compensation model IC TL NEW is added. The newly added second upper-left illumination compensation model is different from the IC TL model, and the difference lies in the selection method of the 4 available pixel pairs set to calculate the scaling compensation coefficient a and the offset compensation coefficient b of the illumination compensation model.

[0291] The IC TL model selects 2 pixel pairs from the upper side and 2 pixel pairs from the left side when both the upper and left pixels are "available", as shown in Fig. 3a; when only the upper pixels are available, 4 pixel pairs are selected from the upper side, and the selected positions are width: 0 / 4, 1 / 4, 2 / 4, 3 / 4, as shown in Fig. 3b; when only the left pixels are available, 4 pixel pairs are selected from the left side, and the selected positions are height: 0 / 4, 1 / 4, 2 / 4, 3 / 4, as shown in Fig. 3c.

[0292] The newly added IC TL NEW model is another IC TL model, which improves the selection method of the 4 available pixel pairs set to calculate the scaling compensation coefficient a and the offset compensation coefficient b of the existing illumination compensation model.

[0293] For the IC TL NEW model, when both the upper and left pixels are "available", 4 pixel pairs are obtained according to the following steps.

[0294] Step 1: 2 pixel pairs are selected from the upper side, and the selected positions are width: 3 / 8, 5 / 8; 2 pixel pairs are selected from the left side, and the selected positions are height: 3 / 8, 5 / 8, as shown in Fig. 4a;

[0295] Step 2: Four new pixel pairs are constructed by weighted average of the four pixel pairs of IC_TL shown in Fig. 3a and the four pixel pairs of IC_TL_NEW shown in Fig. 4a. The four new pixel pairs are used as the derivation of scaling compensation coefficient and offset compensation coefficient of IC_TL_NEW model. Taking Figs. 3a and 4a as examples, the pixels (black circles in the figure) at positions 1 on the upper side of the current block and the reference block in Fig. 3a and the pixels at positions 3 on the upper side of the current block and the reference block in Fig. 4a are respectively weighted and averaged, the pixels (black circles in the figure) at positions 7 on the upper side of the current block and the reference block in Fig. 3a and the pixels at positions 5 on the upper side of the current block and the reference block in Fig. 4a are respectively weighted and averaged, the pixels (black circles in the figure) at positions 1 on the left side of the current block and the reference block in Fig. 3a and the pixels at positions 3 on the left side of the current block and the reference block in Fig. 4a are respectively weighted and averaged, and the pixels (black circles in the figure) at positions 7 on the left side of the current block and the reference block in Fig. 3a and the pixels at positions 5 on the left side of the current block and the reference block in Fig. 4a are respectively weighted and averaged, thereby constructing the four new pixel pairs.

[0296] Only the upper side pixels are available, the four pixel pairs are all selected from the upper side, and the selected positions are: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8 as shown in Fig. 4b. Only the left side pixels are available, the four pixel pairs are all selected from the left side, and the selected positions are: 0 / 4+1 / 8, 1 / 4+1 / 8, 2 / 4+1 / 8, 3 / 4+1 / 8 as shown in Fig. 4c. It should be noted that the selection positions of the four pixel pairs of the newly added IC_TL_NEW model are not limited to the examples listed, and any selection method different from the selection positions of the four pixel pairs of the IC_TL illumination compensation model is acceptable.

[0297] After adding a new illumination compensation model, the number of illumination compensation models changes from three to four, and therefore the block-level syntax identification representation method needs to be adjusted. The adjusted block-level syntax identification representation method is shown in Table 5.

[0298] Table 5: Block-level syntax identification after adding IC_TL_NEW

[0299] IC_flag is the inter prediction value correction flag. When IC_flag is 0, the inter prediction value correction flag is closed. When IC_flag is 1, the inter prediction value correction flag is opened. IC_index0 and IC_index1 together constitute the illumination compensation index.

[0300] At the encoding end, after RDO decision, if it is decided not to enable IPC, IC_flag is directly written into the bitstream. If it is decided to enable IPC, IC_flag and the illumination compensation index corresponding to the selected illumination compensation model are written into the bitstream and sent to the decoding end. At the decoding end, IC_flag is first parsed from the received bitstream. If IC_flag indicates that IPC is not enabled (IC_flag is 0), the prediction block is not corrected by IPC. If IC_flag indicates that IPC is enabled (IC_flag is 1), the illumination compensation index of the selected illumination compensation model also needs to be further parsed from the bitstream, and the prediction block is corrected by the parsed illumination compensation model.

[0301] Example 8:

[0302] On the basis of example 7, the scaling compensation coefficient and the offset compensation coefficient derivation method of IC_T_NEW model and IC_L_NEW are improved. The scaling compensation coefficient a and the offset compensation coefficient b are calculated from the 4 available pixel pairs of the illumination compensation model.

[0303] In this example, 4 new pixel pairs are constructed by weighted average of the pixel pairs of IC_T shown in Fig. 3b and the pixel pairs of IC_T_NEW shown in Fig. 4b. The 4 new pixel pairs are used for derivation of the scaling compensation coefficient and the offset compensation coefficient of IC_T_NEW model. Taking Fig. 3b and Fig. 4b as an example, the pixels at position 0 of the current block and the reference block in Fig. 3b and the pixels at position 1 of the current block and the reference block in Fig. 4b (black circles in the figure) are weighted and averaged; similarly, the pixels at position 2 and position 3, the pixels at position 4 and position 5, and the pixels at position 6 and position 7 are weighted and averaged, and the results of the weighted average are used to construct 4 new pixel pairs.

[0304] 4 new pixel pairs are constructed by weighted average of the pixel pairs of IC_L shown in Fig. 3c and the pixel pairs of IC_L_NEW shown in Fig. 4c. The 4 new pixel pairs are used for derivation of the scaling compensation coefficient and the offset compensation coefficient of IC_TL_NEW model. Similarly, the construction process of the 4 new pixel pairs of IC_L_NEW model is similar to that of IC_T_NEW model. This weighted average construction method can to some extent make up for the problem of large deviation of the compensation coefficient of the illumination compensation model caused by uneven distribution of pixel pairs.

[0305] FIG. 9 is a structural schematic diagram of an electronic device according to an embodiment of the present application. As shown in FIG. 9, the electronic device 2000 includes a memory 2100 and a processor 2200. The number of the memory 2100 and the processor 2200 can be one or more, and one memory 2101 and one processor 2201 are taken as an example in FIG. 9. The memory 2101 and the processor 2201 in the network device can be connected through a bus or other means, and connection through a bus is taken as an example in FIG. 9.

[0306] The memory 2101 is a computer readable storage medium, which can be configured to store software programs, computer executable programs and modules, such as program instructions / modules of the method provided in any embodiment of the present application. The processor 2201 realizes the inter prediction method provided in any embodiment of the present application by running the software programs, instructions and modules stored in the memory 2101.

[0307] The memory 2101 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function. In addition, the memory 2101 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device or other non-volatile solid-state storage device. In some examples, the memory 2101 further includes a memory remotely arranged with respect to the processor 2201, and these remote memories can be connected to the device through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0308] An embodiment of the present application further provides a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions are configured to execute the inter prediction method provided in any embodiment of the present application.

[0309] An embodiment of the present application further provides a computer program product, which includes a computer program or computer instructions, and the computer program or computer instructions are stored in a computer readable storage medium. A processor of a computer device reads the computer program or computer instructions from the computer readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device executes the inter prediction method provided in any embodiment of the present application.

[0310] The system architecture and application scenarios described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. It can be known by those skilled in the art that, with the evolution of system architecture and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0311] It is understood by those of ordinary skill in the art that all or some of the steps in the methods disclosed above, the functional modules / units in the systems, devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0312] In hardware implementation, the division of the functional modules / units referred to in the above description does not necessarily correspond to the division of physical components; for example, one physical component can have multiple functions, or one function or step can be performed by several physical components in cooperation. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on computer readable media, which can include computer storage media (or non-transitory media) and communication media (or transitory media). As is well known to those of ordinary skill in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium which can be configured to store the desired information and which can be accessed by a computer. Further, it is understood by those of ordinary skill in the art that communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.

[0313] The terms "component," "module," "system," and the like are used generically herein to represent a computer-related entity, hardware, firmware, a combination of hardware and software, software, or a combination of software and executable instructions. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, or a computer. By way of illustration, both an application running on a computing device and the computing device can be a component. One or more components can reside within a process and / or thread of execution and a component can be localized, partially localized, or distributed across two or more computers or other processing devices. Also, these components can execute from various computer readable media having various data structures stored thereon. The components can communicate by way of local or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, or across a network such as the Internet with other systems via the signal).

[0314] The above describes some embodiments of the present application with reference to the accompanying drawings, and is not limited to the scope of the present application. Any modification, equivalent replacement and improvement made by those skilled in the art without departing from the scope and essence of the present application shall be within the scope of the present application.

Claims

1. An inter prediction method applied to a decoding end, the method comprising: obtaining a bitstream and parsing the bitstream to obtain an illumination compensation index, motion information and residual information of a current decoding unit; performing motion compensation on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit; correcting the prediction block of the current decoding unit according to target illumination compensation information corresponding to the illumination compensation index to obtain a corrected current decoding unit prediction block; obtaining a reconstructed block of the current decoding unit according to the corrected current decoding unit prediction block and the residual information. wherein the target illumination compensation information is one of illumination compensation models; the illumination compensation models comprise a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, and at least one of the following: a second upper-left illumination compensation model; a second upper illumination compensation model; a second left illumination compensation model.

2. The inter prediction method of claim 1, wherein: the first upper-left illumination compensation model and the second upper-left illumination compensation model are different; the first upper illumination compensation model and the second upper illumination compensation model are different; the first left illumination compensation model and the second left illumination compensation model are different.

3. The inter prediction method of claim 2, wherein, the illumination compensation models at least satisfy one of the following: a pixel point pair corresponding to the first upper-left illumination compensation model is different from a pixel point pair corresponding to the second upper-left illumination compensation model; a pixel point pair corresponding to the first upper illumination compensation model is different from a pixel point pair corresponding to the second upper illumination compensation model; a pixel point pair corresponding to the first left illumination compensation model is different from a pixel point pair corresponding to the second left illumination compensation model; wherein the pixel point pair is obtained according to a reconstructed pixel of the current decoding unit and a reconstructed pixel of a reference unit matched with the current decoding unit.

4. The inter prediction method of claim 3, wherein, the illumination compensation models at least satisfy one of the following: a reconstructed pixel of the current decoding unit corresponding to the first upper-left illumination compensation model is the same as a reconstructed pixel of the current decoding unit corresponding to the second upper-left illumination compensation model, and a reconstructed pixel of a reference unit matched with the current decoding unit corresponding to the first upper-left illumination compensation model is the same as a reconstructed pixel of the reference unit matched with the current decoding unit corresponding to the second upper-left illumination compensation model; a reconstructed pixel of the current decoding unit corresponding to the first upper illumination compensation model is the same as a reconstructed pixel of the current decoding unit corresponding to the second upper illumination compensation model, and a reconstructed pixel of a reference unit matched with the current decoding unit corresponding to the first upper illumination compensation model is the same as a reconstructed pixel of the reference unit matched with the current decoding unit corresponding to the second upper illumination compensation model; a reconstructed pixel of the current decoding unit corresponding to the first left illumination compensation model is the same as a reconstructed pixel of the current decoding unit corresponding to the second left illumination compensation model, and a reconstructed pixel of a reference unit matched with the current decoding unit corresponding to the first left illumination compensation model is the same as a reconstructed pixel of the reference unit matched with the current decoding unit corresponding to the second left illumination compensation model.

5. The inter prediction method of claim 3 or 4, wherein, The illumination compensation model at least meets one of the following conditions: The pixel point pair corresponding to the illumination compensation model is obtained according to the width of the current decoding unit or the width of the reference unit matched with the current decoding unit. The pixel point pair corresponding to the illumination compensation model is obtained according to the height of the current decoding unit or the height of the reference unit matched with the current decoding unit.

6. The inter prediction method of claim 3 or 4, wherein, The illumination compensation model at least meets one of the following conditions: The pixel point pair corresponding to the second upper left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first upper left illumination compensation model; The pixel point pair corresponding to the second upper illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first upper illumination compensation model; The pixel point pair corresponding to the second left illumination compensation model is obtained according to the position information of the pixel point pair corresponding to the first left illumination compensation model.

7. The inter prediction method of claim 6, wherein: The pixel point pair corresponding to the second upper left illumination compensation model at least includes one of the following: The row pixel point pair corresponding to the second upper left illumination compensation model; The column pixel point pair corresponding to the second upper left illumination compensation model; The pixel point pair corresponding to the first upper left illumination compensation model at least includes one of the following: The row pixel point pair corresponding to the first upper left illumination compensation model; The column pixel point pair corresponding to the first upper left illumination compensation model.

8. The inter prediction method of claim 7, wherein: The row pixel point pair corresponding to the second upper left illumination compensation model is obtained by linear operation according to the position information of the row pixel point pair corresponding to the first upper left illumination compensation model; Or the column pixel point pair corresponding to the second upper left illumination compensation model is obtained by linear operation according to the position information of the column pixel point pair corresponding to the first upper left illumination compensation model.

9. The inter prediction method of claim 7, wherein: The row pixel point pair corresponding to the second upper left illumination compensation model is obtained by weighted operation according to the row pixel point pair corresponding to the first upper left illumination compensation model and the initial row pixel point pair of the second upper left illumination compensation model; Or the column pixel point pair corresponding to the second upper left illumination compensation model is obtained by weighted operation according to the column pixel point pair corresponding to the first upper left illumination compensation model and the initial column pixel point pair of the second upper left illumination compensation model; Wherein, the initial row pixel point pair of the second upper left illumination compensation model is obtained by linear operation according to the position information of the row pixel point pair corresponding to the first upper left illumination compensation model; The initial column pixel point pair of the second upper left illumination compensation model is obtained by linear operation according to the position information of the column pixel point pair corresponding to the first upper left illumination compensation model.

10. The inter prediction method of claim 6, wherein: The pixel point pair corresponding to the second upper illumination compensation model is obtained by linear operation according to the position information of the pixel point pair corresponding to the first upper illumination compensation model; or the second left light compensation model corresponding to the pixel pair, according to the first left light compensation model corresponding to the pixel pair, and the second left light compensation model initial pixel pair weighted operation is obtained; wherein the second left light compensation model initial pixel pair, according to the position information of the first left light compensation model corresponding to the linear operation of the pixel pair is obtained.

11. The inter prediction method of claim 6, wherein: the second left light compensation model corresponding to the pixel pair, according to the position information of the first left light compensation model corresponding to the linear operation of the pixel pair is obtained; or the second left light compensation model corresponding to the pixel pair, according to the first left light compensation model corresponding to the pixel pair, and the second left light compensation model initial pixel pair weighted operation is obtained; wherein the second left light compensation model initial pixel pair, according to the position information of the first left light compensation model corresponding to the linear operation of the pixel pair is obtained.

12. The inter prediction method of claim 1, wherein, The code stream is acquired and parsed, and the illumination compensation index, motion information and residual information of the current decoding unit are obtained. The code stream is acquired and parsed, and the illumination compensation index, motion information and residual information of the current decoding unit are obtained. In the case of the inter prediction value correction identifier being turned on, the code stream is parsed to obtain the illumination compensation index.

13. The inter prediction method of claim 1, wherein, The prediction block of the current decoding unit includes a chrominance component and a luminance component; The target illumination compensation information corresponding to the illumination compensation index is obtained. The scaling compensation coefficient and the offset compensation coefficient are obtained according to the target illumination compensation information corresponding to the illumination compensation index. The chrominance component and the luminance component are corrected according to the scaling compensation coefficient and the offset compensation coefficient, respectively, to obtain the corrected chrominance component and the corrected luminance component.

14. An inter prediction method applied to a decoding end, the method comprising: acquiring a code stream and parsing the code stream to obtain an illumination compensation index, motion information and residual information of a current decoding unit; performing motion compensation on the current decoding unit according to the motion information to obtain a prediction block of the current decoding unit; in the case that neither an upper neighboring block nor a left neighboring block of the current decoding unit is available, a reconstructed block of the current decoding unit is obtained according to the prediction block of the current decoding unit and the residual information.

15. An inter prediction method applied to an encoding end, the method comprising: performing motion compensation on a current encoding unit according to motion information to obtain a prediction block of the current encoding unit; performing rate-distortion optimization-based correction on the prediction block of the current encoding unit according to an illumination compensation model; determining target illumination compensation information according to a rate-distortion optimization cost; writing an illumination compensation index corresponding to the target illumination compensation information into a code stream; wherein the target illumination compensation information is one of the illumination compensation models; the illumination compensation models include a first upper-left illumination compensation model, a first upper illumination compensation model, a first left illumination compensation model, and at least one of the following: a second top-left illumination compensation model; a second top illumination compensation model; a second left illumination compensation model.

16. The inter prediction method of claim 15, wherein: the first top-left illumination compensation model and the second top-left illumination compensation model are different; the first top illumination compensation model and the second top illumination compensation model are different; the first left illumination compensation model and the second left illumination compensation model are different.

17. The inter prediction method of claim 16, wherein, the illumination compensation model at least meets one of the following: the pixel pair corresponding to the first top-left illumination compensation model and the pixel pair corresponding to the second top-left illumination compensation model are different; the pixel pair corresponding to the first top illumination compensation model and the pixel pair corresponding to the second top illumination compensation model are different; the pixel pair corresponding to the first left illumination compensation model and the pixel pair corresponding to the second left illumination compensation model are different; wherein the pixel pair is obtained according to reconstructed pixels of the current coding unit and reconstructed pixels of a reference unit matched with the current coding unit.

18. The inter prediction method of claim 17, wherein, the illumination compensation model at least meets one of the following: the reconstructed pixels of the current coding unit corresponding to the first top-left illumination compensation model and the reconstructed pixels of the current coding unit corresponding to the second top-left illumination compensation model are same, and the reconstructed pixels of a reference unit matched with the current coding unit corresponding to the first top-left illumination compensation model and the reconstructed pixels of the reference unit matched with the current coding unit corresponding to the second top-left illumination compensation model are same; the reconstructed pixels of the current coding unit corresponding to the first top illumination compensation model and the reconstructed pixels of the current coding unit corresponding to the second top illumination compensation model are same, and the reconstructed pixels of a reference unit matched with the current coding unit corresponding to the first top illumination compensation model and the reconstructed pixels of the reference unit matched with the current coding unit corresponding to the second top illumination compensation model are same; the reconstructed pixels of the current coding unit corresponding to the first left illumination compensation model and the reconstructed pixels of the current coding unit corresponding to the second left illumination compensation model are same, and the reconstructed pixels of a reference unit matched with the current coding unit corresponding to the first left illumination compensation model and the reconstructed pixels of the reference unit matched with the current coding unit corresponding to the second left illumination compensation model are same.

19. The inter prediction method of claim 17 or 18, wherein, the illumination compensation model at least meets one of the following: the pixel pair corresponding to the illumination compensation model is obtained according to a width of the current coding unit or a width of a reference unit matched with the current coding unit; the pixel pair corresponding to the illumination compensation model is obtained according to a height of the current coding unit or a height of a reference unit matched with the current coding unit.

20. The inter prediction method of claim 17 or 18, wherein, the illumination compensation model at least meets one of the following: the pixel pair corresponding to the second top-left illumination compensation model is obtained according to position information of the pixel pair corresponding to the first top-left illumination compensation model; the pixel pair corresponding to the second top illumination compensation model is obtained according to position information of the pixel pair corresponding to the first top illumination compensation model; the pixel pair corresponding to the second left illumination compensation model is obtained according to position information of the pixel pair corresponding to the first left illumination compensation model.

21. The inter prediction method of claim 20, wherein: the second upper-left illumination compensation model corresponds to a pair of pixel points, which includes at least one of: a pair of row pixel points corresponding to the second upper-left illumination compensation model; a pair of column pixel points corresponding to the second upper-left illumination compensation model. the first upper-left illumination compensation model corresponds to a pair of pixel points, which includes at least one of: a pair of row pixel points corresponding to the first upper-left illumination compensation model; a pair of column pixel points corresponding to the first upper-left illumination compensation model.

22. The inter prediction method of claim 21, wherein: the pair of row pixel points corresponding to the second upper-left illumination compensation model is obtained by linear operation according to the position information of the pair of row pixel points corresponding to the first upper-left illumination compensation model; or the pair of column pixel points corresponding to the second upper-left illumination compensation model is obtained by linear operation according to the position information of the pair of column pixel points corresponding to the first upper-left illumination compensation model.

23. The inter prediction method of claim 21, wherein: the pair of row pixel points corresponding to the second upper-left illumination compensation model is obtained by weighted operation according to the pair of row pixel points corresponding to the first upper-left illumination compensation model and the initial pair of row pixel points of the second upper-left illumination compensation model; or the pair of column pixel points corresponding to the second upper-left illumination compensation model is obtained by weighted operation according to the pair of column pixel points corresponding to the first upper-left illumination compensation model and the initial pair of column pixel points of the second upper-left illumination compensation model; wherein the initial pair of row pixel points of the second upper-left illumination compensation model is obtained by linear operation according to the position information of the pair of row pixel points corresponding to the first upper-left illumination compensation model; the initial pair of column pixel points of the second upper-left illumination compensation model is obtained by linear operation according to the position information of the pair of column pixel points corresponding to the first upper-left illumination compensation model.

24. The inter prediction method of claim 20, wherein: the pair of pixel points corresponding to the second upper illumination compensation model is obtained by linear operation according to the position information of the pair of pixel points corresponding to the first upper illumination compensation model; or the pair of pixel points corresponding to the second upper illumination compensation model is obtained by weighted operation according to the pair of pixel points corresponding to the first upper illumination compensation model and the initial pair of pixel points of the second upper illumination compensation model, wherein the initial pair of pixel points of the second upper illumination compensation model is obtained by linear operation according to the position information of the pair of row pixel points corresponding to the first upper-left illumination compensation model.

25. The inter prediction method of claim 20, wherein: the pair of pixel points corresponding to the second left illumination compensation model is obtained by linear operation according to the position information of the pair of pixel points corresponding to the first left illumination compensation model; or the pair of pixel points corresponding to the second left illumination compensation model is obtained by weighted operation according to the pair of pixel points corresponding to the first left illumination compensation model and the initial pair of pixel points of the second left illumination compensation model, wherein the initial pair of pixel points of the second left illumination compensation model is obtained by linear operation according to the position information of the pair of pixel points corresponding to the first left illumination compensation model.

26. An electronic device, comprising: at least one processor; at least one memory configured to store at least one program; The method according to any one of claims 1 to 25 is implemented when at least one of the programs is executed by at least one of the processors.

27. A computer readable storage medium storing a processor executable program, which, when executed by a processor, is arranged to implement the method according to any one of claims 1 to 25.

28. A computer program product comprising computer programs or computer instructions, wherein, The computer program or the computer instructions are stored in a computer readable storage medium, and the processor of the computer device reads the computer program or the computer instructions from the computer readable storage medium, and the processor executes the computer program or the computer instructions, so that the computer device executes the method according to any one of claims 1 to 25.

Images