Coding unit division method and device, equipment, program product and storage medium

Abstract

The invention discloses a coding unit division method and device, equipment, a program product and a storage medium, and the method comprises the steps: obtaining first information, the first information comprises one or more of the following information: information of an upper-layer CU of a first coding unit (CU), information of a lower-layer CU of the first CU, information of an adjacent CU of the first CU and prediction mode information of the first CU; determining a first judgment result of the first CU according to a first classifier and the first information, wherein the first judgment result represents whether the first CU skips a first division mode or not; and determining a second judgment result of the first CU according to the first information, the second judgment result representing whether the first CU skips a second division mode or not.

Description

Technical Field

[0001] The present application relates to the field of video coding technology, and in particular to a coding unit division method and apparatus, device, program product, and storage medium. Background Art

[0002] In recent years, with the improvement of network bandwidth and the continuous development of video transmission technology, the market demand for ultra-high-definition video technology and VR technology has continued to increase. Especially in recent years, the way people communicate with each other has shifted to online video communication, and with the development of cloud computing technology, the way people work has become more diverse. As a new type of office carrier, cloud computers have great development potential. Whether it is online communication or online office, video communication technology plays a very important role. Especially for online meetings, how to improve user experience and improve the quality of video calls requires in-depth exploration of the transmission technology involved. Summary of the invention

[0003] In order to solve the above technical problems, the present application provides a coding unit division method and device, equipment, program product, and storage medium.

[0004] The present application provides a coding unit division method, the method comprising:

[0005] Acquire first information, where the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of a neighboring CU of the first CU, and prediction mode information of the first CU;

[0006] A first judgment result of the first CU is determined according to the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and a second judgment result of the first CU is determined according to the first information, the second judgment result indicating whether the first CU skips the second division mode.

[0007] The coding unit division device provided in the present application includes:

[0008] An acquiring unit, configured to acquire first information, where the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of an adjacent CU of the first CU, and prediction mode information of the first CU;

[0009] A determination unit is used to determine a first judgment result of the first CU according to the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and to determine a second judgment result of the first CU according to the first information, the second judgment result indicating whether the first CU skips the second division mode.

[0010] The coding unit division device provided in the present application includes: a processor and a memory, the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the above-mentioned coding unit division method.

[0011] A computer program product provided by the present application includes: a computer program, and the computer program implements the above method when executed by a processor.

[0012] The computer-readable storage medium provided in the present application is used to store a computer program, and the computer program enables a computer to execute the above method.

[0013] In the technical solution of the present application, first information is obtained, and the first information includes one or more of the following information: information of the upper CU of the first coding unit CU, information of the lower CU of the first CU, information of the adjacent CU of the first CU, and prediction mode information of the first CU; a first judgment result of the first CU is determined according to the first classifier and the first information, and the first judgment result indicates whether the first CU skips the first division mode; and a second judgment result of the first CU is determined according to the first information, and the second judgment result indicates whether the first CU skips the second division mode. In this way, by determining whether the first CU skips the first division mode through the first classifier and the first information, and judging whether the second division mode is skipped through the first information, the division efficiency of the first CU is improved, and the accuracy of the first CU division is also improved. Furthermore, by efficiently and accurately dividing the first CU, the coding complexity is reduced and the stability of the image quality is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 Schematic diagram of the coding framework of the multifunctional video coding VVC provided in the embodiment of the present application;

[0015] Figure 2 It is a schematic diagram of a multi-type tree partitioning structure adopted by VVC provided in an embodiment of the present application;

[0016] Figure 3 This is a schematic diagram of the QTMT division of the coding tree unit CTU provided in an embodiment of the present application;

[0017] Figure 4 It is a schematic diagram of a VVC intra-frame prediction mode provided in an embodiment of the present application;

[0018] Figure 5 The following is a flow chart of the coding unit division method provided in the embodiment of the present application. Figure 1 ;

[0019] Figure 6 This is a statistical result diagram of different sequences and different division types provided in the embodiments of the present application;

[0020] Figure 7 It is a current CU horizontal partition probability map under different partition types provided in an embodiment of the present application;

[0021] Figure 8 It is a current CU vertical partition probability map under different partition types provided in an embodiment of the present application;

[0022] Fig. 9 The following is a flow chart of the coding unit division method provided in the embodiment of the present application. Figure 2 ;

[0023] Fig.10 It is a rate-distortion curve diagram of different test sequences provided in the embodiments of the present application under VTM6.0 and the algorithm of the present application;

[0024] Fig.11 It is a schematic diagram of the structural composition of the coding unit division device provided in an embodiment of the present application;

[0025] Fig.12 is a schematic structural diagram of a coding unit division device provided in an embodiment of the present application;

[0026] Fig.13 It is a schematic structural diagram of the chip of an embodiment of the present application. DETAILED DESCRIPTION

[0027] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0029] It should also be pointed out that the terms "first\second\third" involved in the embodiments of the present application are only used to distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here. The term "and / or" herein is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the front and back associated objects are in an "or" relationship. It should also be understood that the "indication" mentioned in the embodiments of the present application can be a direct indication, an indirect indication, or an indication of an association relationship. For example, A indicates B, which can mean that A directly indicates B, for example, B can be obtained through A; it can also mean that A indirectly indicates B, for example, A indicates C, and B can be obtained through C; it can also mean that there is an association relationship between A and B. It should also be understood that the "correspondence" mentioned in the embodiments of the present application may indicate a direct or indirect correspondence between the two, or an association between the two, or a relationship of indication and being indicated, configuration and being configured, etc.

[0030] To facilitate understanding of the technical solutions of the embodiments of the present application, the relevant technologies of the embodiments of the present application are described below. The following related technologies can be arbitrarily combined with the technical solutions of the embodiments of the present application as optional solutions, and they all belong to the protection scope of the embodiments of the present application.

[0031] In recent years, with the improvement of network bandwidth and the continuous development of video transmission technology, the market demand for ultra-high-definition video technology and VR technology has continued to increase. Especially in recent years, the way people communicate with each other has shifted to online video communication, and with the development of cloud computing technology, the way people work has become more diverse. As a new type of office carrier, cloud computers have great development potential. Whether it is online communication or online office, video communication technology plays a very important role. Especially for online meetings, how to improve user experience and improve the quality of video calls requires in-depth exploration of the transmission technology involved.

[0032] As a key technology in the video coding process, intra-frame coding is widely used in image coding. Intra-frame coding is derived from differential coding, that is, using the previous similar information to predict itself to reduce redundancy. In recent years, many scholars have improved the intra-frame prediction method and reduced the computational complexity. Fast intra-frame coding methods can be divided into data-driven methods and heuristic methods. Among them, the data-driven method includes a fast algorithm based on a convolutional neural network, which does not depend on the depth of the coding unit (Coding Unit, CU) and the spatial correlation of the CU, is very friendly to parallel processing, and analyzes multi-type tree partitions through statistical methods. For heuristic methods, such as using the spatial correlation of texture features to skip certain partitions, such as decomposing the CU partition problem into a binary classification problem through a support vector machine, and using an offline support vector machine to accelerate the CU partition, in which the rate-distortion cost caused by misclassification is introduced as a weight in the support vector machine training, so that the algorithm has good and stable performance in different configurations and different video contents.

[0033] Figure 1 is a schematic diagram of a coding framework of a multifunctional video coding VVC provided in an embodiment of the present application, Figure 2 Schematic diagram of a multi-type tree partitioning structure used by VVC provided in an embodiment of the present application, such as Figure 2 As shown, part (a) is a horizontal binary tree partition, part (b) is a horizontal ternary tree partition, part (c) is a vertical binary tree partition, and part (d) is a vertical ternary tree partition. Figure 3 This is a schematic diagram of the QTMT division of the coding tree unit CTU provided in an embodiment of the present application. Figure 4 is a schematic diagram of a VVC intra-frame prediction mode provided in an embodiment of the present application, such as Figure 4 As shown in FIG. 1 , there are 67 intra-frame prediction modes. For the intra-frame prediction process of Versatile Video Coding (VVC), it is necessary to traverse the intra-frame prediction modes and all division modes of VVC for each coding unit. Therefore, the coding complexity of VVC is relatively high. However, although some existing technical solutions can reduce the coding complexity to a certain extent, there are still the following problems:

[0034] 1) Compression efficiency: Although existing video coding standards such as H.264, H.265 (HEVC), VP9, ​​etc. have made significant progress in compressing videos, the demand for higher compression efficiency still exists as resolution and bit rate increase, as well as the demand for higher quality videos. This means that more efficient compression algorithms are needed to reduce the size of video files to reduce storage and transmission costs.

[0035] 2) Real-time: Real-time video encoding is very important for many applications, such as video calls, live streaming, and games. Existing encoding algorithms may require a lot of computing resources, resulting in increased latency. Therefore, faster real-time encoding algorithms are needed to reduce latency and provide a better user experience.

[0036] 3) Multi-resolution and multi-device support: Modern video applications need to support multiple resolutions and different types of devices, from mobile phones to HDTVs. This brings challenges of adaptability and transcoding, requiring more efficient video encoding methods.

[0037] 4) The currently widely used fast encoding methods are mainly proposed for the previous generation video coding standard (High Efficiency Video Coding, HEVC), while the structure of the latest VVC video coding method has changed, and many improved methods are not applicable to the VVC standard.

[0038] Therefore, how to improve the encoding speed while ensuring the video encoding quality as much as possible becomes a problem that needs to be considered in video encoding. To this end, the following technical solutions of the embodiments of the present application are proposed.

[0039] To facilitate understanding of the technical solutions of the embodiments of the present application, the technical solutions of the present application are described in detail below through specific embodiments. The above related technologies can be combined arbitrarily with the technical solutions of the embodiments of the present application as optional solutions, and they all belong to the protection scope of the embodiments of the present application. The embodiments of the present application include at least part of the following contents.

[0040] Figure 5 The following is a flow chart of the coding unit division method provided in the embodiment of the present application. Figure 1 ,like Figure 5 As shown, the coding unit division method includes the following steps:

[0041] Step 501: Obtain first information, where the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of an adjacent CU of the first CU, and prediction mode information of the first CU.

[0042] Step 502: Determine a first judgment result of the first CU based on the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and determine a second judgment result of the first CU based on the first information, the second judgment result indicating whether the first CU skips the second division mode.

[0043] In some embodiments, the first CU is a currently encoded CU, the upper CU of the first CU is the CU of the parent node corresponding to the first CU as a child node and the CU at the same level as the parent node, the lower CU of the first CU is the CU of the child node corresponding to the first CU as a parent node and the CU at the same level as the child node, and the information of the adjacent CUs of the first CU includes information of the CU above the first CU and information of the CU below the first CU, wherein the information of the adjacent CUs includes prediction mode information and division method information of the adjacent CUs.

[0044] In some implementations, the first partitioning mode is a vertical binary tree partitioning mode in a VVC encoding process, and the second partitioning mode is a ternary tree partitioning mode.

[0045] In some embodiments, the encoding process of the embodiment of the present application starts according to the encoding process of VVC. First, the coding tree unit (Coding Tree Unit, CTU) division is started according to the encoding process of VVC. After the CTU division, VVC will perform an intra-frame prediction process. After the intra-frame prediction process is completed, binary tree division will be performed. First, horizontal binary tree division will be performed, and then a first judgment result of the first CU is determined according to the first classifier and the first information, and the first judgment result indicates whether the first CU skips the vertical binary tree division mode; a second judgment result of the first CU is determined according to the first information, and the second judgment result indicates whether the first CU skips the ternary tree division mode.

[0046] In some implementations, the CU partitioning process is modeled as a binary classification problem, and thus the first classifier is used as a basis for determining whether to skip the first partitioning mode, that is, the first classifier is used as a basis for determining whether to skip the vertical binary tree partitioning mode.

[0047] Specifically, the event of whether the current CU selects the vertical binary tree partition is defined as w i ,i∈{0,1}, where w0 and w1 represent the skip vertical binary tree partition mode and the non-skip vertical binary tree partition mode, respectively. i |g) is the condition that the given eigenvector g belongs to w i The posterior probability of , which determines whether the current CU skips the vertical binary tree partitioning mode, is calculated according to the Bayesian rule as shown in (1):

[0048]

[0049] Among them, P(g|w i ) is w i The class conditional probability density, P(w i ) and P(g) represent w i The prior probability and total probability density of .

[0050] Therefore, the condition for the current block to skip the vertical binary tree partition mode is as shown in formula (2):

[0051] P(w0|g)>P(w1|g) (2);

[0052] According to formula (1) and formula (2), formula (3) is obtained by simplifying:

[0053] P(g|w0)·P(w0)>P(g|w1)·P(w1) (3);

[0054] Among them, P(w0) and P(w1) can be obtained by directly statistically realizing the data, while P(g|w0) and P(g|w1) can be obtained by looking at the offline estimation and using the histogram-based non-parametric estimation method.

[0055] In some embodiments, the method also includes: obtaining a first data set, the first data set including second information of a CU in a first CU set, the second information including one or more of the following information: prediction mode information of the CU, partition mode information of the CU, and feature information of the CU; the CU in the first CU set is a CU corresponding to a target partition mode obtained after a training video set undergoes a standard video encoding process.

[0056] Specifically, a part of the VVC test sequence is selected, and after the VVC encoding process, the encoding result of the test sequence is obtained, and the encoding result includes the division information of each CU of each frame in each video sequence, the intra-frame prediction mode of each CU and the feature information of each CU.

[0057] In some embodiments, the method also includes: determining first probability information based on the first data set, the first probability information including probability information that the CU in the first CU set skips the first partitioning mode and probability information that the first partitioning mode is not skipped; determining conditional probability information based on the first data set, the conditional probability information including conditional probability information that the CU in the first CU set skips the first partitioning mode and conditional probability information that the first partitioning mode is not skipped; and constructing a first classifier based on the first probability information and the conditional probability information.

[0058] Specifically, the first data set includes CUs that have obtained the target partitioning mode. Therefore, the number of CUs that have been partitioned in the vertical binary tree mode and the number of CUs that have not been partitioned in the vertical binary tree mode can be counted based on the first data set, and then the probability information P(w0) of the CUs in the first CU set skipping the vertical binary tree partitioning mode and the probability information P(w1) of the CUs that have not skipped the vertical binary tree partitioning mode can be obtained. At the same time, the conditional probability information P(g|w0) of skipping the vertical binary tree partitioning mode and the conditional probability information P(g|w1) of not skipping the vertical binary tree partitioning mode can also be determined based on the first data set.

[0059] In some embodiments, the method also includes: constructing a first feature vector, the first feature vector including prediction mode information of each CU in the first CU set, partition mode information of adjacent CUs of each CU, and prediction mode information of adjacent CUs of each CU; determining conditional probability information based on the first data set, including: determining conditional probability information corresponding to the first feature vector based on the first data set.

[0060] In some implementations, the first feature vector is composed of the partition type and intra prediction direction of the adjacent CU, specifically, Figure 2 As shown in Figure 1, since there are 67 intra prediction directions in VVC, in order to reduce the coding complexity, the intra prediction directions are classified, where intra prediction modes 10-26 are defined as horizontal prediction directions, intra prediction modes 40-58 are defined as vertical prediction directions, and the remaining intra prediction modes are defined as diagonal prediction directions. The division mode of the upper sub-block of the current CU is defined as S U_t , the intra prediction direction of the upper sub-block is defined as S U_d , the partition type of the sub-block below is defined as S L_t , the intra prediction direction of the lower sub-block is defined as S L_d , the intra prediction method of the current CU is defined as S C_d , therefore, the first eigenvector g is composed of the above features, as shown in the following formula (4):

[0061] g(S U_t ,S L_t ,S U_d ,S L_d ,S C_d ) (4);

[0062] Therefore, determining the conditional probability information according to the first data set is to determine the conditional probability information corresponding to the first feature vector according to the first data set, that is, to obtain P(g|w0) and P(g|w1).

[0063] Specifically, the first eigenvector is first quantized using the quantization parameter Q, as shown in the following formula (5):

[0064]

[0065] Then construct the matrix and The conditional probability densities P(g|w0) and P(g|w1) are obtained, where Represents the set of feature vectors corresponding to the CU that skips the vertical binary tree partition, The set of feature vectors corresponding to the CU that does not skip the vertical binary tree partition is first initialized, where the initialization method is shown in the following equations (6) and (7):

[0066]

[0067] That is, at the very beginning, and The number of each eigenvector g in is initialized to 0.

[0068] in, and The update method of is shown in the following equations (8) and (9):

[0069]

[0070] When the CU (event x) in the first data set belongs to skipping vertical binary tree partitioning, the number of first feature vectors corresponding to the CU in the event is increased by 1. For example, when the upper sub-block of a CU adopts horizontal binary tree partitioning, the intra-frame prediction direction of the upper sub-block is horizontal prediction direction, the lower sub-block adopts vertical binary tree partitioning, the intra-frame prediction direction of the lower sub-block is vertical prediction direction, and the current CU adopts horizontal intra-frame prediction, then the number of events corresponding to this feature vector is increased by 1. After all CUs in the first data set are counted, all types and corresponding numbers of first feature vectors corresponding to CUs that skip vertical binary tree partitioning, as well as all types and corresponding numbers of first feature vectors of CUs that do not skip vertical binary tree partitioning can be obtained. Based on this, the class conditional probability density P(g|w0) and P(g|w1) corresponding to each first feature vector are obtained, and the calculation formulas are shown in the following equations (10) and (11):

[0071]

[0072] Among them, N0 and N1 represent the number of CUs in the statistical data of skipping vertical binary tree partition mode and not skipping vertical binary tree partition mode, respectively.

[0073] In this way, according to the statistics of the first data set P(w0), P(w1), P(g|w0) and P(g|w1), a first classifier is constructed based on this, and the first judgment result of the first CU is determined by the first classifier and the first information. The first judgment result is used to indicate whether the first CU skips the vertical binary tree partitioning mode.

[0074] In some embodiments, the method also includes: determining third information based on the first data set, the third information including the proportion of different division modes of CUs of different sizes in the first CU set and the relationship between the division mode of the upper CU and the division mode of the lower CU of each CU in the first CU set and the division mode of each CU; determining a first judgment strategy based on the third information, the first judgment strategy is used to determine whether the first CU skips the second division mode; the first information determines the second judgment result of the first CU, including: determining the second judgment result of the first CU based on the first information and the first judgment strategy.

[0075] In some embodiments, the data in the first data set are counted to obtain a distribution table of different partition types in CUs of different sizes, which shows the probability distribution of the partition type for each size of CU. At the same time, the relationship between the partition mode of the upper CU of each CU and the partition mode of the lower CU and the partition mode of each CU is obtained in the first data set. Based on this information, a first judgment strategy is determined to determine whether the first CU skips the ternary tree partition mode, that is, the second judgment result of the first CU is determined based on the first information and the first judgment strategy. Exemplarily, the first judgment strategy is that if both the upper CU and the lower CU are binary-tree-divided in the same direction, the current CU block will skip the ternary tree partition check, otherwise, the ternary tree partition check will continue according to the original partition process.

[0076] In some embodiments, the first information also includes size information of the first CU and feature information of the first CU; the method also includes: determining a third judgment result of the first CU based on the first information and the second judgment strategy; wherein the second judgment strategy is used to determine whether the first CU skips the third partitioning mode, and the third judgment result indicates whether the first CU skips the third partitioning mode.

[0077] In some embodiments, the third division mode is a quadtree division mode, and according to the size information of the first CU and the feature information of the first CU, the first information and the second judgment strategy are used to determine whether the first CU skips the quadtree division mode, wherein the second judgment strategy can be the judgment strategy for determining quadtree division in the VVC video coding standard, or it can be other judgment strategies for determining whether to perform quadtree division, and this application does not make specific limitations on this.

[0078] The technical solution of the embodiment of the present application obtains first information, and the first information includes one or more of the following information: information of the upper CU of the first coding unit CU, information of the lower CU of the first CU, information of the adjacent CU of the first CU, and prediction mode information of the first CU; a first judgment result of the first CU is determined according to the first classifier and the first information, and the first judgment result indicates whether the first CU skips the first division mode; and a second judgment result of the first CU is determined according to the first information, and the second judgment result indicates whether the first CU skips the second division mode. In this way, by determining whether the first CU skips the first division mode through the first classifier and the first information, and judging whether the second division mode is skipped through the first information, the division efficiency of the first CU is improved, and the accuracy of the first CU division is also improved. Furthermore, by efficiently and accurately dividing the first CU, the coding complexity is reduced and the stability of the image quality is guaranteed.

[0079] The technical solution of the embodiment of the present application is illustrated below with reference to specific examples.

[0080] The embodiment of the present application proposes a VVC fast encoding method based on Naive Bayes, which can greatly improve the video compression efficiency and fully meet the video transmission requirements of higher resolution and higher bit rate. On the cloud video platform, the Naive Bayes VVC video encoding technology can be used to compress and optimize the stored and transmitted video content to better reduce bandwidth costs and improve transmission efficiency.

[0081] The specific steps include:

[0082] S1: Start CTU division according to the VVC encoding process, perform intra-frame prediction and horizontal binary tree division;

[0083] S2: Construct a naive Bayes classifier, using the partition type and intra prediction direction as feature vectors to construct a naive Bayes classifier;

[0084] S3: Calculate the probability of skipping vertical division and the probability of not skipping vertical division by using the Naive Bayes theorem, and compare them to decide whether to skip the vertical division process of BT;

[0085] S4: Determine whether to skip TT division based on the feature information and statistical data of the current block;

[0086] S5: Determine whether to perform quadtree division. If quadtree division is performed, perform quadtree division mode. If quadtree division is not performed, determine whether the depth of the current CU is the maximum depth. If not, increase the CU depth by 1 and repeat steps S1 to S5. Otherwise, obtain the best CU division mode and end the CU division process.

[0087] The specific implementation is as follows:

[0088] In step S1, CTU division is mainly started according to the VVC encoding process, where the VVC encoding framework is shown in the figure Figure 1 As shown in Figure 2, compared with the previous generation video coding standard HEVC, the block partitioning method adopted by VVC is a multi-type tree partitioning structure. Figure 2 As shown, the result of the division is shown in Figure 3 As shown, VVC will perform intra-frame prediction after CTU division, and the intra-frame prediction mode is shown in the figure Figure 4 As shown, after the intra-frame prediction is completed, binary tree division will be performed. First, horizontal binary tree division will be performed, and whether to perform vertical binary tree division is determined by steps S2 to S3.

[0089] In step S2, the CU partitioning process is modeled as a binary classification problem, and the event of whether the current CU selects the vertical binary tree partitioning is defined as w i ,i∈{0,1}, where w0 and w1 represent the events of skipping the vertical binary tree partition mode and not skipping the vertical binary tree partition mode respectively. i |g) is the condition that the given eigenvector g belongs to w i The posterior probability of , which determines whether the current CU skips vertical partitioning. According to the Bayesian rule, the calculation formula is shown in formula (1).

[0090] Since there are 67 intra prediction directions in VVC, in order to reduce the coding complexity, the intra prediction directions can be classified, where the intra prediction modes 10-26 are defined as horizontal prediction directions, the intra prediction modes 40-58 are defined as vertical prediction directions, and the remaining intra prediction modes are defined as diagonal prediction directions. The feature vector in the embodiment of the present application is mainly composed of the partition type of the adjacent CU and the intra prediction direction. Specifically, the partition type of the upper sub-block of the current CU is defined as S U_t , the partition type of the sub-block below is defined as S L_t , their intra mode prediction directions are defined as S U_d and S L_d The intra prediction direction of the current CU is defined as S C_d The joint feature vector g is composed of the above features, as shown in formula (4). Then it can be seen that the condition for the current block to skip the vertical binary tree partition mode is shown in formula (2), which can be simplified to formula (3) based on the above formula.

[0091] In step S3, in order to obtain the class conditional probability density P(g|w i ), it is necessary to establish a multivariate data set consisting of JVET sequences. The specific test conditions are shown in Table 1 below.

[0092]

[0093] Table 1 JVET sequence test conditions

[0094] The specific method of non-parametric estimation is as follows: first, the feature vector g is quantized by the quantization parameter Q, as shown in formula (5). Then, by constructing and Initialization is performed, and the initialization formulas are shown in formulas (6) and (7). and The update method of is shown in formula (8) and formula (9). The class conditional probability density estimation is shown in formula (10) and formula (11).

[0095] Since some results may be irregular when estimating non-parametric density using this method, a smoothing filter is needed to regularize the parameter estimation in order to obtain regular data.

[0096] In step S4, the feature information and statistical data of the current block are used to determine whether to skip the ternary tree (TT) partition. The specific process is as follows: In VVC, the entire image is first divided into several blocks of size 128×128, which are called CTUs. Then, these blocks are divided into 4 CUs of size 64×64, and then the CUs are further divided using multi-type tree partitioning. In order to further understand the proportion of different partitioning methods in the block partitioning process, the VVC block partitioning situation can be statistically analyzed. Figure 6 is a statistical result diagram of different sequences and different division types provided in the embodiment of the present application, Figure 6 It can be seen that for CU blocks of size 64×64, there is a probability of nearly 80% that quadtree partitioning will be selected. For CU blocks of size 32×32, the most likely partitioning is still quadtree partitioning, while for small-sized blocks such as 16×8, 8×16, and 8×8, most blocks will choose not to be partitioned (NS).

[0097] In order to further explore the correlation between CU size and partition type, it is necessary to integrate and statistically analyze the above result data. The statistical results are shown in Table 2.

[0098] Table 2 Distribution of different partition types in CUs of different sizes (%)

[0099]

[0100] From the data in Table 2, it can be seen that the quadtree (QT) type is almost only used for larger CUs and accounts for a large proportion. This means that most of the multiple-type tree (MTT) types can be skipped at an early stage. In small CUs (any CU less than or equal to 16×16 pixels), the proportion of NS is about 50% or even higher, indicating that nearly half of the small CUs do not need to be further divided. In addition, since the QT type is not allowed to appear under the MTT node, there is almost no chance that the QT type will appear in a small-sized CU. From the perspective of MT, regardless of the size of the CU, the proportion of the TT type is one-third to one-half of the (binary tree) BT type. According to the above observations, it can also be found that the proportion of horizontal division is almost equal to the proportion of vertical division, and the vertical division is slightly lower, indicating that the complexity of checking the vertical division may be lower than that of the horizontal division to a certain extent. Therefore, targeted strategies can be adopted to optimize the coding performance, such as giving priority to the QT type in large CUs and the NS type in small CUs. At the same time, in terms of the selection of horizontal division and vertical division, the optimal decision can be made according to the actual situation to reduce the coding complexity and improve the coding efficiency.

[0101] In the VVC standard, a leaf CU refers to the smallest block that cannot be further divided after a CU has been divided multiple times. There is a certain relationship between the division type of the current CU and its leaf CU in video coding, that is, the division type of the leaf CU will affect the division type selection of the current CU. In order to study the relationship between them, an analysis can be conducted to further understand this relationship by exploring the probability that the division type of the current CU has different division types in its leaf CU. Figure 7 is a current CU horizontal partition probability map under different partition types provided in an embodiment of the present application, Figure 8 It is a vertical partition probability map of the current CU under different partition types provided in an embodiment of the present application.

[0102] exist Figure 7 and Figure 8 In the figure, the horizontal axis represents the partition type of the upper leaf CU (upper layer CU), the vertical axis represents the probability of the current CU selecting horizontal or vertical partitioning, and the partition type of the lower leaf CU (lower layer CU) is represented by a legend. Figure 7 As shown in . When both the upper leaf CU and the lower leaf CU choose horizontal partitioning, the probability of the current CU choosing horizontal partitioning is more than 70%. When the upper leaf CU and the lower leaf CU choose vertical ternary tree partitioning or no partitioning respectively, the probability of the current CU choosing horizontal partitioning is less than 3%, which is the lowest value. When both the upper leaf CU and the lower leaf CU choose horizontal partitioning, the proportion of horizontal partitioning remains at a high level. Similar data distribution is also shown in Figure 8As shown in the figure, when the upper leaf CU and the lower leaf CU perform vertical division at the same time, the current CU is more likely to choose vertical division. When the upper leaf CU and the lower leaf CU perform vertical ternary tree division at the same time, the probability of the current CU choosing vertical ternary tree division exceeds 80%. However, when both the upper leaf CU and the lower leaf CU choose to divide in the horizontal direction, the current CU almost no longer chooses to divide in the vertical direction, and its minimum probability is less than 3%. Experimental data also show that different division types in the same direction have different effects on the current CU division type. For example, when the lower leaf CU performs vertical ternary tree division, the current CU is very likely to choose vertical division, regardless of the division type of the upper leaf CU. From the above analysis, it can be seen that if you want to skip the CU division stage in advance, you can also use CU spatial correlation to achieve it.

[0103] According to the coding rules, VVC will first perform binary tree division and then ternary tree division when performing block division. When performing ternary tree division, the ternary tree division check stage will be performed first. Only after the ternary tree division check is completed will the ternary tree division be continued. Therefore, the coding information of the binary tree division can be used to decide whether to skip the ternary tree division in the next stage. If both the upper leaf CU and the lower leaf CU are binary-divided in the same direction, the current CU block will skip the ternary tree division check. Otherwise, the ternary tree division check will continue according to the original division process.

[0104] In step S5, a quadtree division judgment is performed. If quadtree division is performed, the quadtree division mode is performed. If quadtree division is not performed, it is determined whether the depth of the current CU is the maximum depth. If it is not the maximum depth, the CU depth is +1, and steps S1 to S5 are repeated. Otherwise, the best CU division mode is obtained and the CU division process is terminated.

[0105] Based on the above, Fig. 9 The following is a flow chart of the coding unit division method provided in the embodiment of the present application. Figure 2 ,like Fig. 9As shown, CTU division is started, and the intra-frame prediction process is performed. Because binary tree division and ternary tree division are judged for CUs of specific sizes, for a CTU, a quadtree division judgment is first performed. When the CTU is divided into a size of 32×32, binary tree and ternary tree division judgments are performed. In the binary tree division, horizontal binary tree division is first performed, and then vertical binary tree division is determined according to the previously established Bayesian classifier. If vertical binary tree division is required, vertical binary tree division is performed. Otherwise, it is determined whether the ternary tree division condition is met. If the ternary tree division condition is met, ternary tree division is performed. Otherwise, it is determined whether the quadtree division condition is met. If the quadtree division condition is met, quadtree division is performed. Otherwise, the quadtree division is terminated, and then the depth of the current CU is judged. If the CU can no longer be divided, that is, the CU depth is the largest, the best CU division mode is obtained, and the CU division is terminated. Otherwise, the depth is increased by 1, and the aforementioned steps are repeated until the CU division is terminated.

[0106] In the VTM6.0 encoder, the test results of the AE class video sequence in the VVC standard general test sequence are shown in Table 3 below. The experimental results show that the average encoding time of the algorithm proposed in the embodiment of the application is reduced by 37.67%, and the BDBR is only increased by 0.73%.

[0107] Table 3 Experimental results of the embodiment of the present application in the VVC general test sequence

[0108]

[0109] Fig.10 : is a rate-distortion curve diagram of different test sequences provided in the embodiments of the present application under VTM6.0 and the algorithm of the present application, such as Fig.10 As shown in Figure 1, the rate-distortion curves of the original encoder VTM6.0 and the encoder after using the optimized algorithm are shown. Fig.10 It can be seen that the two curves are almost overlapped, indicating that the proposed algorithm performs basically the same as the original VTM6.0 in terms of rate-distortion performance. The effectiveness of the proposed algorithm has been verified and is expected to play a role in practical applications.

[0110] The technical solution of the embodiment of the present application proposes a VVC block partition optimization algorithm based on naive Bayes for the redundancy problem of the CU partition process caused by the multi-type tree partition structure. The algorithm uses the CU partition type and CU partition direction as feature vectors. The embodiment of the present application is divided into two parts. The first part is to skip the vertical binary tree partition stage in VVC encoding, and the second part is to select whether to skip the ternary tree partition based on the first part by using the current block encoding information. The experimental results show that the bit rate loss of the algorithm is only 0.73%, and the time complexity is reduced by 37.67%. In addition, by testing on different video sequences, the results show that the embodiment of the present application shows a good optimization effect on most video sequences. Therefore, the embodiment of the present application has practical significance and application prospects for improving the efficiency and performance of VVC encoding. In the embodiment of the present application, how the naive Bayes classifier has an important influence on the experimental results. The embodiment of the present application needs to design a naive Bayes classifier based on the Bayesian principle to make a probabilistic judgment on whether the current block needs to continue to be divided or whether to skip the vertical division. Among them, the input of the classifier includes various features of the current block. The embodiment of the present application selects the CU division type and CU division direction, and obtains reliable experimental results through a large amount of statistical data.

[0111] The advantages of the embodiments of the present application are as follows: Feature vector design: The division type and intra-frame prediction direction are used as feature vectors, and the optimal block division scheme is predicted by the naive Bayes classifier. Compared with the traditional method, not only the size and depth of the image block are considered, but also the vertical or horizontal division method of the block is considered, which further improves the coding efficiency. Model optimization: By adjusting parameters such as the prior probability of the naive Bayes classifier and the proportion of training samples, the accuracy and robustness of the model are optimized. At the same time, a multi-level block division strategy is introduced, combined with the naive Bayes method, to further improve the accuracy and efficiency of block division. Excellent experimental results: The VVC block division optimization method based on naive Bayes showed high block division accuracy and coding efficiency in the experiment. Compared with the traditional method, the embodiments of the present application greatly reduce the coding complexity while ensuring the stability of image quality.

[0112] The embodiment of the present application proposes to utilize features such as naive Bayes, a histogram-based non-parametric estimation method, and statistical parameters for whether the current block should continue to be divided. By establishing a naive Bayes classifier model for CUs of different sizes, the problem of low prediction accuracy of traditional deep learning is solved, and a VVC block division optimization method is proposed. Compared with the traditional method, the VVC video coding compression efficiency is greatly improved. As a new generation of video coding method, it can meet the needs of different application scenarios, especially in terms of high resolution, HDR, WCG and network adaptability. The lower BDBR loss and higher time complexity reduction of the embodiment of the present application have important practical significance and application prospects.

[0113] The VVC fast coding method based on naive Bayes can provide higher quality video content for video streaming, HDTV, 4K and 8K video, virtual reality and other high-resolution applications by improving coding efficiency and performance, thus meeting the growing market demand. This can improve customer satisfaction, reduce transmission and storage costs, and bring considerable commercial profits to enterprises.

[0114] Fig.11 is a schematic diagram of the structure of the coding unit division device provided in an embodiment of the present application, such as Fig.11 As shown, the coding unit division device includes:

[0115] An acquiring unit 1101 is configured to acquire first information, where the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of an adjacent CU of the first CU, and prediction mode information of the first CU;

[0116] The determination unit 1102 is used to determine a first judgment result of the first CU based on the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and to determine a second judgment result of the first CU based on the first information, the second judgment result indicating whether the first CU skips the second division mode.

[0117] In some embodiments, the acquisition unit 1101 is used to acquire a first data set, the first data set includes second information of a CU in a first CU set, the second information includes one or more of the following information: prediction mode information of the CU, partition mode information of the CU, and feature information of the CU; the CU in the first CU set is a CU corresponding to a target partition mode obtained after a training video set undergoes a standard video encoding process.

[0118] In some embodiments, the determination unit 1102 is used to determine first probability information based on the first data set, the first probability information including probability information that the CU in the first CU set skips the first partitioning mode and probability information that the first partitioning mode is not skipped; determine conditional probability information based on the first data set, the conditional probability information including conditional probability information that the CU in the first CU set skips the first partitioning mode and conditional probability information that the first partitioning mode is not skipped; the device also includes: a construction unit 1103; the construction unit 1103 is used to construct a first classifier based on the first probability information and the conditional probability information.

[0119] In some embodiments, the construction unit 1103 is used to construct a first feature vector, which includes prediction mode information of each CU in the first CU set, partition mode information of adjacent CUs of each CU, and prediction mode information of adjacent CUs of each CU; the determination unit 1102 is used to determine conditional probability information corresponding to the first feature vector based on the first data set.

[0120] In some embodiments, the determination unit 1102 is used to determine third information based on the first data set, the third information including the proportion of different division modes of CUs of different sizes in the first CU set and the relationship between the division mode of the upper CU and the division mode of the lower CU of each CU in the first CU set and the division mode of each CU; determine a first judgment strategy based on the third information, the first judgment strategy is used to determine whether the first CU skips the second division mode; determine a second judgment result of the first CU based on the first information and the first judgment strategy.

[0121] In some embodiments, the first information also includes size information of the first CU and feature information of the first CU; the determination unit 1102 is used to determine the third judgment result of the first CU based on the first information and the second judgment strategy; wherein the second judgment strategy is used to determine whether the first CU skips the third division mode, and the third judgment result indicates whether the first CU skips the third division mode.

[0122] Those skilled in the art should understand that Fig.11 The implementation functions of each unit in the coding unit division device shown can be understood by referring to the relevant description of the aforementioned method. Fig.11 The functions of each unit in the coding unit division device shown can be implemented by a program running on a processor, or by a specific logic circuit.

[0123] Fig.12 It is a schematic structural diagram of a coding unit division device 1200 provided in an embodiment of the present application. Fig.12 The coding unit division device 1200 shown includes a processor 1210, which can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0124] Alternatively, if Fig.12 As shown, the coding unit division device 1200 may further include a memory 1220. The processor 1210 may call and run a computer program from the memory 1220 to implement the method in the embodiment of the present application.

[0125] The memory 1220 may be a separate device independent of the processor 1210 , or may be integrated into the processor 1210 .

[0126] Alternatively, if Fig.12 As shown, the coding unit division device 1200 may further include a transceiver 1230, and the processor 1210 may control the transceiver 1230 to communicate with other devices, specifically, may send information or data to other devices, or receive information or data sent by other devices.

[0127] The transceiver 1230 may include a transmitter and a receiver. The transceiver 1230 may further include an antenna, and the number of antennas may be one or more.

[0128] The coding unit division device 1200 can implement the corresponding processes implemented by the coding unit division device in each method of the embodiments of the present application. For the sake of brevity, they are not repeated here.

[0129] Fig.13 It is a schematic structural diagram of the chip of an embodiment of the present application. Fig.13 The chip 1300 shown includes a processor 1310, which can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0130] Alternatively, if Fig.13 As shown, the chip 1300 may further include a memory 1320. The processor 1310 may call and run a computer program from the memory 1320 to implement the method in the embodiment of the present application.

[0131] The memory 1320 may be a separate device independent of the processor 1310 , or may be integrated into the processor 1310 .

[0132] Optionally, the chip 1300 may further include an input interface 1330. The processor 1310 may control the input interface 1330 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.

[0133] Optionally, the chip 1300 may further include an output interface 1340. The processor 1310 may control the output interface 1340 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.

[0134] The chip can implement the corresponding processes implemented by the coding unit division device in each method of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0135] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0136] It should be understood that the processor of the embodiment of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiment can be completed by the hardware integrated logic circuit or software instructions in the processor. The above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor are combined and performed. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, and other mature storage media in the art. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0137] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0138] It should be understood that the above-mentioned memory is exemplary but not restrictive. For example, the memory in the embodiments of the present application may also be static random access memory (static RAM, SRAM), dynamic random access memory (dynamic RAM, DRAM), synchronous dynamic random access memory (synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (double data rate SDRAM, DDR SDRAM), enhanced synchronous dynamic random access memory (enhanced SDRAM, ESDRAM), synchronous link dynamic random access memory (synch link DRAM, SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DR RAM), etc. That is to say, the memory in the embodiments of the present application is intended to include but not limited to these and any other suitable types of memory.

[0139] An embodiment of the present application also provides a computer program product, including a computer program.

[0140] When the computer program is executed by the processor, it implements the corresponding processes implemented by the coding unit division device in each method of the embodiments of the present application. For the sake of brevity, it will not be repeated here.

[0141] An embodiment of the present application also provides a computer-readable storage medium for storing a computer program.

[0142] The computer program enables the computer to execute the corresponding processes implemented by the coding unit division device in each method of the embodiments of the present application. For the sake of brevity, they are not repeated here.

[0143] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0144] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0145] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0146] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0147] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0148] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0149] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A coding unit division method, characterized in that: The method comprises: Acquire first information, where the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of a neighboring CU of the first CU, and prediction mode information of the first CU; A first judgment result of the first CU is determined based on the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and a second judgment result of the first CU is determined based on the first information, the second judgment result indicating whether the first CU skips the second division mode.

2. The method according to claim 1, characterized in that The method further comprises: A first data set is obtained, where the first data set includes second information of CUs in a first CU set, where the second information includes one or more of the following information: prediction mode information of the CUs, partition mode information of the CUs, and feature information of the CUs; ​​the CUs in the first CU set are CUs corresponding to a target partition mode obtained after a training video set undergoes a standard video encoding process.

3. The method according to claim 2, characterized in that The method further comprises: Determine first probability information according to the first data set, where the first probability information includes probability information that a CU in the first CU set skips the first partitioning mode and probability information that the CU does not skip the first partitioning mode; Determine conditional probability information according to the first data set, the conditional probability information including conditional probability information that a CU in the first CU set skips the first partitioning mode and conditional probability information that the CU does not skip the first partitioning mode; The first classifier is constructed according to the first probability information and the conditional probability information.

4. The method according to claim 3, characterized in that The method further comprises: Constructing a first feature vector, wherein the first feature vector includes prediction mode information of each CU in the first CU set, partition mode information of adjacent CUs of each CU, and prediction mode information of adjacent CUs of each CU; The determining conditional probability information according to the first data set includes: Conditional probability information corresponding to the first feature vector is determined based on the first data set.

5. The method according to claim 2, characterized in that: The method further comprises: Determine third information according to the first data set, the third information including proportions of different partition modes of CUs of different sizes in the first CU set and a relationship between a partition mode of an upper layer CU and a partition mode of a lower layer CU of each CU in the first CU set and a partition mode of each CU; Determine a first judgment strategy according to the third information, where the first judgment strategy is used to determine whether the first CU skips the second division mode; The determining a second judgment result of the first CU according to the first information includes: A second judgment result of the first CU is determined according to the first information and the first judgment strategy.

6. The method according to any one of claims 1 to 5, characterized in that: The first information further includes size information of the first CU and feature information of the first CU; the method further includes: The third judgment result of the first CU is determined according to the first information and the second judgment strategy; wherein the second judgment strategy is used to determine whether the first CU skips the third division mode, and the third judgment result indicates whether the first CU skips the third division mode.

7. A coding unit division device, characterized in that: The device comprises: an acquiring unit, configured to acquire first information, wherein the first information includes one or more of the following information: information of an upper layer CU of a first coding unit CU, information of a lower layer CU of the first CU, information of a neighboring CU of the first CU, and prediction mode information of the first CU; A determination unit is used to determine a first judgment result of the first CU based on the first classifier and the first information, the first judgment result indicating whether the first CU skips the first division mode; and to determine a second judgment result of the first CU based on the first information, the second judgment result indicating whether the first CU skips the second division mode.

8. A coding unit division device, characterized in that: include: A processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory to execute the method as claimed in any one of claims 1 to 6.

9. A computer program product, characterized in that include: A computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: Used to store a computer program, wherein the computer program causes a computer to execute the method according to any one of claims 1 to 6.

Images