A decoding, encoding method, apparatus and device thereof

Abstract

The application provides a decoding and encoding method, device and equipment. The method comprises the following steps: obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block; decoding a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a corrected residual feature corresponding to the current block; and determining a reconstructed image block corresponding to the current block based on the corrected residual feature. The application improves the encoding performance and the decoding performance.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of coding and decoding, in particular to a decoding method and an encoding method, a decoding device and an encoding device. BACKGROUND

[0002] In order to save space, video images are transmitted after being encoded. Complete video encoding can include prediction, transformation, quantization, entropy encoding, filtering and the like. For the prediction process, the prediction process can include intra prediction and inter prediction. Inter prediction refers to using the correlation in the time domain of a video to predict the current pixel using the pixels of the adjacent encoded image, so as to effectively remove the temporal redundancy of the video. Intra prediction refers to using the correlation in the spatial domain of a video to predict the current pixel using the pixels of the encoded block of the current frame, so as to remove the spatial redundancy of the video.

[0003] With the rapid development of deep learning, deep learning has achieved success in many high-level computer vision problems, such as image classification, object detection, etc. Deep learning has also gradually begun to be applied in the field of coding and decoding, i.e., a neural network can be used to encode and decode images. Although the neural network-based coding and decoding method exhibits great performance potential, the neural network-based coding and decoding method still has problems such as poor coding performance, poor decoding performance, and high complexity. SUMMARY

[0004] Therefore, the present application provides a decoding method and an encoding method, a decoding device and an encoding device, which improve the coding performance and the decoding performance.

[0005] The present application provides a decoding method applied to a decoding end, the method comprising:

[0006] obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block;

[0007] decoding a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a corrected residual feature corresponding to the current block;

[0008] determining a reconstructed image block corresponding to the current block based on the corrected residual feature.

[0009] The present application provides an encoding method applied to an encoding end, the method comprising:

[0010] correcting a residual feature corresponding to a current block to obtain a corrected residual feature;

[0011] obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block;

[0012] Encode the modified residual feature based on the target probability distribution model to obtain a second code stream corresponding to the current block, and send the second code stream to a decoding end.

[0013] The present application provides a decoding device, the device comprises:

[0014] a memory configured to store video data;

[0015] a decoder configured to implement:

[0016] obtain a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block;

[0017] decode a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a modified residual feature corresponding to the current block;

[0018] determine a reconstructed image block corresponding to the current block based on the modified residual feature.

[0019] The present application provides an encoding device, the device comprises:

[0020] a memory configured to store video data;

[0021] an encoder configured to implement:

[0022] modify a residual feature corresponding to a current block to obtain a modified residual feature;

[0023] obtain a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block;

[0024] encode the modified residual feature based on the target probability distribution model to obtain a second code stream corresponding to the current block, and send the second code stream to a decoding end.

[0025] The present application provides a decoding end device, comprising: a processor and a machine readable storage medium, the machine readable storage medium stores machine executable instructions that can be executed by the processor;

[0026] The processor is used to execute machine executable instructions to implement the above decoding method.

[0027] The present application provides an encoding end device, comprising: a processor and a machine readable storage medium, the machine readable storage medium stores machine executable instructions that can be executed by the processor;

[0028] The processor is used to execute machine executable instructions to implement the above encoding method.

[0029] From the above technical solutions, in the embodiments of the present application, the target probability distribution model corresponding to the first scaling factor is obtained, the second code stream corresponding to the current block is decoded based on the target probability distribution model corresponding to the first scaling factor to obtain the residual feature corresponding to the current block, and the reconstructed image block corresponding to the current block is determined based on the residual feature. When the target probability distribution model is determined, the change of the residual can be fully perceived, and the change of the residual is applied to the target probability distribution model, so that the probability distribution describing the residual is more accurate, and the purpose of reducing the code rate is achieved. The parameters of the target probability distribution model used for encoding the residual and the residual are adjusted synchronously, the scaling factor of the residual on the encoding end acts on the original probability distribution model to obtain the target probability distribution model, and when encoding and decoding are performed based on the target probability distribution model, the encoding performance and the decoding performance can be improved, and the complexity is low. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 is a schematic diagram of a video encoding framework;

[0031] Figure 2 is a flowchart of a decoding method in an embodiment of the present application;

[0032] Figure 3 is a flowchart of an encoding method in an embodiment of the present application;

[0033] Figure 4 is a schematic diagram of a processing process of an encoding end in an embodiment of the present application;

[0034] Figure 5 is a schematic diagram of a G-Unit operation in an embodiment of the present application;

[0035] Figure 6 is a schematic diagram of a processing process of a decoding end in an embodiment of the present application;

[0036] Figure 7 is a schematic diagram of a processing process of an encoding end in an embodiment of the present application;

[0037] Figure 8 is a flowchart of a decoding method in an embodiment of the present application;

[0038] Figure 9A and Figure 9B is a schematic diagram of correction of an initial standard deviation in an embodiment of the present application;

[0039] Figure 10A is a hardware structure diagram of a decoding end device in an embodiment of the present application;

[0040] Figure 10B is a hardware structure diagram of an encoding end device in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The terminology used in the embodiments of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. The singular forms "a," "an," and "the" as used in the embodiments of the present application and the claims are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be understood that the terms "and / or," "comprises" and "comprising," when used in the embodiments of the present application and the claims, are open-ended transitions that specify the presence of zero, one, two or more of the linked listing items but do not exclude the presence of additional items. It will be understood that, although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a temporal sequence, but are used only as labels for the sake of clarity. For example, a first item can be termed a second item, and similarly, a second item can be termed a first item, depending on the context, without departing from the scope of the embodiments of the present application. Furthermore, the word "if' can be interpreted to mean "when" or "upon" or "in response to determining" depending on the context.

[0042] The embodiments of the present application propose a decoding and encoding method, which can involve the following concepts:

[0043] Neural Network (NN): Neural network refers to an artificial neural network. The neural network is an operation model composed of a large number of nodes (or neurons) connected to each other. In the neural network, the neuron processing unit can represent different objects, such as features, letters, concepts, or some meaningful abstract patterns. The types of processing units in the neural network can be divided into three categories: input units, output units, and hidden units. The input units accept signals and data from the external world; the output units output the processing results; and the hidden units are the units between the input and output units, which cannot be observed from the outside of the system. The connection weights between neurons reflect the connection strength between units, and the representation and processing of information are embodied in the connection relationship of processing units. The neural network is a non-programmed, brain-like information processing method. The essence of the neural network is to obtain a parallel distributed information processing function through the transformation and dynamics of the neural network, and to simulate the information processing function of the neural system of the human brain to different degrees and levels. In the field of video processing, the commonly used neural networks can include but are not limited to convolutional neural networks (CNN), recurrent neural networks (RNN), fully connected networks, etc.

[0044] Convolutional Neural Network (CNN): Convolutional Neural Network is a kind of feedforward neural network, which is one of the most representative network structures in deep learning technology. The artificial neurons of the convolutional neural network can respond to a part of the surrounding units within the coverage range, and have excellent performance for large image processing. The basic structure of the convolutional neural network includes two layers. One is the feature extraction layer (also known as the convolution layer), and the input of each neuron is connected to the local receptive field of the previous layer, and the local features are extracted. Once the local features are extracted, the positional relationship between them and other features is also determined. The second is the feature mapping layer (also known as the activation layer), and each calculation layer of the neural network is composed of multiple feature mappings. Each feature mapping is a plane, and all the weights of the neurons on the plane are equal. The feature mapping structure can use Sigmoid function, ReLU function, Leaky-ReLU function, PReLU function, GDN function, etc. as the activation function of the convolutional network. In addition, since the neurons on a mapping plane share weights, the number of free parameters of the network is reduced.

[0045] For example, one of the advantages of the convolutional neural network compared to the image processing algorithm is that it avoids the complex pre-processing process of the image (extracting artificial features, etc.), and can directly input the original image for end-to-end learning. One of the advantages of the convolutional neural network compared to the ordinary neural network is that the ordinary neural network uses full connection, that is, all the neurons from the input layer to the hidden layer are connected. This will result in a large number of parameters, making the network training time-consuming and even difficult to train. The convolutional neural network avoids this difficulty through local connection and weight sharing.

[0046] Deconvolution: Deconvolution is also called transposed convolution layer. The working process of the deconvolution layer is similar to that of the convolution layer, and the main difference is that the deconvolution layer will be padded to make the output larger than the input (of course, it can also be kept the same). If the stride is 1, the output size is equal to the input size. If the stride is N, the width of the output feature is N times the width of the input feature, and the height of the output feature is N times the height of the input feature.

[0047] Generalization Ability: Generalization ability can refer to the adaptability of a machine learning algorithm to new samples. The purpose of learning is to learn the rules hidden in the data pairs. The trained network can also give appropriate output for data outside the learning set with the same rules, and this ability can be called generalization ability.

[0048] Rate-Distortion Optimized (RDBEM) principle: Two main metrics for evaluating coding efficiency are bitrate and PSNR (Peak Signal-to-Noise Ratio). A smaller bitrate results in a higher compression ratio, and a higher PSNR leads to better reconstructed image quality. In mode selection, the decision formula essentially evaluates both factors. For example, the cost of a mode is: J(mode) = D + λ*R, where D represents Distortion, typically measured using the SSE metric (Sum of Mean Squares of Differences between the Reconstructed Image Patch and the Source Image). Alternatively, the SAD metric (Sum of Absolute Differences between the Reconstructed Image Patch and the Source Image) can be used to consider the cost. λ is a Lagrange multiplier, and R is the actual number of bits required to encode the image patch in that mode, including the total number of bits needed for encoding mode information, motion information, residuals, etc. Using the RDBEM principle to compare and decide on coding modes during mode selection usually ensures optimal coding performance.

[0049] Entropy coding: Entropy coding is a coding process that follows the principle of entropy without losing any information. Information entropy is the average amount of information in the source (a measure of uncertainty). For example, common entropy coding methods include, but are not limited to: Shannon coding, Huffman coding, and arithmetic coding.

[0050] Video coding framework: see Figure 1 The diagram shown is a schematic of the video encoding framework at the encoding end. This video encoding framework can be used to implement the encoding end processing flow of the embodiments of this application. The schematic diagram of the video decoding framework can be compared with... Figure 1 Similarly, without repeating the details here, the decoding end processing flow of the embodiments of this application can be implemented using a video decoding framework.

[0051] For example, see Figure 1 As shown, a video coding framework can include modules such as prediction, transform, quantization, entropy encoder, inverse quantization, inverse transform, reconstruction, and filtering. At the encoding end, the cooperation between these modules enables the processing flow. Similarly, a video decoding framework can include modules such as prediction, transform, quantization, entropy decoder, inverse quantization, inverse transform, reconstruction, and filtering. At the decoding end, the cooperation between these modules enables the processing flow.

[0052] A lot of encoding tools are proposed for each module of the encoding end, and each tool often has multiple modes. For different video sequences, the encoding tool that can obtain the optimal encoding performance is often different. Therefore, in the encoding process, the RDO (Rate-Distortion Optimize) is usually used to compare the encoding performance of different tools or modes to select the optimal mode. After determining the optimal tool or mode, the decision information of the tool or mode is transmitted by encoding the marker information in the bit stream. Although this method brings higher encoding complexity, it can adaptively select the optimal mode combination for different contents to obtain the optimal encoding performance. The decoding end can obtain the related mode information by directly parsing the marker information, and the complexity is less affected.

[0053] The embodiment of the present application proposes a decoding method and an encoding method, which can be decoded and encoded based on a probability distribution model of residual correction, so as to achieve the purpose of reducing the code rate by obtaining a probability distribution model more consistent with the data.

[0054] The decoding method and the encoding method in the embodiment of the present application will be described in detail below in combination with several specific embodiments.

[0055] Embodiment 1: The embodiment of the present application proposes a decoding method, as shown in Figure 2 The method can be applied to the decoding end (also referred to as a video decoder), and the method can include the following steps.

[0056] Step 201: Obtain a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block.

[0057] Step 202: Decode a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a corrected residual feature corresponding to the current block.

[0058] Step 203: Determine a reconstructed image block corresponding to the current block based on the corrected residual feature.

[0059] For example, obtaining the target probability distribution model corresponding to the first scaling factor based on the first code stream corresponding to the current block can include but is not limited to: obtaining target probability distribution parameters corresponding to the first scaling factor based on the first code stream corresponding to the current block, and determining the target probability distribution model based on the target probability distribution parameters. The target probability distribution parameters include but are not limited to a target mean and / or a target standard deviation. Of course, the target mean and the target standard deviation are only examples, and are not limited thereto.

[0060] In a possible implementation, if the target probability distribution parameter is a target standard deviation, obtaining the target probability distribution parameter corresponding to the first scaling factor based on the first code stream corresponding to the current block can include but is not limited to: decoding the first code stream to obtain initial coefficient hyperparameter features corresponding to the current block; performing coefficient hyperparameter feature inverse transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation; and correcting the initial standard deviation based on the first scaling factor to obtain the target standard deviation.

[0061] For example, correcting the initial standard deviation based on the first scaling factor to obtain the target standard deviation can include but is not limited to: generating a target scaling factor based on the first scaling factor; and correcting the initial standard deviation based on the target scaling factor to obtain the target standard deviation. The target scaling factor can be the first scaling factor, or the target scaling factor is determined based on the first scaling factor and a second scaling factor. The second scaling factor is used to represent a mapping relationship between a residual and a standard deviation. The second scaling factor can be pre-configured, or the second scaling factor is decoded from the code stream corresponding to the current block.

[0062] In another possible implementation, if the target probability distribution parameter is a target standard deviation, obtaining the target probability distribution parameter corresponding to the first scaling factor based on the first code stream corresponding to the current block can include but is not limited to: decoding the first code stream to obtain initial coefficient hyperparameter features corresponding to the current block; correcting the initial coefficient hyperparameter features based on the first scaling factor to obtain target coefficient hyperparameter features; and performing coefficient hyperparameter feature inverse transformation on the target coefficient hyperparameter features to obtain the target standard deviation.

[0063] For example, correcting the initial coefficient hyperparameter features based on the first scaling factor to obtain the target coefficient hyperparameter features can include but is not limited to: generating a target scaling factor based on the first scaling factor, and correcting the initial coefficient hyperparameter features based on the target scaling factor to obtain the target coefficient hyperparameter features. The target scaling factor can be the first scaling factor, or the target scaling factor is determined based on the first scaling factor and a third scaling factor. The third scaling factor is used to represent a mapping relationship between a residual and a coefficient hyperparameter feature. The third scaling factor can be pre-configured, or the third scaling factor is decoded from the code stream corresponding to the current block.

[0064] In a possible implementation, determining the reconstructed image block corresponding to the current block based on the corrected residual feature can include but is not limited to: correcting the corrected residual feature based on the first scaling factor to obtain a residual feature before correction (i.e., a residual feature before correction at the encoding end), and determining the reconstructed image block corresponding to the current block based on the residual feature. For example, obtaining a mean value corresponding to the current block, and determining the reconstructed image block corresponding to the current block based on the residual feature and the mean value corresponding to the current block.

[0065] In a possible implementation, the initial coefficient hyperparameter feature corresponding to the current block is obtained by decoding the first code stream, which can include but is not limited to: the initial coefficient hyperparameter feature corresponding to the current block is obtained by decoding the first code stream to obtain the hyperparameter quantization feature corresponding to the current block, and then dequantizing the hyperparameter quantization feature. Alternatively, the initial coefficient hyperparameter feature corresponding to the current block can be directly obtained by decoding the first code stream without dequantization operation.

[0066] In a possible implementation, the second code stream corresponding to the current block is decoded based on the target probability distribution model corresponding to the first scaling factor to obtain the corrected residual feature corresponding to the current block, which can include but is not limited to: the second code stream corresponding to the current block is decoded based on the target probability distribution model to obtain the residual quantization feature corresponding to the current block, and then dequantizing the residual quantization feature to obtain the corrected residual feature corresponding to the current block. Alternatively, the corrected residual feature corresponding to the current block can be directly obtained by decoding the second code stream corresponding to the current block based on the target probability distribution model.

[0067] For example, the above execution order is only an example given for convenience of description, and in actual application, the execution order between steps can also be changed, and the execution order is not limited. Moreover, in other embodiments, the steps of the corresponding method are not necessarily performed in the order shown and described in the specification, and the steps included in the method can be more or less than those described in the specification. In addition, a single step described in the specification can be divided into multiple steps for description in other embodiments; multiple steps described in the specification can also be combined into a single step for description in other embodiments.

[0068] As can be seen from the above technical solutions, in the embodiments of the present application, the target probability distribution model corresponding to the first scaling factor is obtained, the second code stream corresponding to the current block is decoded based on the target probability distribution model corresponding to the first scaling factor to obtain the residual feature corresponding to the current block, and the reconstructed image block corresponding to the current block is determined based on the residual feature. When the target probability distribution model is determined, the change of the residual can be fully perceived, and the change of the residual is applied to the target probability distribution model, so that the probability distribution describing the residual is more accurate, and the purpose of reducing the code rate is achieved. The parameters of the target probability distribution model used for encoding the residual and the residual are adjusted synchronously, the scaling factor of the residual on the encoding end acts on the original probability distribution model to obtain the target probability distribution model, and when encoding and decoding based on the target probability distribution model, the encoding performance and decoding performance can be improved, and the complexity is low.

[0069] Embodiment 2: In the embodiments of the present application, an encoding method is proposed, which is described with reference to Figure 3As shown, a flowchart of the encoding method is shown. The method can be applied to an encoding end (also referred to as a video encoder). The method can include the following steps:

[0070] In step 301, the residual feature corresponding to the current block is modified to obtain a modified residual feature.

[0071] In step 302, a target probability distribution model corresponding to the first scaling factor is obtained based on the first code stream corresponding to the current block.

[0072] In step 303, the modified residual feature is encoded based on the target probability distribution model corresponding to the first scaling factor to obtain a second code stream corresponding to the current block, and the second code stream is sent to the decoding end.

[0073] For example, the residual feature corresponding to the current block is modified to obtain a modified residual feature, which can include but is not limited to: the residual feature corresponding to the current block is modified based on the first scaling factor to obtain a modified residual feature.

[0074] For example, the target probability distribution model corresponding to the first scaling factor is obtained based on the first code stream corresponding to the current block, which can include but is not limited to: the target probability distribution parameters corresponding to the first scaling factor are obtained based on the first code stream corresponding to the current block, and the target probability distribution model corresponding to the first scaling factor is determined based on the target probability distribution parameters.

[0075] For example, before the target probability distribution model corresponding to the first scaling factor is obtained based on the first code stream corresponding to the current block, the current block can be transformed to obtain an image feature corresponding to the current block, and the image feature can be transformed to obtain an initial coefficient hyperparameter feature corresponding to the current block. The initial coefficient hyperparameter feature is encoded to obtain the first code stream.

[0076] For example, the modified residual feature is encoded based on the target probability distribution model corresponding to the first scaling factor to obtain a second code stream corresponding to the current block, which can include but is not limited to: the modified residual feature is quantized to obtain a residual quantized feature corresponding to the current block, and the residual quantized feature is encoded based on the target probability distribution model corresponding to the first scaling factor to obtain a second code stream corresponding to the current block. Alternatively, the modified residual feature can be directly encoded based on the target probability distribution model corresponding to the first scaling factor to obtain a second code stream corresponding to the current block, without involving quantization operation.

[0077] Exemplarily, encoding the initial coefficient hyper-parameter feature to obtain the first bitstream can include but is not limited to: quantizing the initial coefficient hyper-parameter feature to obtain a hyper-parameter quantized feature corresponding to the current block, and encoding the hyper-parameter quantized feature to obtain the first bitstream. Alternatively, the initial coefficient hyper-parameter feature can be directly encoded to obtain the first bitstream without involving the quantization operation.

[0078] Exemplarily, when the target probability distribution parameter corresponding to the first scaling factor is obtained based on the first bitstream corresponding to the current block, the target probability distribution parameter includes but is not limited to the target mean and / or the target standard deviation. Of course, the target mean and the target standard deviation are only examples, and are not limited thereto. If the target probability distribution parameter is the target standard deviation, obtaining the target probability distribution parameter corresponding to the first scaling factor based on the first bitstream corresponding to the current block can include but is not limited to: decoding the first bitstream to obtain the initial coefficient hyper-parameter feature corresponding to the current block; performing coefficient hyper-parameter feature inverse transformation on the initial coefficient hyper-parameter feature to obtain an initial standard deviation; and correcting the initial standard deviation based on the first scaling factor to obtain the target standard deviation.

[0079] Exemplarily, correcting the initial standard deviation based on the first scaling factor to obtain the target standard deviation can include but is not limited to: generating a target scaling factor based on the first scaling factor; and correcting the initial standard deviation based on the target scaling factor to obtain the target standard deviation. The target scaling factor can be the first scaling factor; or the target scaling factor is determined based on the first scaling factor and a second scaling factor; wherein the second scaling factor is used to represent the mapping relationship between the residual and the standard deviation.

[0080] In another possible implementation, if the target probability distribution parameter is the target standard deviation, obtaining the target probability distribution parameter corresponding to the first scaling factor based on the first bitstream corresponding to the current block can include but is not limited to: decoding the first bitstream to obtain the initial coefficient hyper-parameter feature corresponding to the current block; correcting the initial coefficient hyper-parameter feature based on the first scaling factor to obtain a target coefficient hyper-parameter feature; and performing coefficient hyper-parameter feature inverse transformation on the target coefficient hyper-parameter feature to obtain the target standard deviation.

[0081] Exemplarily, correcting the initial coefficient hyper-parameter feature based on the first scaling factor to obtain the target coefficient hyper-parameter feature can include but is not limited to: generating a target scaling factor based on the first scaling factor, and correcting the initial coefficient hyper-parameter feature based on the target scaling factor to obtain the target coefficient hyper-parameter feature. The target scaling factor can be the first scaling factor; or the target scaling factor can be determined based on the first scaling factor and a third scaling factor; wherein the third scaling factor is used to represent the mapping relationship between the residual and the coefficient hyper-parameter feature.

[0082] For example, the execution sequence described above is only an example given for the convenience of description, and in actual application, the execution sequence between steps can also be changed, and the execution sequence is not limited. Moreover, in other embodiments, the steps of the corresponding method can not necessarily be executed in the order shown and described in the specification, and the steps included in the method can be more or less than those described in the specification. In addition, a single step described in the specification can be divided into multiple steps for description in other embodiments; multiple steps described in the specification can also be combined into a single step for description in other embodiments.

[0083] From the above technical solutions, in the embodiments of the present application, when determining the target probability distribution model, the change of the residual can be fully perceived, and the change of the residual is applied to the target probability distribution model, so that the probability distribution describing the residual is more accurate, and the purpose of reducing the code rate is achieved. The parameters of the target probability distribution model used for encoding the residual and the residual are adjusted synchronously, the scaling factor of the residual on the encoding end acts on the original probability distribution model to obtain the target probability distribution model, and when encoding and decoding based on the target probability distribution model, the encoding performance and the decoding performance can be improved, and the complexity is low.

[0084] Embodiment 3: For embodiments 1 and 2, the processing process of the encoding end can refer to Figure 4 of course, Figure 4 is only an example of the processing process of the encoding end, and the processing process of the encoding end is not limited.

[0085] After the encoding end obtains the current block x, the feature transformation of the current block x (that is, the analysis transformation of the current block x) can be performed to obtain the image feature y corresponding to the current block x. Wherein, the feature transformation of the current block x refers to: transforming the current block x into the image feature y in the latent domain, so as to facilitate the operation of all subsequent processes in the latent domain.

[0086] After the encoding end obtains the image feature y, the coefficient hyperparameter feature transformation of the image feature y is performed to obtain the initial coefficient hyperparameter feature z. For example, the image feature y can be input to the super encoding network, and the super encoding network is used to perform the coefficient hyperparameter feature transformation on the image feature y to obtain the initial coefficient hyperparameter feature z. Wherein, the super encoding network can be a trained neural network, and the training process of the super encoding network is not limited, as long as it can perform the coefficient hyperparameter feature transformation on the image feature y. Wherein, the image feature y in the latent domain obtains the super-prior latent information z after passing through the super encoding network.

[0087] After the encoding end obtains the initial coefficient hyperparameter feature z, the initial coefficient hyperparameter feature z can be quantized to obtain the hyperparameter quantization feature corresponding to the initial coefficient hyperparameter feature z, that is, Figure 4Q in the above formula represents a quantization process. After obtaining the hyper-parameter quantization feature corresponding to the initial coefficient hyper-parameter feature z, the hyper-parameter quantization feature is encoded to obtain Bitstream#1 (i.e., a first code stream) corresponding to the current block, that is, Figure 4 The AE in the above formula represents an encoding process, such as an entropy encoding process. Alternatively, the encoding end can directly encode the initial coefficient hyper-parameter feature z to obtain Bitstream#1 corresponding to the current block. The hyper-parameter quantization feature or the initial coefficient hyper-parameter feature z carried in Bitstream#1 is mainly used to obtain the parameters of the mean value and the probability distribution model.

[0088] After obtaining Bitstream#1 corresponding to the current block, the encoding end can send Bitstream#1 corresponding to the current block to the decoding end. For the processing process of the decoding end with respect to Bitstream#1 corresponding to the current block, see subsequent embodiments.

[0089] After obtaining Bitstream#1 corresponding to the current block, the encoding end can also decode Bitstream#1 to obtain the hyper-parameter quantization feature, that is, Figure 4 AD in the above formula represents a decoding process, and then the hyper-parameter quantization feature is dequantized to obtain the initial coefficient hyper-parameter feature z_hat. The initial coefficient hyper-parameter feature z_hat can be the same as or different from the initial coefficient hyper-parameter feature z, Figure 4 The IQ in the above formula represents a dequantization process. Alternatively, after obtaining Bitstream#1 corresponding to the current block, the encoding end can also decode Bitstream#1 to obtain the initial coefficient hyper-parameter feature z_hat, without involving the dequantization process.

[0090] After obtaining the initial coefficient hyper-parameter feature z_hat, the encoding end can perform context-based prediction based on the initial coefficient hyper-parameter feature z_hat of the current block and the residual feature y_hat of the previous image block (the determination process of the residual feature y_hat is described in subsequent embodiments) to obtain the mean value mu of the current block. The prediction process is not limited. For the context-based prediction process, the inputs include the initial coefficient hyper-parameter feature z_hat and the decoded residual feature y_hat, and the two are jointly input to obtain a more accurate mean value mu, which is used to obtain the residual by subtracting the original feature and obtain the reconstruction y by adding the decoded residual.

[0091] After obtaining the image feature y, the encoding end determines the residual feature r based on the image feature y and the mean value mu. The residual feature r can be understood as the residual feature r before correction, for example, the difference between the image feature y and the mean value mu is taken as the residual feature r.

[0092] After obtaining the residual feature r, the encoding end can correct the residual feature r to obtain a corrected residual feature r', for example, the encoding end corrects the residual feature r based on a first scaling factor to obtain the corrected residual feature r'. For example, the G-Unit operation can be a scaling operation, and the encoding end can perform a scaling operation (i.e., calculate the product of the first scaling factor and the residual feature r) on the residual feature r based on the first scaling factor to obtain the corrected residual feature r'. Referring to Figure 5 As shown in FIG. 6, the input of the G-Unit operation is the residual feature r, and the output is the residual feature r' after the scaling operation. The first scaling factor can be scale, that is, the residual feature r is scaled by the first scaling factor to obtain the corrected residual feature r'.

[0093] From Figure 4 It can be seen that the residual feature r is subjected to the G-Unit operation before being coded by Bitstream#2. The G-Unit operation and the residual feature r have a relationship as shown in Figure 5 As shown in FIG. 6, the G-Unit includes a trainable matrix, which is denoted as the first scaling factor scale. C represents the number of channels of the latent domain feature, and represents the channel-level product of the residual feature r and the first scaling factor scale, that is, the residual feature r is multiplied by the correction value in the first scaling factor scale corresponding to the channel vector.

[0094] As can be seen from the above, the first scaling factor scale is a correction parameter of the residual feature r. The first scaling factor scale is used to correct the residual feature r to make the corrected residual feature r' more accurate and reliable. The first scaling factor scale can be pre-configured or trained, and the source of the first scaling factor scale is not limited.

[0095] After obtaining the residual feature r', the encoding end can quantize the residual feature r' to obtain a residual quantized feature corresponding to the residual feature r', that is, Figure 4 The Q operation in the above formula (1) is a quantization process. After obtaining the residual quantized feature corresponding to the residual feature r', the encoding end can encode the residual quantized feature to obtain the Bitstream#2 (i.e., the second code stream) corresponding to the current block, that is, Figure 4 The AE operation in the above formula (2) represents an encoding process, such as an entropy encoding process. Alternatively, the encoding end can directly encode the residual feature r' to obtain the Bitstream#2 corresponding to the current block, without involving the quantization process of the residual feature r'.

[0096] After obtaining the Bitstream#2 corresponding to the current block, the encoding end can send the Bitstream#2 corresponding to the current block to the decoding end. For the processing process of the decoding end for the Bitstream#2 corresponding to the current block, see the subsequent embodiments.

[0097] After obtaining the Bitstream#2 corresponding to the current block, the encoding end can also decode the Bitstream#2 to obtain the residual quantization feature, that is, Figure 4 AD in the above formula represents the decoding process. Then, the encoding end can dequantize the residual quantization feature to obtain the modified residual feature r'_hat. The residual feature r'_hat can be the same as or different from the residual feature r'. Figure 4 The IQ operation in the above formula is the dequantization process. Alternatively, after obtaining the Bitstream#2 corresponding to the current block, the encoding end can also decode the Bitstream#2 to obtain the modified residual feature r'_hat, without involving the dequantization process of the residual feature.

[0098] After obtaining the modified residual feature r'_hat, the encoding end can modify the modified residual feature r'_hat to obtain the modified residual feature r_hat. The residual feature r_hat can be the same as or different from the residual feature r. For example, the encoding end modifies the modified residual feature r'_hat based on a first scaling factor to obtain the modified residual feature r_hat. For example, the IG-Unit operation can be a scaling operation, and the encoding end can perform a scaling operation (i.e., calculate the quotient of the modified residual feature r'_hat and the first scaling factor) on the modified residual feature r'_hat based on the first scaling factor to obtain the modified residual feature r_hat. The IG-Unit process can be understood as the inverse process of the G-Unit process, which is used to modify the modified residual feature r'_hat to the modified residual feature r_hat.

[0099] After obtaining the residual feature r_hat, the encoding end determines the image feature y_hat based on the residual feature r_hat and the mean value mu. The image feature y_hat can be the same as or different from the image feature y, for example, the sum of the residual feature r_hat and the mean value mu is taken as the image feature y_hat. After obtaining the image feature y_hat, the image feature y_hat can be subjected to a synthesis transform to obtain the reconstructed image block x_hat corresponding to the current block x. Thus, the image reconstruction process is completed, and the reconstructed image block x_hat is obtained.

[0100] In one possible implementation, when the encoder encodes the residual quantization feature or residual feature r' to obtain Bitstream#2 corresponding to the current block, the encoder needs to first determine the probability distribution model, and then encode the residual quantization feature or residual feature r' based on the probability distribution model. Furthermore, when the encoder decodes Bitstream#2, it also needs to first determine the probability distribution model, and then decode Bitstream#2 based on the probability distribution model.

[0101] To obtain the probability distribution model, which is related to the mean and standard deviation, if the mean of the probability distribution model is 0, then only the standard deviation of the probability distribution model needs to be determined, and then the probability distribution model is determined based on the standard deviation. If the mean of the probability distribution model is not 0, then the mean and standard deviation of the probability distribution model need to be determined, and then the probability distribution model is determined based on the mean and standard deviation. In this embodiment, there are no restrictions on how to determine the probability distribution model.

[0102] To obtain the standard deviation of the probability distribution model, please refer to [link / reference needed]. Figure 4 As shown, after obtaining the initial hyperparameter features z_hat, the encoder performs an inverse hyperparameter transformation on z_hat to obtain the standard deviation σ. Then, based on the standard deviation, the probability distribution model N(0, σ) is determined. Here, we take a probability distribution model with a mean of 0 as an example. If the mean of the probability distribution model is not 0, the mean of the probability distribution model also needs to be determined. For example, the initial hyperparameter features z_hat can be input into a superscale decoding network, which performs an inverse hyperparameter transformation on z_hat to obtain the standard deviation. The superscale decoding network can be a trained neural network; the training process of this superscale decoding network is not restricted, as long as it can perform the inverse hyperparameter transformation on z_hat.

[0103] In one possible implementation, the above-mentioned encoding end processing is performed by a deep learning model or a neural network model to achieve end-to-end image compression and encoding, and no restrictions are placed on this process.

[0104] Example 4: For Examples 1 and 2, the processing procedure at the decoding end can be found in [reference needed]. Figure 6 As shown, of course, Figure 6 This is just one example of the processing procedure at the decoding end, and no restrictions are imposed on the processing procedure at the decoding end.

[0105] After obtaining Bitstream#1 corresponding to the current block, the decoding end can decode Bitstream#1 to obtain the hyperparameter quantization characteristics, i.e. Figure 6AD in the figure represents a decoding process, and then the hyper-quantized feature is dequantized to obtain the initial coefficient hyper-feature z_hat, which can be the same as or different from the initial coefficient hyper-feature z, Figure 6 The IQ operation in the figure is a dequantization process. Alternatively, the decoding end can also decode Bitstream#1 corresponding to the current block after obtaining Bitstream#1, to obtain the initial coefficient hyper-feature z_hat, without involving the dequantization process.

[0106] After the decoding end obtains the initial coefficient hyper-feature z_hat, the initial coefficient hyper-feature z_hat of the current block and the residual feature y_hat of the previous image block (the determination process of the residual feature y_hat is described in subsequent embodiments) can be used to perform context-based prediction to obtain the mean value mu of the current block, and the prediction process is not limited. For the context-based prediction process, the inputs include the initial coefficient hyper-feature z_hat and the decoded residual feature y_hat, and the two are jointly input to obtain a more accurate mean value mu, which is used to obtain the residual by subtracting the original feature and is added to the decoded residual to obtain the reconstruction y.

[0107] After the decoding end obtains Bitstream#2 corresponding to the current block, Bitstream#2 can also be decoded to obtain the residual quantized feature, that is, Figure 6 AD in the figure represents a decoding process, and then the decoding end can dequantize the residual quantized feature to obtain the modified residual feature r'_hat, which can be the same as or different from the residual feature r', Figure 6 The IQ operation in the figure is a dequantization process. Alternatively, the decoding end can also decode Bitstream#2 corresponding to the current block after obtaining Bitstream#2, to obtain the modified residual feature r'_hat, without involving the dequantization process of the residual feature.

[0108] After the decoding end obtains the modified residual feature r'_hat, the modified residual feature r'_hat can be modified to obtain the modified residual feature r_hat, which can be the same as or different from the residual feature r. For example, the decoding end modifies the residual feature r'_hat based on the first scaling factor to obtain the modified residual feature r_hat. For example, the IG-Unit operation can be a scaling operation, and the decoding end can perform a scaling operation (i.e., calculate the quotient of the residual feature r'_hat and the first scaling factor) on the residual feature r'_hat based on the first scaling factor to obtain the modified residual feature r_hat. The IG-Unit process can be understood as the inverse process of the G-Unit process, which is used to modify the residual feature r'_hat to the residual feature r_hat.

[0109] After obtaining the residual feature r_hat, the decoder determines the image feature y_hat based on the residual feature r_hat and the mean mu. The image feature y_hat and the image feature y can be the same or different. For example, the sum of the residual feature r_hat and the mean mu can be used as the image feature y_hat. After obtaining the image feature y_hat, a synthetic transformation can be performed on the image feature y_hat to obtain the reconstructed image block x_hat corresponding to the current block x. At this point, the image reconstruction process is completed, and the reconstructed image block x_hat is obtained.

[0110] When decoding Bitstream#2, the decoding end needs to first determine the probability distribution model, and then decode Bitstream#2 based on this probability distribution model. To obtain the probability distribution model, it is related to the mean and standard deviation. If the mean of the probability distribution model is 0, then only the standard deviation needs to be determined, and the probability distribution model is then determined based on the standard deviation. If the mean of the probability distribution model is not 0, then both the mean and standard deviation need to be determined, and the probability distribution model is then determined based on the mean and standard deviation. This embodiment does not impose restrictions on how the probability distribution model is determined.

[0111] To obtain the standard deviation of the probability distribution model, please refer to [link / reference needed]. Figure 6 As shown, after obtaining the initial hyperparameter features z_hat, the decoder performs an inverse hyperparameter transformation on z_hat to obtain the standard deviation σ. Then, based on the standard deviation, the probability distribution model N(0, σ) is determined. Here, we take a probability distribution model with a mean of 0 as an example. If the mean of the probability distribution model is not 0, then the mean of the probability distribution model also needs to be determined. For example, the initial hyperparameter features z_hat can be input into a superscale decoding network, which then performs an inverse hyperparameter transformation on z_hat to obtain the standard deviation. The superscale decoding network can be a trained neural network; the training process of this superscale decoding network is not restricted, as long as it can perform the inverse hyperparameter transformation on z_hat.

[0112] In one possible implementation, the above-mentioned decoding process is executed by a deep learning model or a neural network model to achieve end-to-end image decompression and decoding, and no restrictions are placed on this process.

[0113] Example 5: Regarding Example 3, when determining the probability distribution model, in order to match the probability distribution model with the G-Unit process, that is, to enable the probability distribution model to perceive the scaling effect of the G-Unit process on the residual features, so that there is no significant deviation between the probability distribution model and the actual residual features to be encoded, in this example, the target probability distribution model corresponding to the first scaling factor (i.e., the first scaling factor used in the G-Unit process) can also be obtained. That is, the probability distribution model is modified based on the first scaling factor to obtain the modified target probability distribution model. The residual quantization feature or residual feature r' is encoded based on the target probability distribution model, and Bitstream#2 is decoded based on the target probability distribution model.

[0114] For example, see Figure 7 As shown, after obtaining the current block x, the encoder can perform feature transformation (i.e., analyze and transform the current block x) to obtain the image feature y corresponding to the current block x. After obtaining the image feature y, the encoder performs coefficient hyperparameter feature transformation on the image feature y to obtain the initial coefficient hyperparameter feature z. After obtaining the initial coefficient hyperparameter feature z, the encoder can quantize the initial coefficient hyperparameter feature z to obtain the hyperparameter quantization feature corresponding to the initial coefficient hyperparameter feature z. The hyperparameter quantization feature is then encoded to obtain the Bitstream#1 (i.e., the first bitstream) corresponding to the current block. Alternatively, the encoder can directly encode the initial coefficient hyperparameter feature z to obtain the Bitstream#1 corresponding to the current block.

[0115] After obtaining Bitstream#1 corresponding to the current block, the encoder can also decode Bitstream#1 to obtain the hyperparameter quantization feature, and then dequantize the hyperparameter quantization feature to obtain the initial coefficient hyperparameter feature z_hat. Alternatively, the encoder can also decode Bitstream#1 to obtain the initial coefficient hyperparameter feature z_hat without involving the dequantization process.

[0116] After obtaining the initial coefficient hyperparameter feature z_hat, the encoder can perform context-based prediction based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain the mean mu corresponding to the current block.

[0117] After obtaining the image features y, the encoder determines the residual features r based on the image features y and the mean mu. The encoder then corrects the residual features r to obtain the corrected residual features r', i.e. Figure 7 The residual scaling process in the process.

[0118] After obtaining the residual feature r', the encoding end can quantize the residual feature r' to obtain a residual quantized feature corresponding to the residual feature r', encode the residual quantized feature to obtain a Bitstream#2 (i.e., a second code stream) corresponding to the current block, or directly encode the residual feature r' to obtain the Bitstream#2 corresponding to the current block.

[0119] After obtaining the Bitstream#2 corresponding to the current block, the encoding end can further decode the Bitstream#2 to obtain a residual quantized feature, dequantize the residual quantized feature to obtain a corrected residual feature r'_hat, or decode the Bitstream#2 to obtain the corrected residual feature r'_hat after obtaining the Bitstream#2 corresponding to the current block.

[0120] After obtaining the corrected residual feature r'_hat, the encoding end can further correct the corrected residual feature r'_hat to obtain a residual feature r_hat before correction, i.e., the residual inverse scaling process shown in the following formula (3). Figure 7

[0121] After obtaining the residual feature r_hat, the encoding end determines an image feature y_hat based on the residual feature r_hat and the mean value mu, and performs a synthesis transform on the image feature y_hat to obtain a reconstructed image block x_hat corresponding to the current block x.

[0122] In a possible implementation, when the encoding end encodes the residual quantized feature or the residual feature r' to obtain the Bitstream#2 corresponding to the current block, the encoding end needs to first determine a probability distribution model, and then encodes the residual quantized feature or the residual feature r' based on the probability distribution model. In addition, when the encoding end decodes the Bitstream#2, the encoding end also needs to first determine a probability distribution model, and then decodes the Bitstream#2 based on the probability distribution model.

[0123] ​In this embodiment, to ensure the probability distribution model matches the G-Unit process—that is, to ensure the probability distribution model is aware of the scaling effect of the G-Unit process on the residual features, and that there is no significant deviation between the probability distribution model and the actual residual features to be encoded—the probability distribution model PDM can be modified to obtain the modified probability distribution model PDM_hat. This modified probability distribution model PDM_hat can be referred to as the target probability distribution model. Specifically, to ensure the probability distribution model matches the G-Unit process, a first scaling factor (i.e., the scaling factor of the G-Unit process) can be used to modify the probability distribution model PDM, resulting in the modified target probability distribution model PDM_hat.

[0124] In this embodiment, see Figure 7 As shown, the inputs to the probability model correction module are mainly: the first scaling factor (scale) of the residual features at the encoder and the original probability distribution model confirmed by Bitstream#1 decoding. The output is the adjusted target probability distribution model PDM_hat. The main purpose of the probability model correction module is to optimize the probability distribution model of Bitstream#2's entropy encoding. By adjusting the probability distribution parameters of the probability distribution model, the bitrate of the encoded bitstream is reduced. For example, by correcting the probability distribution model PDM output by the super-prior network, the probability distribution model becomes aware of all scaling changes of the residual features at the encoder. The scaling factor of the residual features and the probability distribution model decoded by Bitstream#2 are jointly input into the probability model correction module, which outputs a new target probability distribution model PDM_hat. The target probability distribution model PDM_hat describes the distribution of the residual features more accurately than the previous probability distribution model, thus achieving the goal of reducing the bitrate.

[0125] For example, since the target probability distribution model PDM_hat is determined based on probability distribution parameters (such as mean and / or standard deviation), the target probability distribution parameters corresponding to the first scaling factor can be obtained, and the target probability distribution model PDM_hat corresponding to the first scaling factor can be determined based on the target probability distribution parameters. For instance, the initial probability distribution parameters can be corrected using the first scaling factor to obtain the corrected target probability distribution parameters.

[0126] For example, if the mean of the probability distribution model is 0, only the standard deviation of the probability distribution model needs to be determined, and then the probability distribution model is determined based on the standard deviation, in which case, the target standard deviation corresponding to the first scaling factor can be obtained, such as the initial standard deviation can be corrected by the first scaling factor to obtain the corrected target standard deviation, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target standard deviation.

[0127] For example, if the mean of the probability distribution model is not 0, the mean and the standard deviation of the probability distribution model need to be determined, and then the probability distribution model is determined based on the mean and the standard deviation, in which case, the target mean and the target standard deviation corresponding to the first scaling factor can be obtained, such as the initial mean can be corrected by the first scaling factor to obtain the corrected target mean, and the initial standard deviation can be corrected by the first scaling factor to obtain the corrected target standard deviation, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target mean and the target standard deviation.

[0128] For example, if the mean of the probability distribution model is not 0, but the standard deviation of the probability distribution model is 0, the mean of the probability distribution model needs to be determined, and then the probability distribution model is determined based on the mean, in which case, the target mean corresponding to the first scaling factor can be obtained, such as the initial mean can be corrected by the first scaling factor to obtain the corrected target mean, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target mean.

[0129] For example, to obtain the mean of the probability distribution model, context-based prediction can be performed based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain the mean mu corresponding to the current block.

[0130] For example, to obtain the standard deviation of the probability distribution model, after obtaining the initial coefficient hyperparameter feature z_hat, the initial coefficient hyperparameter feature z_hat is subjected to coefficient hyperparameter feature inverse transformation to obtain the standard deviation σ corresponding to the current block.

[0131] For example, the encoding end can obtain the target probability distribution parameter corresponding to the first scaling factor, and determine the target probability distribution model corresponding to the first scaling factor based on the target probability distribution parameter. For example, the encoding end can correct the initial probability distribution parameter based on the first scaling factor to obtain the target probability distribution parameter corresponding to the first scaling factor.

[0132] In a possible implementation, the target probability distribution parameter can include, but is not limited to, a target mean and / or a target standard deviation. Of course, the target mean and the target standard deviation are only examples, and the target probability distribution parameter is not limited thereto.

[0133] For example, if the target probability distribution parameter is a target standard deviation, the encoding end can decode the first code stream to obtain an initial coefficient hyperparameter feature corresponding to the current block, perform coefficient hyperparameter feature inverse transformation on the initial coefficient hyperparameter feature to obtain an initial standard deviation, and correct the initial standard deviation based on the first scaling factor to obtain the target standard deviation.

[0134] For example, the encoding end can generate a target scaling factor based on the first scaling factor, correct the initial standard deviation based on the target scaling factor to obtain the target standard deviation, for example, take the product of the target scaling factor and the initial standard deviation as the target standard deviation. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and the second scaling factor, for example, determined based on the product of the first scaling factor and the second scaling factor, or a bias value is added to the product value of the first scaling factor and the second scaling factor to obtain the target scaling factor.

[0135] The second scaling factor is used to represent the mapping relationship between the residual and the standard deviation. For example, the residual and the standard deviation are values of different magnitudes, and the second scaling factor can adjust the magnitude of the standard deviation to the magnitude of the residual. That is, the second scaling factor is used as an adjustment coefficient to adjust the magnitude of the standard deviation to the magnitude of the residual, so that the residual and the standard deviation are adjusted to the same magnitude for operation, thereby improving the operation accuracy.

[0136] For example, if the target probability distribution parameter is a target standard deviation, the encoding end can decode the first code stream to obtain an initial coefficient hyperparameter feature corresponding to the current block, correct the initial coefficient hyperparameter feature based on the first scaling factor to obtain a target coefficient hyperparameter feature, and perform coefficient hyperparameter feature inverse transformation on the target coefficient hyperparameter feature to obtain the target standard deviation.

[0137] For example, the encoding end can generate a target scaling factor based on the first scaling factor, and correct the initial coefficient hyperparameter feature based on the target scaling factor to obtain a target coefficient hyperparameter feature. For example, the encoding end can take the product of the target scaling factor and the initial coefficient hyperparameter feature as the target coefficient hyperparameter feature. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and a third scaling factor, for example, the target scaling factor can be determined based on the product of the first scaling factor and the third scaling factor, or a certain bias value is added to the product value of the first scaling factor and the third scaling factor to obtain the target scaling factor. The third scaling factor is used to represent the mapping relationship between the residual and the coefficient hyperparameter feature. For example, the residual and the coefficient hyperparameter feature are values of different magnitudes, and the third scaling factor can adjust the magnitude of the coefficient hyperparameter feature to the magnitude of the residual, that is, the third scaling factor as an adjustment coefficient can adjust the magnitude of the coefficient hyperparameter feature to the magnitude of the residual, so as to adjust the residual and the coefficient hyperparameter feature to the same magnitude for operation, so that the residual and the coefficient hyperparameter feature can better match the operation and improve the operation accuracy.

[0138] For example, if the target probability distribution parameter is a target mean, the encoding end can perform context-based prediction based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain an initial mean corresponding to the current block. The encoding end can correct the initial mean based on the first scaling factor to obtain the target mean.

[0139] For example, the encoding end can generate a target scaling factor based on the first scaling factor, and correct the initial mean based on the target scaling factor to obtain the target mean, for example, taking the product of the target scaling factor and the initial mean as the target mean. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and a second scaling factor, for example, based on the product of the first scaling factor and the second scaling factor, or a certain bias value is added to the product value of the first scaling factor and the second scaling factor to obtain the target scaling factor.

[0140] The second scaling factor is used to represent the mapping relationship between the residual and the mean. For example, the residual and the mean are values of different magnitudes, and the second scaling factor can adjust the magnitude of the mean to the magnitude of the residual, that is, the second scaling factor as an adjustment coefficient can adjust the magnitude of the mean to the magnitude of the residual, so as to adjust the residual and the mean to the same magnitude for operation, so that the residual and the mean can better match the operation and improve the operation accuracy.

[0141] Embodiment 6: For Embodiment 4, when the decoding end determines the probability distribution model, in order to make the probability distribution model match the G-Unit process, that is, the probability distribution model can perceive the scaling effect of the G-Unit process on the residual feature, so that there is no obvious deviation between the probability distribution model and the actual residual feature to be encoded, then the target probability distribution model corresponding to the first scaling factor (that is, the first scaling factor adopted by the G-Unit process) can be obtained, that is, the probability distribution model is modified based on the first scaling factor, and a modified target probability distribution model is obtained, and Bitstream#2 is decoded based on the target probability distribution model.

[0142] For example, after the decoding end obtains Bitstream#1 corresponding to the current block, it can decode Bitstream#1 to obtain the hyperparameter quantized feature, and perform inverse quantization on the hyperparameter quantized feature to obtain the initial coefficient hyperparameter feature z_hat. Alternatively, the decoding end can also decode Bitstream#1 to obtain the initial coefficient hyperparameter feature z_hat without involving the inverse quantization process.

[0143] After the decoding end obtains the initial coefficient hyperparameter feature z_hat, it can perform context-based prediction based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain the mean value mu corresponding to the current block.

[0144] After the decoding end obtains Bitstream#2 corresponding to the current block, it can also decode Bitstream#2 to obtain the residual quantized feature, and perform inverse quantization on the residual quantized feature to obtain the modified residual feature r’_hat. Alternatively, after the decoding end obtains Bitstream#2 corresponding to the current block, it can decode Bitstream#2 to obtain the modified residual feature r’_hat.

[0145] After the decoding end obtains the modified residual feature r’_hat, it can also modify the modified residual feature r’_hat to obtain the modified residual feature r_hat, that is, the residual inverse scaling process shown in the following formula (3). Figure 7

[0146] After the decoding end obtains the residual feature r_hat, it determines the image feature y_hat based on the residual feature r_hat and the mean value mu, and performs synthesis transformation on the image feature y_hat to obtain the reconstructed image block x_hat corresponding to the current block x.

[0147] ​In a possible implementation, when decoding Bitstream#2, the decoding end needs to determine a probability distribution model first, and then decode Bitstream#2 based on the probability distribution model. In order to match the probability distribution model with the G-Unit process, that is, the probability distribution model can perceive the scaling effect of the G-Unit process on the residual feature, so that there is no obvious deviation between the probability distribution model and the actual residual feature to be encoded, the probability distribution model PDM can be corrected to obtain a corrected probability distribution model PDM_hat, and the probability distribution model PDM_hat can be referred to as a target probability distribution model. In order to match the probability distribution model with the G-Unit process, the first scaling factor (the scaling factor of the G-Unit process) can be used to correct the probability distribution model PDM to obtain the corrected target probability distribution model PDM_hat.

[0148] In this embodiment, the input of the probability model correction module mainly includes the first scaling factor scale of the residual feature at the decoding end and the original probability distribution model confirmed by decoding Bitstream#1, and the output is the target probability distribution model PDM_hat after adjustment. The main purpose of the probability model correction module is to optimize the probability distribution model of the entropy coding of Bitstream#2, and to make the code rate of the encoded code stream smaller by adjusting the probability distribution parameters of the probability distribution model. For example, by correcting the probability distribution model PDM output by the hyper-prior network, the probability distribution model can perceive all scaling changes of the residual feature at the decoding end, the scaling factor of the residual feature and the probability distribution model decoded by Bitstream#2 are jointly input into the probability model correction module, and a new target probability distribution model PDM_hat is output. The target probability distribution model PDM_hat is more accurate in describing the distribution of the residual feature than the previous probability distribution model, and can achieve the purpose of reducing the code rate.

[0149] For example, since the target probability distribution model PDM_hat is determined based on the probability distribution parameters (such as mean and / or standard deviation), the target probability distribution parameters corresponding to the first scaling factor can be obtained, and the target probability distribution model PDM_hat corresponding to the first scaling factor can be determined based on the target probability distribution parameters. For example, the initial probability distribution parameters can be corrected by using the first scaling factor to obtain the corrected target probability distribution parameters.

[0150] For example, if the mean of the probability distribution model is 0, only the standard deviation of the probability distribution model needs to be determined, and then the probability distribution model is determined based on the standard deviation, in which case, the target standard deviation corresponding to the first scaling factor can be obtained, such as the initial standard deviation can be corrected by the first scaling factor to obtain the corrected target standard deviation, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target standard deviation.

[0151] For example, if the mean of the probability distribution model is not 0, the mean and the standard deviation of the probability distribution model need to be determined, and then the probability distribution model is determined based on the mean and the standard deviation, in which case, the target mean and the target standard deviation corresponding to the first scaling factor can be obtained, such as the initial mean can be corrected by the first scaling factor to obtain the corrected target mean, and the initial standard deviation can be corrected by the first scaling factor to obtain the corrected target standard deviation, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target mean and the target standard deviation.

[0152] For example, if the mean of the probability distribution model is not 0, but the standard deviation of the probability distribution model is 0, the mean of the probability distribution model needs to be determined, and then the probability distribution model is determined based on the mean, in which case, the target mean corresponding to the first scaling factor can be obtained, such as the initial mean can be corrected by the first scaling factor to obtain the corrected target mean, and then the target probability distribution model PDM_hat corresponding to the first scaling factor is determined based on the target mean.

[0153] For example, to obtain the mean of the probability distribution model, context-based prediction can be performed based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain the mean mu corresponding to the current block.

[0154] For example, to obtain the standard deviation of the probability distribution model, after obtaining the initial coefficient hyperparameter feature z_hat, the initial coefficient hyperparameter feature z_hat is subjected to coefficient hyperparameter feature inverse transformation to obtain the standard deviation σ corresponding to the current block.

[0155] For example, the decoding end can obtain the target probability distribution parameter corresponding to the first scaling factor, and determine the target probability distribution model corresponding to the first scaling factor based on the target probability distribution parameter. For example, the decoding end can correct the initial probability distribution parameter based on the first scaling factor to obtain the target probability distribution parameter corresponding to the first scaling factor.

[0156] In a possible implementation, the target probability distribution parameter can include, but is not limited to, a target mean and / or a target standard deviation. Of course, the target mean and the target standard deviation are only examples, and the target probability distribution parameter is not limited thereto.

[0157] For example, if the target probability distribution parameter is a target standard deviation, the decoding end can decode the first code stream to obtain an initial coefficient hyperparameter feature corresponding to the current block, perform coefficient hyperparameter feature inverse transformation on the initial coefficient hyperparameter feature to obtain an initial standard deviation, and correct the initial standard deviation based on the first scaling factor to obtain the target standard deviation.

[0158] For example, the decoding end can generate a target scaling factor based on the first scaling factor, and correct the initial standard deviation based on the target scaling factor to obtain the target standard deviation, for example, taking the product of the target scaling factor and the initial standard deviation as the target standard deviation. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and the second scaling factor, for example, based on the product of the first scaling factor and the second scaling factor, or adding a certain bias value to the product value of the first scaling factor and the second scaling factor to obtain the target scaling factor.

[0159] The second scaling factor is used to represent the mapping relationship between the residual and the standard deviation. For example, the residual and the standard deviation are values of different magnitudes, and the second scaling factor can adjust the magnitude of the standard deviation to the magnitude of the residual. That is, the second scaling factor is used as an adjustment coefficient to adjust the magnitude of the standard deviation to the magnitude of the residual, so that the residual and the standard deviation are adjusted to the same magnitude for operation, thereby better matching the operation of the residual and the standard deviation and improving the operation accuracy.

[0160] The second scaling factor can be preconfigured, or the second scaling factor can be decoded from the code stream corresponding to the current block, for example, decoded from the first code stream corresponding to the current block, or decoded from the second code stream corresponding to the current block, or decoded from other code streams, and the present disclosure is not limited thereto.

[0161] For example, if the target probability distribution parameter is a target standard deviation, the decoding end can decode the first code stream to obtain an initial coefficient hyperparameter feature corresponding to the current block, correct the initial coefficient hyperparameter feature based on the first scaling factor to obtain a target coefficient hyperparameter feature, and perform coefficient hyperparameter feature inverse transformation on the target coefficient hyperparameter feature to obtain the target standard deviation.

[0162] For example, the decoding end can generate a target scaling factor based on the first scaling factor, and correct the initial coefficient hyperparameter feature based on the target scaling factor to obtain a target coefficient hyperparameter feature. For example, the decoding end can take the product of the target scaling factor and the initial coefficient hyperparameter feature as the target coefficient hyperparameter feature. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and a third scaling factor, for example, the target scaling factor can be determined based on the product of the first scaling factor and the third scaling factor, or a bias value is added to the product value of the first scaling factor and the third scaling factor to obtain the target scaling factor. The third scaling factor is used to represent the mapping relationship between the residual and the coefficient hyperparameter feature. For example, the residual and the coefficient hyperparameter feature are values of different magnitudes, and the third scaling factor can adjust the magnitude of the coefficient hyperparameter feature to the magnitude of the residual, that is, the third scaling factor is used as an adjustment coefficient to adjust the magnitude of the coefficient hyperparameter feature to the magnitude of the residual, so that the residual and the coefficient hyperparameter feature are adjusted to the same magnitude for operation, so that the residual and the coefficient hyperparameter feature can better match the operation and improve the operation accuracy.

[0163] The third scaling factor can be pre-configured, or the third scaling factor can be decoded from the code stream corresponding to the current block, such as decoding the third scaling factor from the first code stream corresponding to the current block, or decoding the third scaling factor from the second code stream corresponding to the current block, or decoding the third scaling factor from other code streams. This is not limited.

[0164] For example, if the target probability distribution parameter is a target mean, the decoding end can perform context-based prediction based on the initial coefficient hyperparameter feature z_hat of the current block and the residual feature y_hat of the previous image block to obtain an initial mean corresponding to the current block. The decoding end can correct the initial mean based on the first scaling factor to obtain the target mean.

[0165] For example, the decoding end can generate a target scaling factor based on the first scaling factor, and correct the initial mean based on the target scaling factor to obtain a target mean, for example, taking the product of the target scaling factor and the initial mean as the target mean. The target scaling factor can be the first scaling factor, or the target scaling factor can be determined based on the first scaling factor and a second scaling factor, for example, based on the product of the first scaling factor and the second scaling factor, or a bias value is added to the product value of the first scaling factor and the second scaling factor to obtain the target scaling factor.

[0166] The second scaling factor is used to represent a mapping relationship between the residual and the mean value. For example, the residual and the mean value are values of different magnitudes. The second scaling factor can adjust the magnitude of the mean value to the magnitude of the residual, that is, the second scaling factor is used as an adjustment coefficient to adjust the magnitude of the mean value to the magnitude of the residual, so that the residual and the mean value are adjusted to the same magnitude for operation, so that the residual and the mean value can be better matched for operation, and the operation accuracy is improved.

[0167] The second scaling factor can be pre-configured, or the second scaling factor can be decoded from the code stream corresponding to the current block, such as decoding the second scaling factor from the first code stream corresponding to the current block, or decoding the second scaling factor from the second code stream corresponding to the current block, or decoding the second scaling factor from other code streams, which is not limited.

[0168] Embodiment 7: see Figure 8 As shown, a flowchart of a decoding method is shown. The decoding end can receive Bitstream#1 (that is, the first code stream) and Bitstream#2 (that is, the second code stream). The decoding end can parse the initial coefficient hyperparameter feature z_hat from Bitstream#1. Based on the initial coefficient hyperparameter feature z_hat, the standard deviation of all feature points can be extracted, that is, the initial standard deviation σ of all feature points is obtained. After the initial standard deviation is corrected by a probability model, the target standard deviation σ_hat after correction is obtained.

[0169] The decoding end can generate a target probability distribution model PDM_hat based on the target standard deviation, and decode Bitstream#2 based on the target probability distribution model PDM_hat to obtain the residual feature corresponding to the current block. For example, in the entire decoding process, the parameters of the target probability distribution model PDM_hat of each residual are set as σ_hat 2 The variance is a distribution used for entropy decoding of Bitsstream#2 to obtain the residual feature r_hat. Therefore, the decoding end can decode Bitstream#2 based on the target probability distribution model PDM_hat to obtain the residual feature r_hat corresponding to the current block.

[0170] The decoding end can parse the initial coefficient hyperparameter feature z_hat from Bitstream#1. Based on the initial coefficient hyperparameter feature z_hat, the mean value of the current point can be extracted. The residual feature r_hat plus the mean value can obtain the reconstruction value y_hat.

[0171] Embodiment 8: For the encoding end, the encoding method can include: determining a zero-mean Gaussian distribution The zero-mean Gaussian distribution can be a target probability distribution model, and the residual feature r_hat is obtained by using arithmetic coding on the second bit stream (Bitstream #2). The encoding end learns a distribution representing the probability of each feature point according to the image content. The selection of the probability distribution model is confirmed as a hyperparameter before training, which can be a Gaussian distribution, a mixed Gaussian, etc., and the type of the probability distribution model is not limited. For example, assuming that each feature point obeys a normal distribution with a mean of 0 and a variance of σ 2 To estimate the distribution of each feature point, 1 parameter σ, i.e., the initial standard deviation, needs to be learned for each parameter. The related information of the initial standard deviation is written in Bitstream #1 in the hyper-prior network. In this embodiment, the initial standard deviation can be synchronously corrected according to the first scaling factor, and therefore, when the target probability distribution model is used to predict the entropy coding, the target standard deviation σ_hat needs to be obtained through a probability model correction module. The target probability distribution model can be represented as That is, the target probability distribution model is related to the mean 0 and the target standard deviation σ_hat.

[0172] For example, since the entropy model is not fixed, the encoding end and the decoding end need to know the probability distribution model parameter σ of each feature point, and the information related to σ needs to be transmitted to the decoding end. Precise modeling can reduce the bit rate of the latent domain feature.

[0173] Embodiment 9: In the encoding process, the residual feature r will undergo various scaling processes, such as the G-unit scaling process, which will cause the distribution of the residual feature r to change. However, in the entire coding framework, the input parameter of the initial standard deviation σ generation network is the original feature y of the latent domain, and in this process, the initial standard deviation σ cannot perceive the change of the residual, resulting in a deviation between the distribution described by the initial standard deviation σ and the real distribution of the residual. In order to enable the initial standard deviation σ to synchronously perceive the transformation of the residual, a probability model correction module is added after the initial standard deviation σ, which aims to map the change of the residual to the initial standard deviation σ. As shown in Figure 9A The input is the initial standard deviation σ and the first scaling factor scale (i.e., the first scaling factor of the G-unit process), and the output is the target standard deviation σ_hat. Obviously, the target standard deviation σ_hat can be determined based on the initial standard deviation σ and the first scaling factor scale, such as the product of the initial standard deviation σ and the first scaling factor scale.

[0174] In Embodiment 9, the mapping relationship adopts a 1:1 scaling relationship, that is, a first scaling factor of the G-Unit is used to generate a target standard deviation σ_hat after operating on the initial standard deviation σ once in the same way, as shown in the following formula: σ_hat = Scale * σ, so that the target probability distribution pattern can be generated The distribution of the prediction residual. The first scaling factor Scale can be obtained by training, or can be obtained by other means without training, which is not limited.

[0175] Embodiment 10: Based on the scaling strategy of the initial standard deviation σ in Embodiment 9, in Embodiment 10, not only the numerical scaling of the residual feature res at the encoding end is considered, but also the difference in the original physical meaning of the residual feature and the initial standard deviation σ is considered. For example, as shown in Figure 9B The input of the probability model correction module includes not only the initial standard deviation σ and the first scaling factor scale (that is, the first scaling factor of the G-unit process), but also the second scaling factor Scale(σ, res) for describing the original physical meaning difference between the residual and the standard deviation. The second scaling factor Scale(σ, res) is used to represent the mapping relationship between the residual and the standard deviation, and the second scaling factor can be pre-configured; or the second scaling factor can be decoded from the code stream corresponding to the current block. The output is the target standard deviation σ_hat after scaling correction. Obviously, the target standard deviation σ_hat can be determined based on the initial standard deviation σ, the first scaling factor scale and the second scaling factor Scale(σ, res), for example, the product of the initial standard deviation σ, the first scaling factor scale and the second scaling factor Scale(σ, res) can be the target standard deviation σ_hat.

[0176] Since there is a difference between the residual and the parameter σ of the normal distribution itself, 1:1 is not the best mapping relationship. Therefore, in order to more accurately perceive the change of the residual to the initial standard deviation σ, as shown in the following formula: σ_hat = Scale(σ, res) * Scale * σ, the initial standard deviation σ is corrected by the second scaling factor Scale(σ, res) in addition to the first scaling factor scale.

[0177] Embodiment 11: In Embodiments 9 and 10, the scaling factor correction of the initial standard deviation σ is directly performed, and in Embodiment 11, the input z_hat of the hyper-scale decoding network used to generate the initial standard deviation σ is corrected, that is, the initial coefficient hyperparameter feature z_hat is corrected to obtain the target coefficient hyperparameter feature In the following formula, the target coefficient hyperparameter feature After inputting the hyper-scale decoding network, the target standard deviation σ_hat can be obtained. For example, in order to make the target standard deviation σ_hat can be synchronized to perceive the transformation of the residual, a probability model correction module is added after the initial coefficient hyperparameter feature z_hat, which aims to map the change of the residual to the initial coefficient hyperparameter feature z_hat. The input is the initial coefficient hyperparameter feature z_hat and the first scaling factor scale (i.e. the first scaling factor of the G-unit process), and the output is the target coefficient hyperparameter feature After obtaining the target coefficient hyperparameter feature After inputting the hyper-scale decoding network, the target standard deviation σ_hat can be obtained. Obviously, the target coefficient hyperparameter feature As the product of the initial coefficient hyperparameter feature z_hat and the first scaling factor scale is the target coefficient hyperparameter feature Based on the target coefficient hyperparameter feature The target standard deviation σ_hat is determined.

[0178] In embodiment 11, the mapping relationship adopts a 1:1 scaling relationship, that is, the first scaling factor of the G-unit is operated on the initial coefficient hyperparameter feature z_hat once in the same way to generate the target coefficient hyperparameter feature The process can be seen from the following formula, In this way, the target probability distribution mode The distribution of the residual is predicted.

[0179] Embodiment 12: On the basis of the scaling strategy of the initial coefficient hyperparameter feature z_hat in embodiment 11, in embodiment 12, not only the numerical scaling of the residual feature res at the encoding end is considered, but also the difference in the original physical meaning between the residual feature and the initial coefficient hyperparameter feature z_hat is considered. For example, the input of the probability model correction module not only includes the initial coefficient hyperparameter feature z_hat and the first scaling factor scale (i.e. the first scaling factor of the G-unit process), but also increases the third scaling factor Scale(z_hat, res) for describing the original physical meaning difference between the residual and the coefficient hyperparameter feature. The third scaling factor Scale(z_hat, res) is used to represent the mapping relationship between the residual and the coefficient hyperparameter feature, and the third scaling factor can be pre-configured; or the third scaling factor can be decoded from the code stream corresponding to the current block, and the output is the target coefficient hyperparameter feature Obviously, the target coefficient hyperparameter feature The product of z_hat, scale, and Scale(z_hat, res) can be a target coefficient hyperparameter feature

[0180] Since the residual and the coefficient hyperparameter feature are different, 1:1 is not the best mapping relationship. Therefore, in order to more accurately perceive the change of the residual in the coefficient hyperparameter feature, see the following formula, The initial coefficient hyperparameter feature z_hat is modified by a third scaling factor Scale(z_hat, res) in addition to the first scaling factor scale.

[0181] For example, each of the above embodiments can be implemented alone, or can be implemented in combination, for example, each of embodiments 1-12 can be implemented alone, and at least two of embodiments 1-12 can be implemented in combination.

[0182] For example, in each of the above embodiments, the content of the encoding end can also be applied to the decoding end, that is, the decoding end can be processed in the same way, and the content of the decoding end can also be applied to the encoding end, that is, the encoding end can be processed in the same way.

[0183] Based on the same application concept as the above method, the present embodiment also proposes a decoding device, which is applied to a decoding end, and the device comprises: a memory configured to store video data; and a decoder configured to implement the decoding method in embodiments 1-12 above, that is, the processing flow of the decoding end.

[0184] For example, in a possible implementation, the decoder is configured to implement:

[0185] Obtain a target probability distribution model corresponding to the first scaling factor based on a first code stream corresponding to the current block;

[0186] Decode a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a modified residual feature corresponding to the current block;

[0187] Determine a reconstructed image block corresponding to the current block based on the modified residual feature.

[0188] Based on the same application concept as the above method, the present embodiment also proposes an encoding device, which is applied to an encoding end, and the device comprises: a memory configured to store video data; and an encoder configured to implement the encoding method in embodiments 1-12 above, that is, the processing flow of the encoding end.

[0189] For example, in a possible implementation, the encoder is configured to implement:

[0190] correcting the residual feature corresponding to the current block to obtain a corrected residual feature;

[0191] obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to the current block;

[0192] encoding the corrected residual feature based on the target probability distribution model to obtain a second code stream corresponding to the current block, and sending the second code stream to a decoding end.

[0193] Based on the same application concept as the above method, the decoding end device (also can be called as a video decoder) provided by the embodiments of the present application can be seen from the hardware architecture diagram of the hardware layer as shown in Figure 10A The processor 1001 and the machine readable storage medium 1002, the machine readable storage medium 1002 stores machine executable instructions that can be executed by the processor 1001; the processor 1001 is used to execute the machine executable instructions to realize the decoding method of the above-mentioned embodiments 1-12 of the present application. For example, in a possible implementation manner, the decoding end device is used to realize:

[0194] obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to the current block;

[0195] decoding the second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a corrected residual feature corresponding to the current block;

[0196] determining a reconstructed image block corresponding to the current block based on the corrected residual feature.

[0197] Based on the same application concept as the above method, the decoding end device (also can be called as a video decoder) provided by the embodiments of the present application can be seen from the hardware architecture diagram of the hardware layer as shown in Figure 10B The processor 1011 and the machine readable storage medium 1012, the machine readable storage medium 1012 stores machine executable instructions that can be executed by the processor 1011; the processor 1011 is used to execute the machine executable instructions to realize the encoding method of the above-mentioned embodiments 1-12 of the present application. For example, in a possible implementation manner, the decoding end device is used to realize:

[0198] obtaining a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to the current block;

[0199] decoding the second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain a corrected residual feature corresponding to the current block;

[0200] determine a reconstructed image block corresponding to the current block based on the modified residual feature.

[0201] Based on the same application concept as the above method, the embodiments of the present application provide an electronic device. It includes a processor and a machine readable storage medium, the machine readable storage medium stores machine executable instructions which can be executed by the processor; the processor is used to execute the machine executable instructions to realize the decoding method or the encoding method of the above embodiments 1-12 of the present application.

[0202] Based on the same application concept as the above method, the embodiments of the present application also provide a machine readable storage medium, the machine readable storage medium stores a plurality of computer instructions, the computer instructions are executed by the processor, and the decoding method or the encoding method disclosed in the above examples of the present application can be realized.

[0203] Based on the same application concept as the above method, the embodiments of the present application also provide a computer application program, which can realize the decoding method or the encoding method disclosed in the above examples of the present application when executed by the processor.

[0204] Based on the same application concept as the above method, the embodiments of the present application also provide a decoding device, which is applied to a decoding end, and includes: an acquisition module, which is used to acquire a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to a current block; a decoding module, which is used to decode a second code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor, to obtain a modified residual feature corresponding to the current block; and a determination module, which is used to determine a reconstructed image block corresponding to the current block based on the modified residual feature.

[0205] For example, when the acquisition module acquires the target probability distribution model corresponding to the first scaling factor based on the first code stream corresponding to the current block, it is specifically used for: acquiring target probability distribution parameters corresponding to the first scaling factor based on the first code stream corresponding to the current block; and determining the target probability distribution model based on the target probability distribution parameters.

[0206] For example, the target probability distribution parameters include target mean and / or target standard deviation.

[0207] For example, when the target probability distribution parameter is a target standard deviation, the obtaining module is specifically configured to: decode the first code stream corresponding to the current block to obtain initial coefficient hyperparameter features corresponding to the current block; perform coefficient hyperparameter feature inverse transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation; and correct the initial standard deviation based on the first scaling factor to obtain the target standard deviation.

[0208] For example, when the obtaining module corrects the initial standard deviation based on the first scaling factor to obtain the target standard deviation, the obtaining module is specifically configured to: generate a target scaling factor based on the first scaling factor; and correct the initial standard deviation based on the target scaling factor to obtain the target standard deviation.

[0209] For example, the target scaling factor is the first scaling factor.

[0210] Alternatively, the target scaling factor is determined based on the first scaling factor and a second scaling factor.

[0211] The second scaling factor is used to represent a mapping relationship between a residual and a standard deviation.

[0212] For example, the second scaling factor is preconfigured; or

[0213] The second scaling factor is decoded from a code stream corresponding to the current block.

[0214] For example, when the target probability distribution parameter is a target standard deviation, the obtaining module is specifically configured to: decode the first code stream corresponding to the current block to obtain initial coefficient hyperparameter features corresponding to the current block; correct the initial coefficient hyperparameter features based on the first scaling factor to obtain target coefficient hyperparameter features; and perform coefficient hyperparameter feature inverse transformation on the target coefficient hyperparameter features to obtain the target standard deviation.

[0215] For example, when the obtaining module corrects the initial coefficient hyperparameter features based on the first scaling factor to obtain target coefficient hyperparameter features, the obtaining module is specifically configured to: generate a target scaling factor based on the first scaling factor; and correct the initial coefficient hyperparameter features based on the target scaling factor to obtain the target coefficient hyperparameter features.

[0216] For example, the target scaling factor is the first scaling factor; or

[0217] The target scaling factor is determined based on the first scaling factor and a third scaling factor.

[0218] The third scaling factor is used to represent a mapping relationship between the residual and the coefficient hyperparameter feature.

[0219] The third scaling factor is pre-configured.

[0220] The third scaling factor is decoded from the code stream corresponding to the current block.

[0221] Based on the same application concept as the above method, an encoding device is also proposed in the embodiments of the present application. The device is applied to an encoding end. The device comprises a correction module, which is used to correct a residual feature corresponding to a current block to obtain a corrected residual feature; an acquisition module, which is used to acquire a target probability distribution model corresponding to a first scaling factor based on a first code stream corresponding to the current block; and an encoding module, which is used to encode the corrected residual feature based on the target probability distribution model to obtain a second code stream corresponding to the current block, and send the second code stream to a decoding end.

[0222] The correction module is specifically used to correct the residual feature corresponding to the current block based on the first scaling factor to obtain the corrected residual feature.

[0223] The acquisition module is specifically used to acquire target probability distribution parameters corresponding to the first scaling factor based on the first code stream corresponding to the current block, and determine the target probability distribution model based on the target probability distribution parameters.

[0224] The encoding module is further used to perform feature transformation on the current block to obtain an image feature corresponding to the current block, perform coefficient hyperparameter feature transformation on the image feature to obtain an initial coefficient hyperparameter feature corresponding to the current block, and encode the initial coefficient hyperparameter feature to obtain the first code stream.

[0225] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. The present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Embodiments of the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage etc.) embodying computer readable program code. The foregoing is merely illustrative of the principles of the application. Numerous modifications and adaptations will be apparent to those skilled in the art without departing from the spirit and scope of the application. Any modifications, equivalents, improvements, etc., which are within the spirit and scope of the application are intended to be encompassed by the claims.

Claims

1. An image decoding method characterized by, The method comprises: decoding a code stream corresponding to a current block to obtain initial coefficient hyperparameter features corresponding to the current block, performing inverse coefficient hyperparameter feature transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation, correcting the initial standard deviation based on a first scaling factor to obtain a target standard deviation, and determining a target probability distribution parameter corresponding to the first scaling factor based on the target standard deviation; determining a target probability distribution model based on the target probability distribution parameter; decoding the code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain corrected residual features corresponding to the current block; determining a reconstructed image block corresponding to the current block based on the corrected residual features.

2. The method of claim 1, wherein, The target probability distribution parameter further comprises a target mean value.

3. The method of claim 2, wherein, The target mean value is 0.

4. The method of claim 1, wherein the correcting the initial standard deviation based on the first scaling factor to obtain the target standard deviation comprises: determining a product value of the initial standard deviation and the first scaling factor as the target standard deviation.

5. An image coding method characterized by, The method comprises: correcting residual features corresponding to a current block to obtain corrected residual features; obtaining initial coefficient hyperparameter features, performing inverse coefficient hyperparameter feature transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation, correcting the initial standard deviation based on a first scaling factor to obtain a target standard deviation, and determining a target probability distribution parameter corresponding to the first scaling factor based on the target standard deviation; determining a target probability distribution model based on the target probability distribution parameter; encoding the corrected residual features based on the target probability distribution model to obtain a code stream corresponding to the current block, and sending the code stream to a decoding end.

6. An image decoding apparatus characterized by comprising: The apparatus comprises: an obtaining module configured to decode a code stream corresponding to a current block to obtain initial coefficient hyperparameter features corresponding to the current block, perform inverse coefficient hyperparameter feature transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation, correct the initial standard deviation based on a first scaling factor to obtain a target standard deviation, and determine a target probability distribution parameter corresponding to the first scaling factor based on the target standard deviation; and determine a target probability distribution model based on the target probability distribution parameter; a decoding module configured to decode the code stream corresponding to the current block based on the target probability distribution model corresponding to the first scaling factor to obtain corrected residual features corresponding to the current block; a determining module configured to determine a reconstructed image block corresponding to the current block based on the corrected residual features.

7. An image coding apparatus characterized by comprising: The apparatus comprises: a correcting module configured to correct residual features corresponding to a current block to obtain corrected residual features; an obtaining module configured to obtain initial coefficient hyperparameter features, perform inverse coefficient hyperparameter feature transformation on the initial coefficient hyperparameter features to obtain an initial standard deviation, correct the initial standard deviation based on a first scaling factor to obtain a target standard deviation, and determine a target probability distribution parameter corresponding to the first scaling factor based on the target standard deviation; and determine a target probability distribution model based on the target probability distribution parameter; The encoding module is configured to encode the modified residual feature based on the target probability distribution model to obtain a code stream corresponding to the current block, and send the code stream to a decoding end.

8. A decoding-side device, comprising: The method comprises: a processor and a machine readable storage medium, the machine readable storage medium storing machine executable instructions executable by the processor; the processor is configured to execute the machine executable instructions to implement the method of any one of claims 1-4.

9. An encoding end device, comprising: The method comprises: a processor and a machine readable storage medium, the machine readable storage medium storing machine executable instructions executable by the processor; the processor is configured to execute the machine executable instructions to implement the method of claim 5.

10. A machine-readable storage medium, characterized in that, The machine readable storage medium stores a plurality of computer instructions, and the computer instructions are executed by the processor to implement the method of any one of claims 1-4, or the computer instructions are executed by the processor to implement the method of claim 5.

11. A computer program product, characterised in that, The computer program product comprises a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1-5.

Images