Inter-frame prediction method and device, electronic equipment and computer program product

Abstract

The invention provides an inter-frame prediction method and device, electronic equipment and a computer program product, and relates to the technical field of video coding and decoding. The method comprises the steps that candidate motion information of a current to-be-predicted block is acquired, a reference block corresponding to the current to-be-predicted block is determined based on the candidate motion information, and the candidate motion information comprises an initial identification value corresponding to a local illumination compensation identification; determining a target template of the current block to be predicted, a reference template corresponding to the reference block, and template distortion between the target template and the reference template; performing identification updating operation on the initial identification value according to the template distortion to obtain an updated identification value; and performing inter-frame prediction processing on the current to-be-predicted block according to the update identification value to obtain a predicted value of the current to-be-predicted block. According to the implicit LIC identifier exporting method provided by the invention, on the premise of not additionally increasing the identifier cost, a more accurate LIC opening state is decided, and the prediction error caused by directly inheriting the LIC identifier is reduced.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of video coding, and particularly relates to an inter prediction method, an inter prediction device, an electronic device and a computer program product. BACKGROUND

[0002] As a core module in the hybrid video coding framework, inter prediction plays a very important role in most video coding systems. Based on the temporal correlation of video sequences, inter prediction is to use the information in the already coded video frames to predict the unknown information in the current frame, and the already coded frame is called the reference frame of the current frame, so as to effectively remove the temporal redundancy information in the video.

[0003] The inter prediction process in video coding can be roughly summarized as three modules: (1) motion information estimation. According to a motion prediction model, the corresponding position of the current block to be predicted in the reference frame is located, that is, an estimated motion of the current block relative to the reference frame is made, and finally the reference block in one or more reference frames is obtained, that is, the prediction block; (2) motion vector coding. After the estimation of the first step, an estimated motion vector is obtained, and the motion vector (Motion Vector, MV) needs to be identified into the code stream and transmitted to the decoding end; (3) motion compensation. The motion information is decoded at the decoding end, and the motion is aligned according to the predefined motion prediction model.

[0004] A large amount of motion information is needed to complete the inter prediction process. In order to derive these motion information, two main prediction modes are designed in the standard: the ordinary inter prediction mode and the inter merge prediction (Merge) mode. The difference between the two modes is that the ordinary inter prediction mode requires the coding block to write the motion information (motion vector, prediction direction, reference frame index, etc.) of the block into the code stream; and the inter merge prediction mode allows the coding block to reuse the motion information of the surrounding already coded blocks, so it is not necessary to write the motion information into the code stream, and only a merge index value (merge index) needs to be coded, which greatly saves the code word cost of coding motion information and improves the inter prediction efficiency. SUMMARY

[0005] The present disclosure provides an inter prediction method, device, electronic device, computer readable storage medium and computer program product to at least solve the problems in the related art that the rate-distortion cost is increased due to the additional code word cost recording the LIC identifier in the inter prediction process, or the accuracy of the prediction block is reduced due to the LIC identifier directly inheriting the adjacent motion information. The technical solutions of the present disclosure are as follows:

[0006] According to a first aspect of the embodiments of the present disclosure, a method for inter prediction is provided, including: obtaining candidate motion information of a current block to be predicted, determining a reference block corresponding to the current block to be predicted based on the candidate motion information, wherein the candidate motion information includes an initial identification value corresponding to a local illumination compensation identification; determining a target template of the current block to be predicted and a reference template corresponding to the reference block, and a template distortion between the target template and the reference template; performing identification update operation on the initial identification value according to the template distortion to obtain an updated identification value; and performing inter prediction processing on the current block to be predicted according to the updated identification value to obtain a prediction value of the current block to be predicted.

[0007] In an example embodiment of the present disclosure, the determining the target template of the current block to be predicted and the reference template corresponding to the reference block includes: constructing the target template based on a specified number of rows and columns of first pixels adjacent to the current block to be predicted; determining reference block position information of the reference block according to the candidate motion information; determining a specified number of rows and columns of second pixels adjacent to the reference block based on the reference block position information; and constructing the reference template based on the second pixels.

[0008] In an example embodiment of the present disclosure, the performing identification update operation on the initial identification value according to the template distortion to obtain an updated identification value includes: determining a non-illumination model distortion value corresponding to a non-illumination model distortion and an illumination model distortion value corresponding to an illumination model distortion according to the initial identification value and the template distortion; and performing identification update operation on the initial identification value according to a comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identification value.

[0009] In an example embodiment of the present disclosure, the template distortion includes a first template distortion value and a second template distortion value; and the determining a non-illumination model distortion value corresponding to a non-illumination model distortion and an illumination model distortion value corresponding to an illumination model distortion according to the initial identification value and the template distortion includes: when the initial identification value is a first illumination compensation identification value, determining a first distortion value corresponding to the non-illumination model distortion according to the first template distortion value, and determining a second distortion value corresponding to the illumination model distortion according to the second template distortion value; and when the initial identification value is a second illumination compensation identification value, determining a third distortion value corresponding to the non-illumination model distortion according to the first template distortion value, and determining a fourth distortion value corresponding to the illumination model distortion according to the second template distortion value.

[0010] In an example embodiment of the present disclosure, the updated identification value includes a first updated identification value, and the identification updating operation on the initial identification value according to the comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identification value includes: when the non-illumination model distortion value is greater than the illumination model distortion value, performing the identification updating operation on the initial identification value to take the second illumination compensation identification value as the first updated identification value; and when the non-illumination model distortion value is less than or equal to the illumination model distortion value, performing the identification updating operation on the initial identification value to take the first illumination compensation identification value as the first updated identification value.

[0011] In an example embodiment of the present disclosure, the updated identification value further includes a second updated identification value, and the method further includes: when the first updated identification value is the first illumination compensation identification value, performing a second updating operation on the updated identification value to obtain the second updated identification value.

[0012] In an example embodiment of the present disclosure, the second updating operation on the updated identification value to obtain the second updated identification value includes: performing splitting processing on the target template to obtain a first target splitting template and a second target splitting template; performing splitting processing on the reference template to obtain a first reference splitting template and a second reference splitting template; determining an illumination compensation parameter according to the first reference splitting template; performing brightness illumination compensation on the second reference splitting template based on the illumination compensation parameter to obtain a second reference illumination template; and determining the second updated identification value according to a splitting template distortion comparison result between the second reference illumination template, the second reference splitting template and the second target splitting template.

[0013] In an example embodiment of the present disclosure, the brightness illumination compensation on the second reference splitting template based on the illumination compensation parameter to obtain a second reference illumination template includes: determining an illumination reference pixel point based on the second reference splitting template, determining an illumination compensation identification value corresponding to the illumination reference pixel point; when the illumination compensation identification value is a first illumination compensation identification value, taking a pixel value of the illumination reference pixel point as a target illumination reference pixel value; and when the illumination compensation identification value is a second illumination compensation identification value, performing brightness illumination compensation on the illumination reference pixel point based on the illumination compensation parameter, and taking an obtained illumination compensation pixel value as the target illumination reference pixel value.

[0014] In an example embodiment of the present disclosure, the determining the second update identification value according to the split template distortion comparison result between the second reference split template, the second reference illumination template and the second target split template comprises: taking the split template distortion between the second reference split template and the second target split template as the non-illumination model distortion value; taking the split template distortion between the second reference illumination template and the second target split template as the illumination model distortion value; when the non-illumination model distortion value is greater than the illumination model distortion value, determining the second update identification value as the second illumination compensation identification value; when the non-illumination model distortion value is less than or equal to the illumination model distortion value, determining the second update identification value as the first illumination compensation identification value.

[0015] In an example embodiment of the present disclosure, the inter-prediction processing of the current to-be-predicted block according to the update identification value comprises: when the update identification value is in an open state, performing brightness illumination compensation processing on the current to-be-predicted block to obtain the prediction value of the current to-be-predicted block; when the update identification value is in a non-open state, determining a motion vector based on the candidate motion information; determining an interpolation prediction block based on the motion vector, and taking the interpolation prediction block as the prediction value of the current to-be-predicted block.

[0016] According to a second aspect of the embodiments of the present disclosure, an inter-prediction device is provided, which comprises: a reference block determination module configured to acquire candidate motion information of a current to-be-predicted block, and determine a reference block corresponding to the current to-be-predicted block based on the candidate motion information, wherein the candidate motion information comprises an initial identification value corresponding to a local illumination compensation identification; a template distortion determination module configured to determine a target template of the current to-be-predicted block and a reference template corresponding to the reference block, and determine a template distortion between the target template and the reference template; an identification update module configured to perform an identification update operation on the initial identification value according to the template distortion to obtain an update identification value; and an inter-prediction module configured to perform inter-prediction processing on the current to-be-predicted block according to the update identification value to obtain a prediction value of the current to-be-predicted block.

[0017] In an example embodiment of the present disclosure, the template distortion determination module comprises a template determination unit configured to: construct the target template based on a specified number of rows and columns of first pixels adjacent to the current to-be-predicted block; determine reference block position information corresponding to the reference block according to the candidate motion information; determine a specified number of rows and columns of second pixels adjacent to the reference block based on the reference block position information; and construct the reference template based on the second pixels.

[0018] In an example embodiment of the present disclosure, the identity updating module comprises a first identity updating unit configured to: determine a non-illumination model distortion value corresponding to the non-illumination model distortion and an illumination model distortion value corresponding to the illumination model distortion according to the initial identity value and the template distortion; and perform an identity updating operation on the initial identity value according to a comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identity value.

[0019] In an example embodiment of the present disclosure, the template distortion comprises a first template distortion value and a second template distortion value; and the first identity updating unit comprises a distortion value determining sub-unit configured to: when the initial identity value is a first illumination compensation identity value, determine a first distortion value corresponding to the non-illumination model distortion according to the first template distortion value and a second distortion value corresponding to the illumination model distortion according to the second template distortion value; and when the initial identity value is a second illumination compensation identity value, determine a third distortion value corresponding to the non-illumination model distortion according to the first template distortion value and a fourth distortion value corresponding to the illumination model distortion according to the second template distortion value.

[0020] In an example embodiment of the present disclosure, the updated identity value comprises a first updated identity value; the first identity updating unit comprises a first identity updating sub-unit configured to: when the non-illumination model distortion value is greater than the illumination model distortion value, perform an identity updating operation on the initial identity value to obtain the first updated identity value as the second illumination compensation identity value; and when the non-illumination model distortion value is less than or equal to the illumination model distortion value, perform an identity updating operation on the initial identity value to obtain the first updated identity value as the first illumination compensation identity value.

[0021] In an example embodiment of the present disclosure, the updated identity value further comprises a second updated identity value; and the identity updating module further comprises a second identity updating unit configured to: when the first updated identity value is the first illumination compensation identity value, perform a secondary updating operation on the updated identity value to obtain the second updated identity value.

[0022] In an example embodiment of the present disclosure, the second identity updating unit comprises a second identity updating subunit, configured to: split the target template to obtain a first target split template and a second target split template; split the reference template to obtain a first reference split template and a second reference split template; determine an illumination compensation parameter according to the first reference split template; perform brightness illumination compensation on the second reference split template based on the illumination compensation parameter to obtain a second reference illumination template; and determine the second update identity value according to a split template distortion comparison result between the second reference illumination template, the second reference split template and the second target split template.

[0023] In an example embodiment of the present disclosure, the second identity updating subunit comprises an illumination compensation subunit, configured to: determine an illumination reference pixel point based on the second reference split template, and determine an illumination compensation identity value corresponding to the illumination reference pixel point; when the illumination compensation identity value is a first illumination compensation identity value, take a pixel value of the illumination reference pixel point as a target illumination reference pixel value; and when the illumination compensation identity value is a second illumination compensation identity value, perform brightness illumination compensation on the illumination reference pixel point based on the illumination compensation parameter, and take a resulting illumination compensation pixel value as the target illumination reference pixel value.

[0024] In an example embodiment of the present disclosure, the second identity updating subunit is configured to perform: taking a split template distortion between the second reference split template and the second target split template as the non-illumination model distortion value; taking a split template distortion between the second reference illumination template and the second target split template as the illumination model distortion value; when the non-illumination model distortion value is greater than the illumination model distortion value, determining the second update identity value as the second illumination compensation identity value; and when the non-illumination model distortion value is less than or equal to the illumination model distortion value, determining the second update identity value as the first illumination compensation identity value.

[0025] In an example embodiment of the present disclosure, the inter prediction module comprises an inter prediction unit, configured to: when the update identity value is in an open state, perform brightness illumination compensation processing on the current prediction block to obtain a prediction value of the current to-be-predicted block; when the update identity value is in a non-open state, determine a motion vector based on the candidate motion information; determine an interpolation prediction block based on the motion vector, and take the interpolation prediction block as the prediction value of the current to-be-predicted block.

[0026] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the inter prediction method according to any one of the preceding aspects.

[0027] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the inter prediction method according to any one of the preceding aspects.

[0028] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the inter prediction method according to any one of the preceding aspects.

[0029] The technical solutions provided by the embodiments of the present disclosure at least have the following beneficial effects:

[0030] On the one hand, by providing an implicit LIC flag derivation method, a more accurate LIC on state is determined for the inter prediction mode decision by simulating the prediction process of LIC on and off without additional flag cost. On the other hand, by updating the LIC flag value, the prediction error caused by directly inheriting the LIC flag is reduced, which helps to improve the video coding performance.

[0031] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0032] The accompanying drawings incorporated in the specification and forming a part of the specification illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure, and do not limit the present disclosure.

[0033] Figure 1 is a flowchart of an inter prediction method according to an exemplary embodiment.

[0034] Figure 2 is a schematic diagram of a spatial candidate list establishment method according to an exemplary embodiment.

[0035] Figure 3 is a schematic diagram of a temporal candidate list construction method according to an exemplary embodiment.

[0036] Figure 4 is a schematic diagram of determining a temporal candidate position according to an exemplary embodiment.

[0037] Figure 5 is a schematic diagram of a reference block and motion information of a current frame in a reference frame in inter prediction according to an exemplary embodiment.

[0038] Figure 6 is a schematic diagram of a target template and a reference template of a current block to be predicted in inter prediction according to an exemplary embodiment.

[0039] Figure 7 is a schematic diagram of a template matching based implicit LIC identification derivation process according to an example embodiment.

[0040] Figure 8 is a schematic diagram of a local illumination compensation prediction mode according to an example embodiment.

[0041] Figure 9 is a block diagram of an inter prediction apparatus according to an example embodiment.

[0042] Figure 10 a block diagram of an electronic device according to an example embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0043] In order for those skilled in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings.

[0044] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described herein can be implemented in an order other than that illustrated or described herein. The implementation described in the following example embodiments does not represent all implementations consistent with the present disclosure. Rather, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0045] In a related scheme, Merge is a special inter prediction mode, and the encoding and decoding ends use the same way to construct a motion information candidate list. At the encoding end, the optimal motion information index is selected by rate-distortion criteria, and only the index needs to be transmitted to the decoding end, without the need for encoding reference frame index, motion vector difference (MVD) and other motion information.

[0046] In the Merge mode of Versatile Video Coding (VVC), motion information candidate lists are constructed using temporal or spatial neighboring blocks, and there are only 6 candidate motion information in the list. The candidate motion information has 5 types, in the order as follows: 1) spatial candidate based on neighboring blocks; 2) temporal candidate based on co-located blocks; 3) candidate based on the FIFO table constructed by historical information; 4) paired average candidate; 5) zero candidate.

[0047] A block-based inter prediction needs multiple motion information to work together. Based on the above-mentioned Merge derivation process, these information can be derived for each block to be coded for performing inter prediction process. The motion information derived by Merge mode can be summarized as follows: inter prediction direction (ID), motion vector, reference index (RI), bi-prediction with CU-level weight index (BCW index), half-pel interpolation filter (HF) and local illumination compensation flag (LIC flag) and so on.

[0048] The advantage of Merge is that it can derive the motion information needed for motion compensation at a very low cost of code words. However, compared with the general inter prediction process, Merge mode lacks the process of traversing and comparing multiple prediction modes, so the derived motion information is not necessarily the most accurate. When the motion information derived based on Merge deviates greatly from the real optimal motion information, it will result in low efficiency of Merge prediction, and ultimately the general inter prediction mode will be selected.

[0049] In another related scheme, after deriving the motion information, the prediction block position is determined on the corresponding reference frame according to the indication of the motion information, and the final prediction pixel is obtained through interpolation. The conventional inter prediction method is to directly take the interpolated pixel as the prediction pixel. In order to make up for the pixel difference caused by illumination change in the motion process, a prediction technology based on local illumination compensation (LIC) is proposed. The core principle is to establish a local illumination compensation prediction model, which contains a scaling coefficient and an offset coefficient. The illumination compensation parameters are solved by using the neighboring pixels of the to-be-predicted block and the corresponding neighboring pixels of the prediction block in the reference frame, and finally used for the brightness illumination compensation of the prediction value of the current block.

[0050] As an inter prediction mode, LIC also performs the conventional motion vector derivation and inter motion compensation process first. The difference is that LIC further corrects the pixels of the interpolated prediction block based on the brightness illumination compensation model. In the existing standard, if it is a general inter prediction mode, an additional bit is needed to represent whether LIC mode is used for pixel correction. If it is a Merge mode, the LIC flag in the derived motion information is directly inherited.

[0051] In the common inter prediction mode, the LIC identifies the decision of rate-distortion optimization (RDO), has higher accuracy, but needs extra code word cost, increases the rate-distortion cost; in the Merge mode, although the LIC identification of the neighboring motion information can be directly inherited to omit the code word, the current block can have different illumination characteristics from the neighboring block, at this time, the LIC state directly inherited is not accurate, and the illumination compensation correction on the prediction block is reduced. Therefore, the accurate decision of the LIC state is the key to improve the LIC prediction efficiency.

[0052] Based on this, according to the embodiments of the present disclosure, an inter prediction method, an inter prediction device, an electronic device, a computer readable storage medium and a computer program product are provided.

[0053] The principles and spirits of the present disclosure are explained in detail below with reference to several representative embodiments of the present disclosure.

[0054] Figure 1 is a flow chart of an inter prediction method according to an example embodiment, as Figure 1 As shown, the inter prediction method can be used in a computer device, wherein the computer device described in the present disclosure can include mobile terminal devices such as mobile phones, tablet computers, notebook computers, palm computers, personal digital assistants (PDAs) and the like, and fixed terminal devices such as desktop computers and the like. The present example embodiment is exemplified by the method applied to the computer device, and it can be understood that the method can also be applied to a server, and can also be applied to a system including the computer device and the server, and is realized through the interaction of the computer device and the server. Specifically, the following steps are included.

[0055] Step S110, acquiring candidate motion information of a current to-be-predicted block, determining a reference block corresponding to the current to-be-predicted block based on the candidate motion information, the candidate motion information including an initial identification value corresponding to a local illumination compensation identification;

[0056] Step S120, determining a target template of the current to-be-predicted block and a reference template corresponding to the reference block, and a template distortion between the target template and the reference template;

[0057] Step S130, performing identification update operation on the initial identification value according to the template distortion, to obtain an updated identification value;

[0058] Step S140, performing inter prediction processing on the current to-be-predicted block according to the updated identification value, to obtain a prediction value of the current to-be-predicted block.

[0059] According to the inter prediction method in the example embodiment, on one hand, by providing an implicit LIC flag derivation method, a more accurate LIC on / off state is determined for the inter prediction mode fusion by simulating the prediction process of LIC on and off without additional identification cost. On the other hand, by updating the LIC flag value, the prediction error caused by directly inheriting the LIC flag is reduced, which helps to improve the video coding performance.

[0060] In the following, the inter prediction method in the example embodiment will be further described.

[0061] In an example embodiment of the disclosure, for step S110, candidate motion information of the current to-be-predicted block is obtained, and a reference block corresponding to the current to-be-predicted block is determined based on the candidate motion information, wherein the candidate motion information includes an initial flag value corresponding to a local illumination compensation flag.

[0062] The current to-be-predicted block can be an image block being processed by the video coding terminal, which is the block in the "encoding" state when the video frame is divided into a plurality of small blocks (such as coding units, prediction units, etc.) and processed one by one in a certain order. The candidate motion information can be the reference motion information used to determine the prediction value of the current to-be-predicted block. The reference block can be a reference area used to predict the pixel value of the current to-be-predicted block. The local illumination compensation flag (LIC flag) can be a flag used to indicate whether the local illumination compensation is used for the current to-be-predicted block; the flag takes a value of 0 or 1, 0 indicating not to use and 1 indicating to use. The local illumination compensation flag can be represented as LIC hereinafter. The initial flag value can be a specific value of the local illumination compensation flag determined based on the candidate motion information.

[0063] In the Merge mode, the candidate motion information mainly includes five types. For the spatial candidate list, the reference Figure 2 , Figure 2 is a schematic diagram of a spatial candidate list establishment method according to an example embodiment. As Figure 2 indicated, A1 is a neighboring CU at the lower left of the current CU, B1 is a neighboring CU at the upper right of the current CU, B2 is a neighboring CU at the upper left of the current CU, and A0 and B0 are the nearest CUs to A1 and B1, respectively. In VVC, a maximum of 4 candidates are provided in the spatial domain, i.e., a maximum of 4 candidate blocks in the 5 candidate blocks are used for motion information, and the candidate list is established in the order of B1->A1->B0->A0->(B2), wherein B2 is a substitute, and B2 is added to the candidate list only when one or more of the first 4 candidates do not exist.

[0064] For the temporal candidate list, the temporal dimension provides at most one candidate motion information. Unlike the spatial construction, the temporal candidate list can directly use all the motion information except the MV, which needs to be scaled according to the position relationship. Referring to Figure 3 , Figure 3 is a schematic diagram of the construction of the temporal candidate list according to an exemplary embodiment. Figure 3 In Td, the distance between the current picture and the reference picture is measured by POC, and Tb is the distance between the collocated picture and its reference picture.

[0065] Referring to Figure 4 , Figure 4 is a schematic diagram of determining the position of the temporal candidate according to an exemplary embodiment. Taking the current CU shown in Figure 4 as an example, the temporal candidate motion information is obtained based on the following C0 and C1 positions. As shown in Figure 4 , if C0 is not available, the collocated CU of C1 is used instead, where Y represents the collocated CU (col_CU) of the current to-be-predicted block (current CU, curr_CU) on the temporal reference frame.

[0066] Multiple motion information needs to be coordinated for a block to perform inter prediction. Based on the inter prediction Merge mode, candidate motion information can be derived for each to-be-predicted block for performing the inter prediction process. The candidate motion information derived by the Merge mode can include, but is not limited to, the inter prediction direction, the motion vector, the reference index, the bi-prediction weight prediction index, the half-pixel interpolation index, and the local illumination compensation flag, etc. The details are as follows:

[0067] The inter prediction direction is used to identify whether the current block is forward predicted (ID = 1), backward predicted (ID = 2), or bi-predicted (ID = 3), so the value range is [1, 3]. The motion vector represents the offset of the reference block relative to the current block. The reference index is used to identify which reference picture is referred to by the current block, and the value of RI depends on the number of reference frames. When the number of reference frames is M (M is a natural number greater than 0, and the value of RI is [0, M-1]); and when ID = 3, there are two RI values, which are respectively used to represent the reference frame index of the forward prediction and the backward prediction.

[0068] The bidirectional weight prediction index is related to an algorithm BCW, which is an algorithm for adjusting the forward and backward prediction weights of the inter bidirectional prediction CU, so the BCW index only takes effect when ID=3. The original default forward and backward prediction weight of the inter bidirectional prediction is 1 / 2, and the algorithm can provide multiple unbalanced weight for the CU to select. For low-delay pictures (all the POCs of the reference frames are less than the POC of the current frame), there are currently 5 candidates for selection, {-2 / 8, 3 / 8, 4 / 8, 5 / 8, 10 / 8}, among which 4 / 8 is the original 1 / 2 weight, and the value of the BCW index is [0, 4] at this time; and for non-low-delay pictures (there are reference frames whose POCs are greater than the POC of the current frame), there are currently 3 candidates for selection: {3 / 8, 4 / 8, 5 / 8}, and the value of the BCW index is [0, 2] at this time, corresponding to the 3 weights above.

[0069] The half-pixel interpolation index is used to indicate whether a half-pixel interpolation filter is used in the stage of generating a prediction block by motion compensation for the current block, and HF takes a value of [0, 1], HF=1 indicates that the half-pixel interpolation filter is used, otherwise HF=0 indicates that the half-pixel interpolation filter is not used. Based on the above candidate motion information, the position of the prediction block on the reference frame, i.e. the position of the reference block, can be determined. Figure 5 , Figure 5 A schematic diagram of a reference block and motion information of a current frame in inter prediction is shown according to an exemplary embodiment. Figure 5 The position information of the reference block of the current to-be-predicted frame in the reference frame 0 and the reference frame 1, the MVD information, and the motion vector prediction (MVP) information are shown in the figure.

[0070] In an exemplary embodiment of the present disclosure, for step S120, the target template of the current to-be-predicted block and the reference template corresponding to the reference block are determined, including: constructing the target template based on a specified number of rows and columns of first pixels adjacent to the current to-be-predicted block; determining the reference block position information corresponding to the reference block according to the candidate motion information; determining a specified number of rows and columns of second pixels adjacent to the reference block based on the reference block position information; and constructing the reference template based on the second pixels.

[0071] The first pixels can be the pixels used to construct the target template of the current to-be-predicted block. The reference block position information can be the position information corresponding to the reference block. The second pixels can be the pixels used to construct the reference template of the current to-be-predicted block. The reference template can be a pixel template located in the reference frame.

[0072] For the current to-be-predicted block, the n rows and n columns of pixels adjacent to the current to-be-predicted block can be defined as the first pixels, and the target template T is constructed based on the first pixelst n can be defined according to the coding requirement, for example, n can be 4, 6, 8, etc. Similarly, n rows and n columns of pixels around the prediction block in the reference frame can be defined as the second pixels, and the reference template T c is constructed based on the second pixels c .

[0073] Reference Figure 6 , Figure 6 is a schematic diagram of a target template of a current block to be predicted and a reference template in inter prediction according to an example embodiment. Figure 6 n = 4, the target template T t and the reference template T c are constructed. When the current block to be predicted is encoded, T t and T c are known information after encoding, so the LIC identifier can be derived based on the relationship between the two.

[0074] In an example embodiment of the present disclosure, for step S130, the initial identifier value is updated according to the template distortion to obtain an updated identifier value, including: determining a non-illumination model distortion value corresponding to the non-illumination model distortion and an illumination model distortion value corresponding to the illumination model distortion according to the initial identifier value and the template distortion; and performing an identifier update operation on the initial identifier value according to a comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identifier value.

[0075] The template distortion can be a distortion calculated based on the difference between the target template and the reference template. The non-illumination model distortion can be a distortion used in the inter prediction process without using the LIC model for prediction, which can be marked as D non-LIC . The non-illumination model distortion value can be a specific value of the non-illumination model distortion. The illumination model distortion can be a distortion used in the inter prediction process using the LIC model for prediction, which can be marked as D LIC . The illumination model distortion value can be a specific value of the illumination model distortion. The updated identifier value can be the LIC identifier value obtained after the identifier update operation.

[0076] Based on the candidate motion information, the initial identifier value of the LIC identifier can be obtained, and for the template distortion between the target template and the reference template, T t and T cthe sum of absolute difference (SAD) distortion and the mean-removed sum of absolute difference (MRSAD) distortion, denoted as D SAD and D MRSAD .

[0077] After the initial identification value and the template distortion are determined, the non-illumination model distortion value and the illumination model distortion value are determined according to the initial identification value and the template distortion. Then, the numerical comparison result between the two values is used to determine whether the initial identification value needs to be updated, and an updated identification value is obtained. The comparison result of the template distortion is used in the identification update operation, which can improve the accuracy of the identification decision.

[0078] In an exemplary embodiment of the present disclosure, the non-illumination model distortion value corresponding to the non-illumination model distortion and the illumination model distortion value corresponding to the illumination model distortion are determined according to the initial identification value and the template distortion, including: when the initial identification value is a first illumination compensation identification value, determining a first distortion value corresponding to the non-illumination model distortion according to a first template distortion value, and determining a second distortion value corresponding to the illumination model distortion according to a second template distortion value; when the initial identification value is a second illumination compensation identification value, determining a third distortion value corresponding to the non-illumination model distortion according to the first template distortion value, and determining a fourth distortion value corresponding to the illumination model distortion according to the second template distortion value.

[0079] The first illumination compensation identification value and the second illumination compensation identification value can be two different values of LIC, which can be 0 or 1; the first illumination compensation identification value can be 0, and the second illumination compensation identification value can be 1. The first template distortion value and the second template distortion value are template distortion values calculated by different calculation methods, in the present disclosure, the first template distortion value can be D SAD , and the second template distortion value can be D MRSAD . The first distortion value, the second distortion value, the third distortion value, and the fourth distortion value are specific numerical values corresponding to the non-illumination model distortion and the illumination model distortion under different conditions.

[0080] For the initial identification value obtained based on the candidate motion information, if the LIC is 0, the first distortion value corresponding to the non-illumination model distortion is calculated according to the first template distortion value, that is, D non-LIC =D SAD / 2; and the second distortion value corresponding to the illumination model distortion is determined according to the second template distortion value, that is, D LIC =D MRSAD ; otherwise, when the LIC is 1, the first template distortion value determines the third distortion value corresponding to the non-illumination model distortion, that is, D non-LIC=D SAD ; and determining a fourth distortion value D LIC =D MRSAD / 2. After obtaining the non-illumination model distortion value and the illumination model distortion value, a comparison result of the two values can be used as a basis for an identification update operation.

[0081] In an example embodiment of the present disclosure, the initial identification value is updated according to a comparison result between the non-illumination model distortion value and the illumination model distortion value, to obtain an updated identification value, including: when the non-illumination model distortion value is greater than the illumination model distortion value, the initial identification value is updated, and the second illumination compensation identification value is taken as the first updated identification value; and when the non-illumination model distortion value is less than or equal to the illumination model distortion value, the initial identification value is updated, and the first illumination compensation identification value is taken as the first updated identification value.

[0082] The first updated identification value can be an updated identification value obtained by updating the LIC identification based on the SAD distortion and the MRSAD distortion comparison mode.

[0083] After obtaining D non-LIC and D LIC , the values of D non-LIC and D LIC are compared. When the non-illumination model distortion value is greater than the illumination model distortion value, i.e., D LIC <D non-LIC , the LIC identification is updated to 1, otherwise, when the non-illumination model distortion value is less than or equal to the illumination model distortion value, the LIC identification is updated to 0. Based on the comparison result between the non-illumination model distortion value and the illumination model distortion value, the LIC identification can be updated, and a more accurate LIC decision mechanism is provided.

[0084] In an example embodiment of the present disclosure, when the first updated identification value is the first illumination compensation identification value, the updated identification value is updated again to obtain a second updated identification value. The second updated identification value can be an updated identification value obtained by updating the LIC identification using the template matching mode.

[0085] After the above identification update operation, if the LIC is still the first illumination compensation identification value, i.e., the LIC is still 0, the updated identification value is updated again using the template matching mode to obtain a second updated identification value. The LIC identification result can be further improved in accuracy by updating the LIC identification twice using the template matching mode.

[0086] It should be noted that the template matching method and the MRSAD distortion comparison method are applied in sequence in the process of the identity updating operation in the present disclosure. In some other exemplary embodiments of the present disclosure, only one of the two methods can be used to perform the identity updating operation, and the sequence of the two identity updating operations can be changed and modified. The present disclosure does not make any special limitation on this.

[0087] In an exemplary embodiment of the present disclosure, the second updating of the updating identity value to obtain a second updating identity value comprises: performing splitting processing on the target template to obtain a first target splitting template and a second target splitting template; performing splitting processing on the reference template to obtain a first reference splitting template and a second reference splitting template; determining an illumination compensation parameter according to the first reference splitting template; performing brightness illumination compensation on the second reference splitting template based on the illumination compensation parameter to obtain a second reference illumination template; and determining the second updating identity value according to a splitting template distortion comparison result between the second reference illumination template, the second reference splitting template and the second target splitting template.

[0088] The first target splitting template and the second target splitting template can be templates obtained by splitting the pixels of the target template. The first reference splitting template and the second reference splitting template can be templates obtained by splitting the pixels of the reference template. The illumination compensation parameter can be a parameter used for performing brightness illumination compensation on the pixels in the second reference splitting template. The second reference illumination template can be a template obtained by performing brightness illumination compensation on the pixels in the second reference splitting template using the illumination compensation parameter. The splitting template distortion comparison result can be a comparison result of the pixel distortion between the splitting templates at the corresponding positions of the current frame and the reference frame.

[0089] Continuing to refer to Figure 6 In the present embodiment, the splitting processing is performed on the pixels in T Figure 6 t and T c In the present embodiment, the splitting processing is performed on the pixels in T Figure 7 As shown in FIG. 6B, the splitting processing is performed on the target template to obtain a first target splitting template (pixels in block 610) and a second target splitting template (pixels in block 620). Similarly, the splitting processing is performed on the reference template to obtain a first reference splitting template (pixels in block 630) and a second reference splitting template (pixels in block 640).

[0090] ​Then, the illumination compensation parameters are determined according to the first reference split template. Specifically, a set of illumination compensation parameters (s1, s3) is inferred in a conventional LIC manner using the pixels in block 630. The pixels in the second reference split template are subjected to brightness illumination compensation based on the parameters (s1, s3), and a template after illumination compensation is obtained. Thus, two sets of split templates of the reference blocks are obtained through the above steps. One set is the original template T non-LIC , i.e., the second reference split template; and the other set is the template T LIC' , i.e., the second reference illumination template.

[0091] After the two sets of reference split templates are obtained, the second update identifier is determined according to the split template distortion comparison results between the second reference illumination template and the second target split template, and between the second reference split template and the second target split template. The above steps provide an identifier update operation implemented in a template matching manner, which further improves the accuracy of the LIC identifier.

[0092] In other exemplary embodiments of the present disclosure, other split methods can also be used to split the target template and the reference template, for example, the target template and the reference template are respectively split into two rows and two columns of adjacent pixels and the remaining two rows and two columns of pixels, and the present disclosure does not make any special limitation on the specific split method of the template. In the template splitting process, the split methods of the target template and the reference template need to be consistent.

[0093] In an exemplary embodiment of the present disclosure, the brightness illumination compensation is performed on the second reference split template based on the illumination compensation parameters to obtain the second reference illumination template, including: determining an illumination reference pixel point based on the second reference split template, and determining an illumination compensation identifier value corresponding to the illumination reference pixel point; when the illumination compensation identifier value is a first illumination compensation identifier value, the pixel value of the illumination reference pixel point is taken as a target illumination reference pixel value; and when the illumination compensation identifier value is a second illumination compensation identifier value, the illumination reference pixel point is subjected to brightness illumination compensation based on the illumination compensation parameters, and the obtained illumination compensation pixel value is taken as the target illumination reference pixel value.

[0094] The illumination reference pixel point can be a pixel point contained in the second reference split template. The illumination compensation identifier value can be a specific value of an illumination compensation identifier corresponding to the illumination reference pixel point. The illumination compensation identifier can be an identifier indicating whether a certain illumination reference pixel point adopts a local illumination compensation calculation method of the LIC model. The pixel value of the illumination reference pixel point can be the pixel value of the illumination reference pixel itself. The target illumination reference pixel value can be a pixel value used to constitute the illumination reference pixel point in the second reference illumination template.

[0095] For all illumination reference pixels included in the second reference split template, the following steps are used one by one to perform brightness illumination compensation calculation, specifically including: determining the illumination compensation identification value corresponding to the illumination reference pixel, where the illumination compensation identification value can be 0 or 1. When the illumination compensation identification value is the first illumination compensation identification value (i.e., 0), the pixel value of the illumination reference pixel can be directly used as the target illumination reference pixel value; when the illumination compensation identification value is the second illumination compensation identification value (i.e., 1), the calculated illumination compensation parameters (s1, s3) are used to perform brightness illumination compensation on the illumination reference pixel, and the obtained illumination compensation pixel value is used as the target illumination reference pixel value.

[0096] refer to Figure 7 , Figure 7 FIG is a schematic diagram showing a process of deriving an implicit LIC identifier based on template matching according to an exemplary embodiment. Figure 8 As shown, for any illumination reference pixel in the second reference split template, the LIC identifier of the pixel can be obtained. When LIC = 0, the pixel value of the illumination reference pixel is directly used as the target illumination reference pixel value; when LIC = 1, the LIC model is used to perform brightness illumination compensation on the illumination reference pixel to obtain the target illumination reference pixel value. The specific calculation process is shown in Formulas 1 to 3.

[0097] First, the difference between the target template and the reference template is calculated, and the least squares method is used to calculate the illumination compensation parameter, as shown in Formula 1.

[0098]

[0099] Among them, (s1, s3) can be illumination compensation parameters; It can be the pixel value of the pixel in the target template; It can be the pixel value of the pixel at the corresponding position in the reference template.

[0100] The illumination compensation parameters and the pixels in the reference template are used to perform illumination compensation on the pixels in the target template, as shown in Formula 2.

[0101] P t1 =P rT1 ·s1+s3 (Formula 2)

[0102] Among them, (s1, s3) can be illumination compensation parameters; P T1 It can be the pixel value of each pixel in the target template; P rT1 It can be the pixel value of each pixel in the reference template.

[0103] Next, the illumination compensation parameters are used in the calculation of the pixel values of the illumination reference pixels in the second reference split template based on the parameter multiplexing manner, to obtain a second reference illumination template, as shown in formula 3.

[0104] min(P rT2 ·s1+s3,P rT2 ) (formula 3)

[0105] Wherein, (s1, s3) can be illumination compensation parameters; P rT2 may be the pixel value of each pixel point in the second reference illumination template.

[0106] Through the above steps, the second reference illumination template after the illumination compensation processing can be constructed, so as to perform the identification updating operation based on the template distortion between the second reference illumination template and the second reference split template, and improve the accuracy of the identification value.

[0107] In an exemplary embodiment of the present disclosure, according to the split template distortion comparison result between the second reference illumination template, the second reference split template and the second target split template, the second updated identification value is determined, including: taking the split template distortion between the second reference split template and the second target split template as a non-illumination model distortion value; taking the split template distortion between the second reference illumination template and the second target split template as an illumination model distortion value; when the non-illumination model distortion value is greater than the illumination model distortion value, the second updated identification value is determined as the second illumination compensation identification value; when the non-illumination model distortion value is less than or equal to the illumination model distortion value, the second updated identification value is determined as the first illumination compensation identification value. Wherein, the split template distortion can be the distortion value between the split templates obtained after the split processing.

[0108] The template distortion between the second reference split template T non-LIC and the second reference illumination template T LIC' and the second target split template corresponding to the current block is calculated respectively as the split template distortion. For example, the split template distortion of the above two groups can be calculated by SAD distortion, and are respectively denoted as a non-illumination model distortion value D non-LIC and an illumination model distortion value D LIC If D non-LIC >D LIC , the LIC identification is updated as 1; otherwise, D non-LIC ≤D LIC , the LIC identification is updated as 0. The above steps provide an identification updating scheme based on template matching.

[0109] In an example embodiment of the present disclosure, for step S140, the inter-frame prediction processing is performed on the current to-be-predicted block according to the update identification value to obtain a prediction value of the current to-be-predicted block, including: when the update identification value is in the open state, performing brightness illumination compensation processing on the current prediction block to obtain the prediction value of the current to-be-predicted block; when the update identification value is in the non-open state, determining a motion vector based on the candidate motion information; determining an interpolation prediction block based on the motion vector, and taking the interpolation prediction block as the prediction value of the current to-be-predicted block. The interpolation prediction block can be a prediction block calculated by interpolation processing based on the motion vector.

[0110] For each Merge candidate, after the above identification update operation is performed, the final LIC identification, i.e., the update identification value, is determined, and the update identification value is exported, which can be the first update identification value or the first update identification value. The exported LIC identification is used to replace the original LIC identification of the Merge candidate as the LIC identification of the current to-be-predicted block.

[0111] If the LIC identification is in the open state, i.e., LIC=0, the brightness illumination compensation processing is performed on the current prediction block to obtain the prediction value of the current to-be-predicted block. In order to compensate for the pixel difference caused by illumination change in the motion process, when the LIC processing is performed on the current to-be-predicted block, a brightness illumination compensation prediction model can be established, and the model parameters include a scaling factor a and an offset b, and then the pixel values in the current to-be-predicted block are corrected according to the brightness illumination compensation prediction model to obtain the final prediction value.

[0112] Reference Figure 8 , Figure 9 FIG. 1 is a schematic diagram of a local brightness illumination compensation prediction method according to an example embodiment. The template pixels of the current block and the reference block are obtained; the scaling factor a and the offset b of the linear transformation are determined by using the least square method according to the template pixels; the pixels of the current block are substituted into the linear transformation to perform illumination compensation calculation to obtain the prediction value of the current to-be-predicted block.

[0113] When the update identification value is in the non-open state, i.e., LIC=1, the motion vector is obtained from the candidate motion information, and the interpolation prediction block corresponding to the MV is directly used as the prediction value of the current to-be-predicted block. Through the above steps, the prediction operation on the current to-be-predicted block is realized. By using the inter-frame prediction method of the present disclosure, based on the codec test, 0.3% of the coding performance can be improved under the random access (RA) configuration.

[0114] In summary, the inter prediction method of the present disclosure, the candidate motion information of the current to-be-predicted block is obtained, the reference block corresponding to the current to-be-predicted block is determined based on the candidate motion information, the candidate motion information includes an initial identification value corresponding to a local illumination compensation identification; the target template of the current to-be-predicted block and the reference template corresponding to the reference block, and the template distortion between the target template and the reference template are determined; the initial identification value is updated according to the template distortion to obtain an updated identification value; the current to-be-predicted block is processed according to the updated identification value to obtain the prediction value of the current to-be-predicted block. On the one hand, by providing an implicit LIC identification derivation method, the LIC opening state is more accurately determined for the fusion inter prediction mode decision by simulating the prediction process of LIC opening and closing without additional identification cost. On the other hand, by updating the LIC identification value, the prediction error caused by directly inheriting the LIC identification is reduced, which helps to improve the video coding performance. On the other hand, a plurality of identification update operation schemes are provided, which can further improve the accuracy of the updated identification value decision.

[0115] The present disclosure also provides an inter prediction device, which refers to Figure 9 , Figure 9 is a block diagram of an inter prediction device according to an exemplary embodiment. Referring to Figure 10 , the inter prediction device 900 includes a reference block determination module 910, a template distortion determination module 920, an identification update module 930, and an inter prediction module 940.

[0116] Specifically, the reference block determination module 910 is configured to obtain the candidate motion information of the current to-be-predicted block, and determine the reference block corresponding to the current to-be-predicted block based on the candidate motion information, wherein the candidate motion information includes an initial identification value corresponding to a local illumination compensation identification; the template distortion determination module 920 is configured to determine the target template of the current to-be-predicted block and the reference template corresponding to the reference block, and the template distortion between the target template and the reference template; the identification update module 930 is configured to perform identification update operation on the initial identification value according to the template distortion to obtain an updated identification value; and the inter prediction module 940 is configured to perform inter prediction processing on the current to-be-predicted block according to the updated identification value to obtain the prediction value of the current to-be-predicted block.

[0117] In an exemplary embodiment of the present disclosure, the template distortion determination module 920 includes a template determination unit configured to: construct a target template based on a specified number of rows and columns of first pixels adjacent to the current to-be-predicted block; determine reference block position information corresponding to the reference block according to the candidate motion information; determine a specified number of rows and columns of second pixels adjacent to the reference block based on the reference block position information; and construct a reference template based on the second pixels.

[0118] In an example embodiment of the present disclosure, the identification updating module 930 comprises a first identification updating unit configured to: determine a non-illumination model distortion value corresponding to the non-illumination model distortion and an illumination model distortion value corresponding to the illumination model distortion according to the initial identification value and the template distortion; and perform an identification updating operation on the initial identification value according to a comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain an updated identification value.

[0119] In an example embodiment of the present disclosure, the template distortion comprises a first template distortion value and a second template distortion value; and the first identification updating unit comprises a distortion value determining sub-unit configured to: when the initial identification value is a first illumination compensation identification value, determine a first distortion value corresponding to the non-illumination model distortion according to the first template distortion value and a second distortion value corresponding to the illumination model distortion according to the second template distortion value; and when the initial identification value is a second illumination compensation identification value, determine a third distortion value corresponding to the non-illumination model distortion according to the first template distortion value and a fourth distortion value corresponding to the illumination model distortion according to the second template distortion value.

[0120] In an example embodiment of the present disclosure, the updated identification value comprises a first updated identification value; the first identification updating unit comprises a first identification updating sub-unit configured to: when the non-illumination model distortion value is greater than the illumination model distortion value, perform an identification updating operation on the initial identification value to obtain the first updated identification value as the second illumination compensation identification value; and when the non-illumination model distortion value is less than or equal to the illumination model distortion value, perform an identification updating operation on the initial identification value to obtain the first updated identification value as the first illumination compensation identification value.

[0121] In an example embodiment of the present disclosure, the updated identification value further comprises a second updated identification value; and the identification updating module 930 further comprises a second identification updating unit configured to: when the first updated identification value is the first illumination compensation identification value, perform a second updating operation on the updated identification value to obtain the second updated identification value.

[0122] In an example embodiment of the present disclosure, the second identification updating unit comprises a second identification updating sub-unit configured to: split a target template to obtain a first target split template and a second target split template; split a reference template to obtain a first reference split template and a second reference split template; determine an illumination compensation parameter according to the first reference split template; perform a brightness illumination compensation on the second reference split template based on the illumination compensation parameter to obtain a second reference illumination template; and determine the second updated identification value according to a split template distortion comparison result between the second reference illumination template, the second reference split template and the second target split template.

[0123] In an example embodiment of the present disclosure, the second identification updating subunit comprises an illumination compensation subunit, configured to: determine an illumination reference pixel point based on the second reference split template, determine an illumination compensation identification value corresponding to the illumination reference pixel point; when the illumination compensation identification value is the first illumination compensation identification value, take the pixel value of the illumination reference pixel point as a target illumination reference pixel value; when the illumination compensation identification value is the second illumination compensation identification value, perform brightness illumination compensation on the illumination reference pixel point based on the illumination compensation parameter, and take the obtained illumination compensation pixel value as the target illumination reference pixel value.

[0124] In an example embodiment of the present disclosure, the second identification updating subunit is configured to perform: taking the split template distortion between the second reference split template and the second target split template as a non-illumination model distortion value; taking the split template distortion between the second reference illumination template and the second target split template as an illumination model distortion value; when the non-illumination model distortion value is greater than the illumination model distortion value, determining the second update identification value as the second illumination compensation identification value; when the non-illumination model distortion value is less than or equal to the illumination model distortion value, determining the second update identification value as the first illumination compensation identification value.

[0125] In an example embodiment of the present disclosure, the inter prediction module 940 comprises an inter prediction unit, configured to: when the update identification value is in an open state, perform brightness illumination compensation processing on the current prediction block to obtain a prediction value of the current to-be-predicted block; when the update identification value is in a non-open state, determine a motion vector based on the candidate motion information; determine an interpolation prediction block based on the motion vector, and take the interpolation prediction block as the prediction value of the current to-be-predicted block.

[0126] As to the apparatus in the above-mentioned embodiments, the specific manners in which various modules perform operations have been described in detail in the embodiments of the method, and will not be described here in detail.

[0127] The electronic device 1000 according to this embodiment of the present disclosure will be described below with reference to Figure 10 Figure 10 The display electronic device 1000 is merely an example, and should not impose any limitation on the functions and use range of the embodiments of the present disclosure.

[0128] As ​ shown, the electronic device 1000 is in the form of a general computing device. The components of the electronic device 1000 can include, but are not limited to, the above-mentioned at least one processing unit 1010, the above-mentioned at least one storage unit 1020, a bus 1030 connecting different system components (including the storage unit 1020 and the processing unit 1010), and a display unit 1040.

[0129] ​The storage unit stores program codes which can be executed by the processing unit 1010, so that the processing unit 1010 performs the steps described in the above "Exemplary Method" section according to various exemplary embodiments of the present disclosure.

[0130] The storage unit 1020 can include a readable medium in the form of volatile storage such as random access memory (RAM) 1021 and / or cache memory 1022, and also can include a non-volatile storage such as read-only memory (ROM) 1023.

[0131] The storage unit 1020 can also include program / utility 1024 having a set of programs / modules 1025, including an operating system, one or more application programs, other programs, and programmatic data, each or any combination thereof, which can include implementation of a network environment.

[0132] The bus 1030 can represent one or more of several types of bus structures, including a storage bus or bus controller, a peripheral bus, a graphics bus (e.g., AGP or Accelerated Graphics Port), an internal bus, or a local bus using any of a variety of bus architectures.

[0133] The electronic device 1000 can also communicate with one or more external devices 1070 such as a keyboard or pointing device, a Bluetooth device, etc.; other devices that enable a user to interact with the electronic device 1000; and / or one or more devices that enable the electronic device 1000 to communicate with one or more other computing devices. Such communication can be facilitated by an Input / Output (I / O) interface 1050. Still yet, the electronic device 1000 can communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or the Internet, through a network adapter 1060. As depicted, the network adapter 1060 can communicate with the other components of the electronic device 1000 through the bus 1030. It should be appreciated that although the network adapter 1060 is depicted as a single component, the network adapter 1060 can comprise two or more components that work together to facilitate communications between the electronic device 1000 and one or more other computing devices.

[0134] In exemplary embodiments, a computer readable storage medium is also provided, such as a memory including instructions, which can be executed by a processor of an apparatus to complete the above-mentioned inter-prediction method. Optionally, the computer readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0135] In an example embodiment, there is also provided a computer program product comprising a computer program which, when executed by a processor, implements the inter prediction method of any of the above.

[0136] Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the features disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the disclosure being indicated by the following claims.

[0137] It will be understood that the present disclosure is not limited to the precise structures herein described and illustrated in the drawings, and that various modifications and changes can be made without departing from its scope. The scope of the present disclosure is limited only by the claims that follow.

Claims

1. An inter prediction method, characterized by, The method comprises: obtaining candidate motion information of a current block to be predicted, determining a reference block corresponding to the current block to be predicted based on the candidate motion information, and the candidate motion information comprising an initial identification value corresponding to a local illumination compensation identification; determining a target template of the current block to be predicted and a reference template corresponding to the reference block, and a template distortion between the target template and the reference template; performing identification update operation on the initial identification value according to the template distortion to obtain an updated identification value; performing inter-frame prediction processing on the current block to be predicted according to the updated identification value to obtain a prediction value of the current block to be predicted.

2. The method of claim 1, wherein, The determination of the target template of the current block to be predicted and the reference template corresponding to the reference block comprises: constructing the target template based on a specified number of first pixels adjacent to the current block to be predicted; determining reference block position information corresponding to the reference block according to the candidate motion information; determining a specified number of second pixels adjacent to the reference block based on the reference block position information; constructing the reference template based on the second pixels.

3. The method of claim 1, wherein, The identification update operation on the initial identification value according to the template distortion to obtain an updated identification value comprises: determining a non-illumination model distortion value corresponding to a non-illumination model distortion and an illumination model distortion value corresponding to an illumination model distortion according to the initial identification value and the template distortion; performing identification update operation on the initial identification value according to a comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identification value.

4. The method of claim 3, wherein, The template distortion comprises a first template distortion value and a second template distortion value; and the determination of the non-illumination model distortion value corresponding to the non-illumination model distortion and the illumination model distortion value corresponding to the illumination model distortion according to the initial identification value and the template distortion comprises: when the initial identification value is a first illumination compensation identification value, determining a first distortion value corresponding to the non-illumination model distortion according to the first template distortion value, and determining a second distortion value corresponding to the illumination model distortion according to the second template distortion value; when the initial identification value is a second illumination compensation identification value, determining a third distortion value corresponding to the non-illumination model distortion according to the first template distortion value, and determining a fourth distortion value corresponding to the illumination model distortion according to the second template distortion value.

5. The method of claim 3, wherein, The updated identification value comprises a first updated identification value; and the identification update operation on the initial identification value according to the comparison result between the non-illumination model distortion value and the illumination model distortion value to obtain the updated identification value comprises: when the non-illumination model distortion value is greater than the illumination model distortion value, performing identification update operation on the initial identification value, and taking the second illumination compensation identification value as the first updated identification value; when the non-illumination model distortion value is less than or equal to the illumination model distortion value, performing identification update operation on the initial identification value, and taking the first illumination compensation identification value as the first updated identification value.

6. The method of claim 5, wherein, The updated identification value further comprises a second updated identification value, and the method further comprises: When the first update identifier value is the first illumination compensation identifier value, a secondary update operation is performed on the update identifier value to obtain the second update identifier value.

7. The method of claim 6, wherein, The secondary update operation performed on the update identifier value to obtain the second update identifier value comprises: splitting the target template to obtain a first target split template and a second target split template; splitting the reference template to obtain a first reference split template and a second reference split template; determining an illumination compensation parameter according to the first reference split template; performing brightness illumination compensation on the second reference split template based on the illumination compensation parameter to obtain a second reference illumination template; determining the second update identifier value according to a split template distortion comparison result between the second reference illumination template, the second reference split template and the second target split template.

8. The method of claim 7, wherein, The brightness illumination compensation performed on the second reference split template based on the illumination compensation parameter to obtain a second reference illumination template comprises: determining an illumination reference pixel point based on the second reference split template, and determining an illumination compensation identifier value corresponding to the illumination reference pixel point; when the illumination compensation identifier value is a first illumination compensation identifier value, taking a pixel value of the illumination reference pixel point as a target illumination reference pixel value; when the illumination compensation identifier value is a second illumination compensation identifier value, performing brightness illumination compensation on the illumination reference pixel point based on the illumination compensation parameter, and taking an obtained illumination compensation pixel value as the target illumination reference pixel value.

9. The method of claim 7, wherein, The determination of the second update identifier value according to a split template distortion comparison result between the second reference illumination template, the second reference split template and the second target split template comprises: taking split template distortion between the second reference split template and the second target split template as the non-illumination model distortion value; taking split template distortion between the second reference illumination template and the second target split template as the illumination model distortion value; when the non-illumination model distortion value is greater than the illumination model distortion value, determining the second update identifier value as the second illumination compensation identifier value; when the non-illumination model distortion value is less than or equal to the illumination model distortion value, determining the second update identifier value as the first illumination compensation identifier value.

10. The method according to any one of claims 1 to 9, characterized in that, The inter prediction processing of the current to-be-predicted block according to the update identifier value to obtain a prediction value of the current to-be-predicted block comprises: when the update identifier value is in an open state, performing brightness illumination compensation processing on the current prediction block to obtain the prediction value of the current to-be-predicted block; when the update identifier value is in a non-open state, determining a motion vector based on the candidate motion information; determining an interpolation prediction block based on the motion vector, and taking the interpolation prediction block as the prediction value of the current to-be-predicted block.

11. An inter prediction apparatus, characterized by comprising: comprises: a reference block determination module configured to acquire candidate motion information of a current to-be-predicted block, and determine a reference block corresponding to the current to-be-predicted block based on the candidate motion information, wherein the candidate motion information comprises an initial identifier value corresponding to a local illumination compensation identifier. a template distortion determination module, configured to determine a target template of the current block to be predicted and a reference template corresponding to the reference block, and a template distortion between the target template and the reference template; an identification update module, configured to perform an identification update operation on the initial identification value according to the template distortion, to obtain an updated identification value; an inter prediction module, configured to perform an inter prediction process on the current block to be predicted according to the updated identification value, to obtain a prediction value of the current block to be predicted.

12. An electronic device, comprising: comprise: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement the inter prediction method according to any one of claims 1 to 10.

13. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the inter prediction method according to any one of claims 1 to 10.

Images