JP6968228B2 - Methods and equipment for image coding and decoding via inter-prediction - Google Patents

Description

The present invention relates to the field of video imaging technology and, in particular, to methods and devices for image coding and decoding via inter-prediction.

Video coding and decoding frameworks typically use a hybrid coding structure for coding and decoding video sequences. The coder side of the hybrid coded structure typically includes a predictor module, a transform module, a quantization module and an entropy coded module. The decoder side of the hybrid coded structure typically includes an entropy decoding module, an inverse quantization module, an inverse transformation module and a predictive compensation module. These coding and decoding modules can be combined to effectively remove redundant information in the video sequence and ensure that the coded image of the video sequence is obtained on the decoder side.

In a video coding and decoding framework, images in a video sequence are usually divided into image blocks for coding. One image is divided into several image blocks, which are encoded and decoded by using the modules described above.

In the above-mentioned module, the prediction module is configured to acquire the prediction block information of the image block of the coded image of the video sequence on the encoder side, whereby the residual of the image block is based on a specific mode. Determine if it needs to be retrieved. The predictive compensation module acquires the predicted block information of the current decoded image block on the decoder side, and then, based on the specific mode, based on the image block residuals obtained via decoding, the current decoded image. It is configured to decide whether to get the block. Prediction modules and predictive compensation modules typically include two technologies: intra-prediction and inter-prediction. In inter-prediction techniques, an image adjacent to the current image and used for inter-prediction is called a reference image.

In the inter-prediction technology, in order to effectively remove the redundant information of the current image block, the encoder side and the decoder side are to reduce the residual of the current image block from the reference image to the current image block. You need to get the matching image block that most closely resembles. When inter-prediction is performed on the current image block, a reference image is searched for the appropriate matching image block based on the predicted motion information. In this case, the motion vector information in the predicted motion information is determined by the motion vector information of the adjacent block. In other words, all the information in the previous reference image may be used to encode and decode the current image block. It can be seen that in interframe coding and decoding, the motion vector information of the adjacent block is used directly as the motion vector predictor of the current block. This method of directly acquiring the motion vector predictor of the inter-prediction inevitably causes a deviation of the predicted motion accuracy.

Embodiments of the present invention provide methods and devices for image coding and decoding via inter-prediction to improve the motion accuracy of inter-prediction.

According to the first aspect, a method for image coding via inter-prediction is provided, wherein the method is a step of determining the predicted motion information of the first coded block, wherein the predicted motion information is. Predictive reference image information, including motion vector predictors and predictive reference image information, is bidirectional to the first coded block based on the steps and predictive motion information used to represent the predictive reference image block. Steps to perform a prediction to get the initial coded predictive block of the first coded block, and based on the initial coded predictive block, get the first coded predictive block of the first coded block, and the prediction reference. A step of performing a first precision motion search on an image block to obtain at least one second coded predictor block, where the motion search search positions are the motion vector predictor and the second. Calculate the difference between the step and the first coded predictive block and each second coded predictive block, as determined by using the precision of 1, the first coded block and the first coded block with the smallest difference. Both the step of using the motion vector predictor between the two coded predictors as the target motion vector predictor of the first coded block and the first coded block based on the target motion vector predictor. The step of executing the direction prediction to obtain the third coded prediction block of the first coded block and the target coded prediction block of the first coded block based on the third coded prediction block are obtained. Includes a step of encoding the first coded block based on the target coded predictive block.

A beneficial effect is that when encoding the first coded block, the encoder side can update the target motion vector predictor via motion search to improve the motion accuracy of the inter-prediction. ..

Referring to the first aspect, in a possible design, the predictive motion information includes forward predictive motion information and backward predictive motion information, and the motion vector predictor includes forward motion vector predictor and backward motion vector predictor. The predictive reference image block includes a forward predictive reference image block and a backward predictive reference image block.

Referring to the first aspect, in a possible design, the initial coded prediction block includes a forward initial coded predictive block and a backward initial coded predictive block, and is based on the predicted motion information with respect to the first coded block. The step of executing bidirectional prediction to obtain the initial coded prediction block of the first coded block is to perform forward prediction for the first coded block based on the forward predicted motion information, and the first step is to perform forward prediction. Based on the step of acquiring the forward initial coding prediction block of one coded block and the backward prediction motion information, the backward prediction is executed for the first coded block, and the backward initial coding of the first coded block is performed. Includes a step to get the prediction block.

Referring to the first aspect, in a possible design, based on the initial coded predictive block, the step of obtaining the first coded predictive block of the first coded block is the forward initial coded predictive block and the backward initial code. The step of performing weighting addition to the coded prediction block to obtain the first coded predictive block of the first coded block, or the forward initial coded predictive block is the first coded block of the first coded block. It includes a method of a step used as a prediction block or a step method in which a backward initial coded prediction block is used as a first coded prediction block of the first coded block.

The beneficial effect is that after performing a bidirectional prediction on the first coded block based on the predicted motion information, the encoder side combines the forward initial coded predictive block and the backward initial coded predictive block. , The first coded prediction block can be obtained, or the first coded prediction block can be obtained based on the unidirectional initial coded prediction block. Implementations are diversified and may be flexibly selected based on actual circumstances.

Referring to the first aspect, in a possible design, the second coded prediction block includes a forward second coded predictive block and a backward second coded predictive block, with a first accuracy relative to the predictive reference image block. The step of performing a motion search in to obtain at least one second coded prediction block performs a first-precision motion search for the forward prediction reference image block based on the forward motion vector predictor. , Using each found forward-coded prediction block as a forward second-coded prediction block to obtain at least one second-coded prediction block, and a backward prediction reference based on the backward motion vector predictor. Perform a first-precision motion search on the image block and use each found backward-coded predictive block as the backward second-coded predictive block to get at least one second-coded predictive block. Including steps, where the first accuracy includes integer pixel accuracy, 1/2 pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Referring to the first aspect, in a possible design, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, a first coded predictor block and each second coded predictor block. Calculates the difference between and uses the motion vector predictor between the first and second coded blocks with the smallest difference as the target motion vector predictor for the first coded block. The step compares the difference between the forward second coded prediction block and the first coded predictive block, and the target forward between the first coded block and the forward second coded predictive block with the smallest difference. The step of using the motion vector predictor as the target motion vector predictor and / or the first with the smallest difference comparing the differences between the backward second coded predictor block and the first coded predictor block. It comprises the step of using the target backward motion vector predictor between the coded block and the backward second coded predictor as the target motion vector predictor.

The beneficial effect is that the difference between the second coded predictive block obtained by performing a motion search on the predictive reference image block and the first coded predictive block is the target of the first coded block. The motion vector predictors are compared to update, thereby improving the motion accuracy of the inter-prediction.

Referring to the first aspect, in a possible design, the third coded predictor block of the first coded block includes a forward third coded predictive block and a backward third coded predictor block, and is a target motion vector predictor. Based on, the step of performing bidirectional prediction on the first coded block to obtain the third coded prediction block of the first coded block is the first step based on the target forward motion vector predictor. For the first coded block, based on the step of performing forward prediction on the coded block to get the forward third coded predictive block of the first coded block and the target backward motion vector predictor. Includes a step of performing a backward prediction to obtain a backward third coded predictive block of the first coded block.

Referring to the first aspect, in a possible design, based on the third coded predictive block, the step of acquiring the target coded predictive block of the first coded block is the forward third coded predictive block and the backward third coded predictive block. The step of performing weighting addition to the 3 coded prediction block to obtain the target coded predictive block of the 1st coded block, or the forward 3rd coded predictive block is the target code of the 1st coded block. Includes a step to use as a conversion prediction block or a step of using a backward third coding prediction block as the target coding prediction block of the first coding block.

The beneficial effect is that after performing a bidirectional prediction on the 3rd coded block based on the target motion vector predictor and getting the 3rd coded predictor block, the encoder side has a forward 3rd code. The target coding prediction block of the first coding block can be obtained by combining the conversion prediction block and the rear third coding prediction block, or based on the unidirectional third coding prediction block. It is possible to acquire the target coding prediction block of the first coding block. Implementations are diversified and may be flexibly selected based on actual circumstances.

According to the second aspect, a method for image decoding via inter-prediction is provided, wherein the method is a step of determining the predicted motion information of the first decoding block, wherein the predicted motion information is a motion vector. Predictive reference image information, including predictors and predictive reference image information, performs bidirectional predictions on the first decoded block based on the steps and predictive motion information used to represent the predictive reference image block. Then, the step of acquiring the initial decoding prediction block of the first decoding block and acquiring the first decoding prediction block of the first decoding block based on the initial decoding prediction block, and the first step with respect to the prediction reference image block. A step of performing a motion search with precision to obtain at least one second decoding prediction block, where the search position for the motion search is determined by using the motion vector predictor and the first precision. The step and the motion vector predictor between the first and second decoding prediction blocks with the smallest difference is calculated by calculating the difference between the first decoding prediction block and each second decoding prediction block. To perform bidirectional prediction on the first decoding block based on the step of using as the target motion vector predictor of the first decoding block and the third decoding of the first decoding block. It includes a step of acquiring a prediction block and a step of acquiring a target decoding prediction block of the first decoding block based on the third decoding prediction block and decoding the first decoding block based on the target decoding prediction block.

A beneficial effect is that when encoding the first decoding block, the decoder side can update the target motion vector predictor via motion search in order to improve the motion accuracy of the inter-prediction.

Referring to the second aspect, in a possible design, the predictive motion information includes forward predictive motion information and backward predictive motion information, and the motion vector predictor includes forward motion vector predictor and backward motion vector predictor. The predictive reference image block includes a forward predictive reference image block and a backward predictive reference image block.

Referring to the second aspect, in a possible design, the initial decoding prediction block includes a forward initial decoding prediction block and a backward initial decoding prediction block, and is bidirectionally predicted with respect to the first decoding block based on the predicted motion information. To obtain the initial decoding prediction block of the first decoding block by executing It includes a step of acquiring a prediction block and a step of performing backward prediction on the first decoding block based on the backward prediction motion information to acquire the backward initial decoding prediction block of the first decoding block.

Referring to the second aspect, in a possible design, the step of acquiring the first decoding prediction block of the first decoding block based on the initial decoding prediction block is for the forward initial decoding prediction block and the rear initial decoding prediction block. To obtain the first decoding prediction block of the first decoding block by performing weighting addition, or to use the forward initial decoding prediction block as the first decoding prediction block of the first decoding block, or the rear. The method including using the initial decoding prediction block as the first decoding prediction block of the first decoding block is included.

The beneficial effect is that after performing a bidirectional prediction on the first decoding block based on the predicted motion information, the decoder side combines the forward initial decoding prediction block and the backward initial decoding prediction block, and the first It is possible to obtain a decryption prediction block, or to obtain a first decoding prediction block based on a unidirectional initial decoding prediction block. Implementations are diversified and may be flexibly selected based on actual circumstances.

Referring to the second aspect, in a possible design, the second decoding prediction block includes a front second decoding prediction block and a rear second decoding prediction block, and a motion search with first accuracy with respect to the prediction reference image block. Was found by performing a first-precision motion search on the forward-predicted reference image block based on the forward motion vector predictor. The first for the backward prediction reference image block, based on the step of using each forward decoding prediction block as the forward second decoding prediction block to obtain at least one second decoding prediction block and the backward motion vector predictor. A step of performing a motion search with the accuracy of and using each found backward decoding prediction block as a backward second decoding prediction block to obtain at least one second decoding prediction block, wherein the first The accuracy of includes integer pixel accuracy, 1/2 pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Referring to the second aspect, in a possible design, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, with a first decoding prediction block and each second decoding prediction block. The step of calculating the difference between and using the motion vector predictor between the first and second decoding block with the smallest difference as the target motion vector predictor of the first decoding block is forward. 2 Compares the difference between the decoding prediction block and the first decoding prediction block, and sets the target forward motion vector predictor between the first decoding block and the forward second decoding prediction block having the smallest difference as the target motion vector. The step used as a predictor and / or the difference between the backward second decoding prediction block and the first decoding prediction block is compared, and the first decoding block and the backward second decoding prediction block having the smallest difference are compared. Includes steps to use the target backward motion vector predictor between them as the target motion vector predictor.

The beneficial effect is that the difference between the second decoding prediction block obtained by performing a motion search on the prediction reference image block and the first decoding prediction block is the target motion vector prediction of the first decoding block. It is to be compared to update the children, thereby improving the motion accuracy of the inter-prediction.

Referring to the second aspect, in a possible design, the third decoding prediction block of the first decoding block includes a forward third decoding prediction block and a rear third decoding prediction block, based on the target motion vector predictor. The step of performing bidirectional prediction on the first decoding block to get the third decoding prediction block on the first decoding block is a forward prediction on the first decoding block based on the target forward motion vector predictor. To get the third forward decoding predictive block of the first decoding block, and to perform backward prediction on the first decoding block based on the target backward motion vector predictor, the first decoding block. Includes a step to obtain the backward third decryption prediction block of.

Referring to the second aspect, in a possible design, based on the third decoding prediction block, the step of acquiring the target decoding prediction block of the first decoding block is the forward third decoding prediction block and the rear third decoding prediction block. A method of acquiring the target decoding prediction block of the first decoding block by performing weighting addition on the target, or a method of using the forward third decoding prediction block as the target decoding prediction block of the first decoding block, or A method is included in which the rear third decoding prediction block is used as the target decoding prediction block of the first decoding block.

The beneficial effect is that after performing a bidirectional prediction on the third decoding block based on the target motion vector predictor to get the third decoding prediction block, the decoder side will see the forward third decoding prediction block. And the backward third decoding prediction block can be combined to obtain the target decoding prediction block of the first decoding block, or based on the unidirectional third decoding prediction block, the target decoding of the first decoding block. It is possible to get a prediction block. Implementations are diversified and may be flexibly selected based on actual circumstances.

According to a third aspect, a device for image coding via inter-prediction is provided, wherein the device is a determination unit configured to determine the predicted motion information of the first coded block. , Predictive motion information includes motion vector predictors and predictive reference image information, and predictive reference image information is a first code based on a determination unit and predictive motion information used to represent a predictive reference image block. Perform bidirectional prediction on the coded block to get the initial coded predictive block of the first coded block, and based on the initial coded predictive block, the first coded predictive block of the first coded block. Acquires and performs a first precision motion search on the predictive reference image block to acquire at least one second coded predictor block, where the motion search search positions are the motion vector predictor and Determined by using the first precision, the difference between the first coded predictive block and each second coded predictive block is calculated and the first and second coded blocks with the smallest difference are calculated. A motion vector predictor to and from the prediction block is used as the target motion vector predictor for the first coded block to perform bidirectional predictions for the first coded block based on the target motion vector predictor. Then, based on the processing unit configured to acquire the third coded prediction block of the first coded block and the third coded predicted block, the target coded predicted block of the first coded block is acquired. , Includes a coding unit configured to encode the first coding block based on the target coding prediction block.

Referring to the third aspect, in a possible design, the predicted motion information includes forward predicted motion information and backward predicted motion information, and the motion vector predictor includes forward motion vector predictor and backward motion vector predictor. The predictive reference image block includes a forward predictive reference image block and a backward predictive reference image block.

Referring to the third aspect, in a possible design, the initial coded prediction block includes a forward initial coded predictive block and a backward initial coded predictive block, and is based on the predicted motion information with respect to the first coded block. When performing bidirectional prediction to obtain the initial coded prediction block of the first coded block, the processing unit specifically refers to the first coded block based on the forward predicted motion information. Perform forward prediction to get the forward initial coding prediction block of the first coded block, perform backward prediction on the first coded block based on the backward prediction motion information, and perform the first coding. It is configured to get the backward initial coded prediction block of the block.

Referring to the third aspect, in a possible design, when acquiring the first coded predictive block of the first coded block based on the initial coded predictive block, the processing unit is specifically forward-initial. Perform weighting addition on the coded prediction block and the backward initial coded prediction block to get the first coded predictive block of the first coded block, or the forward initial coded predictive block with the first code. It is configured to be used as the first coded predictive block of the coded block, or to use the backward initial coded predictive block as the first coded predictive block of the first coded block.

Referring to the third aspect, in a possible design, the second coded prediction block includes a forward second coded predictive block and a backward second coded predictive block, based on the motion vector predictor, the predictive reference image. When performing a first-precision motion search on a block to obtain at least one second coded prediction block, the processing unit specifically predicts forward based on the forward motion vector predictor. Perform a first-precision motion search on the reference image block and use each found forward-coded predictive block as the forward second-coded predictive block to obtain at least one second-coded predictive block. Then, based on the backward motion vector predictor, the motion search with the first accuracy is performed for the backward predictive reference image block, and each found backward coded predictive block is used as the backward second coded predictive block. And is configured to acquire at least one second coded prediction block, where the first precision is integer pixel precision, 1/2 pixel precision, 1/4 pixel precision or 1/8 pixel precision. include.

Referring to the third aspect, in a possible design, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, a first coded predictor block and each second coded predictor block. Calculates the difference between and uses the motion vector predictor between the first and second coded blocks with the smallest difference as the target motion vector predictor for the first coded block. When, the processing unit specifically compares the difference between the forward second coded prediction block and the first coded predictive block, and has the smallest difference between the first coded block and the forward second coded block. Use the target forward motion vector predictor with the predictor block as the target motion vector predictor, or compare the difference between the backward second coded predictor block and the first coded predictor block to minimize the difference. The target backward motion vector predictor between the first coded block and the posterior second coded predictor with the difference is configured to be used as the target motion vector predictor.

Referring to the third aspect, in a possible design, the third coded predictor block of the first coded block includes a forward third coded predictive block and a backward third coded predictor block, and is a target motion vector predictor. Based on, when performing a bidirectional prediction on the first coded block to get the third coded predictive block of the first coded block, the processing unit specifically has a target forward motion vector. Based on the predictor, perform forward prediction on the first coded block to get the forward third coded predictive block of the first coded block, and based on the target backward motion vector predictor, the first It is configured to perform backward prediction on the coded block and acquire the backward third coded predictive block of the first coded block.

Referring to the third aspect, in a possible design, when acquiring the target coded predictive block of the first coded block based on the third coded predictive block, the processing unit is specifically forward-first. The 3rd coded prediction block and the 3rd backward coded predictive block are weighted and added to obtain the target coded predictive block of the 1st coded block, or the 3rd forward coded predictive block is obtained. It is configured to be used as the target coding prediction block of one coding block, or to use the rear third coding prediction block as the target coding prediction block of the first coding block.

According to a fourth aspect, a device for image decoding via inter-prediction is provided, wherein the device is a determination unit configured to determine the predicted motion information of the first decoding block, wherein the prediction is made. The motion information includes a motion vector predictor and a predictive reference image information, and the predictive reference image information is used to represent the predictive reference image block in the first decoding block based on the decision unit and the predictive motion information. On the other hand, bidirectional prediction is executed to obtain the initial decoding prediction block of the first decoding block, and based on the initial decoding prediction block, the first decoding prediction block of the first decoding block is acquired and used as the prediction reference image block. On the other hand, a motion search with the first accuracy is performed to obtain at least one second decoding prediction block, where the search position of the motion search is determined by using the motion vector predictor and the first accuracy. The motion vector predictor between the first and second decryption predictor blocks that is determined and has the smallest difference is calculated by calculating the difference between the first decode predictor block and each second decode predictor block. 1 Used as the target motion vector predictor of the decoding block, and based on the target motion vector predictor, performs bidirectional prediction on the first decoding block to obtain the third decoding prediction block of the first decoding block. Based on the processing unit configured to Includes a decryption unit.

Referring to the fourth aspect, in a possible design, the predicted motion information includes forward predicted motion information and backward predicted motion information, and the motion vector predictor includes forward motion vector predictor and backward motion vector predictor. The predictive reference image block includes a forward predictive reference image block and a backward predictive reference image block.

Referring to the fourth aspect, in a possible design, the initial decoding prediction block includes a forward initial decoding prediction block and a backward initial decoding prediction block, and is bidirectionally predicted with respect to the first decoding block based on the predicted motion information. To obtain the initial decoding prediction block of the first decoding block, the processing unit specifically performs forward prediction on the first decoding block based on the forward prediction motion information. To acquire the forward initial decoding prediction block of the first decoding block, execute backward prediction for the first decoding block based on the backward prediction motion information, and acquire the backward initial decoding prediction block of the first decoding block. It is composed of.

Referring to the fourth aspect, in a possible design, when acquiring the first decoding prediction block of the first decoding block based on the initial decoding prediction block, the processing unit specifically, the forward initial decoding prediction block. And the backward initial decoding prediction block is weighted and added to obtain the first decoding prediction block of the first decoding block, or the front initial decoding prediction block is used as the first decoding prediction block of the first decoding block. Used or configured to use the backward initial decoding prediction block as the first decoding prediction block of the first decoding block.

Referring to the fourth aspect, in a possible design, the second decoding prediction block includes a front second decoding prediction block and a rear second decoding prediction block, based on the motion vector predictor, with respect to the prediction reference image block. When performing a first-precision motion search to obtain at least one second decoding prediction block, the processing unit specifically enters the forward prediction reference image block based on the forward motion vector predictor. On the other hand, a motion search with the first accuracy is performed, and each found forward decoding prediction block is used as a forward second decoding prediction block to obtain at least one second decoding prediction block, and a backward motion vector predictor is obtained. Performs a first-precision motion search on the backward-predicted reference image block based on, and uses each found backward-decoded predictive block as the backward-second decoded predicted block to at least one second-decided predictor. It is configured to acquire a block, where the first precision includes an integer pixel precision, a 1/2 pixel precision, a 1/4 pixel precision or a 1/8 pixel precision.

Referring to the fourth aspect, in a possible design, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, with a first decoding prediction block and each second decoding prediction block. When calculating the difference between and using the motion vector predictor between the first and second decoding blocks with the smallest difference as the target motion vector predictor for the first decoding block, the processing unit , Specifically, the difference between the forward second decoding prediction block and the first decoding prediction block is compared, and the target forward movement between the first decoding block and the forward second decoding prediction block having the smallest difference. Use the vector predictor as the target motion vector predictor, or compare the differences between the backward second decode predictor block and the first decode predictor block and compare the differences between the first decode block and the backward second with the smallest difference. The target backward motion vector predictor to and from the decryption predictor block is configured to be used as the target motion vector predictor.

Referring to the fourth aspect, in a possible design, the third decoding prediction block of the first decoding block includes a forward third decoding prediction block and a rear third decoding prediction block, based on the target motion vector predictor. When performing bidirectional prediction on the first decoding block to get the third decoding prediction block on the first decoding block, the processing unit specifically bases it on the target forward motion vector predictor. Perform forward prediction on one decoding block, get the third forward decoding prediction block on the first decoding block, and perform backward prediction on the first decoding block based on the target backward motion vector predictor. Therefore, it is configured to acquire the rear third decoding prediction block of the first decoding block.

Referring to the fourth aspect, in a possible design, when acquiring the target decoding prediction block of the first decoding block based on the third decoding prediction block, the processing unit specifically predicts the forward third decoding. Weighted addition is performed on the block and the rear third decoding prediction block to obtain the target decoding prediction block of the first decoding block, or the front third decoding prediction block is the target decoding prediction block of the first decoding block. Or the backward third decoding prediction block is configured to be used as the target decoding prediction block of the first decoding block.

According to the fifth aspect, a coding device is provided, wherein the device includes a processor and a memory, where the memory stores a computer-readable program and the processor executes a program in the memory to the first aspect. Such a coding method is carried out.

According to a sixth aspect, a decryption device is provided, wherein the device includes a processor and a memory, where the memory stores a computer-readable program and the processor executes a program in the memory according to the second aspect. Implement the decryption method.

According to a seventh aspect, a computer storage medium is provided, the computer storage medium being configured to store computer software instructions for the first and second aspects, wherein the computer software instructions are: Includes a program designed to perform the aforementioned embodiments.

The third to seventh aspects of the embodiments of the present invention are consistent with the technical solutions of the first and second embodiments of the present invention, and the beneficial effects and corresponding benefits obtained by all the embodiments. It should be understood that the feasible design schemes are similar. Details will not be explained again.

1A and 1B are schematics for selecting predicted motion information for the current block in merged and non-merged modes, respectively. 1A and 1B are schematics for selecting predicted motion information for the current block in merged and non-merged modes, respectively. FIG. 2 is a schematic block diagram of a video codec device or electronic device. FIG. 3 is a schematic block diagram of a video codec system. FIG. 4 is a flowchart of a method for image coding via inter-prediction according to an embodiment of the present invention. FIG. 5 is a flowchart of a method for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 6A is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 6B is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 7 is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 8A is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 8B is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 8C is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 8D is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 9A is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 9B is a schematic diagram of a process for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 10 is a block diagram of an apparatus for image coding via inter-prediction according to an embodiment of the present invention. FIG. 11 is a block diagram of an image encoder using inter-prediction according to an embodiment of the present invention. FIG. 12 is a configuration diagram of an apparatus for image decoding via inter-prediction according to an embodiment of the present invention. FIG. 13 is a block diagram of an image decoder using inter-prediction according to an embodiment of the present invention.

Hereinafter, the technical solutions in the embodiments of the present invention will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the embodiments described are not all, but some of the embodiments of the invention. All other embodiments obtained by one of ordinary skill in the art based on embodiments of the invention without creative effort should be included in the scope of protection of the invention.

Inter-prediction techniques are categorized into Merge mode and non-merge mode, and non-merge mode is further categorized into SKIP mode and non-skip mode.

In non-merge mode and non-skip mode of merge mode in inter-prediction technology, the redundant information of the current image block is removed based on the pixel information of the coded or decoded image adjacent to the current image in order to obtain the residuals. Will be done.

In the inter-prediction technology, in order to effectively remove the redundant information of the current image block, the encoder side or the decoder side can reduce the residual of the current image block from the reference image to the current image block. You need to get the image block that most closely resembles. The encoder side or the decoder side usually acquires an image block of the above-mentioned reference image via motion estimation. Existing video coding and decoding techniques typically search for a reference image for the appropriate matching image block on an image block basis during coding and decoding in the motion estimation process. In some embodiments, the image block is a predictive unit of the image to be encoded by the encoder side or a predictive unit of the image to be reconstructed by the decoder side. In some embodiments, the image block is a conversion unit of the image to be encoded by the encoder side or a conversion unit of the image to be reconstructed by the decoder side. In some embodiments, the image block is a coding unit or coding sub-unit of the image to be encoded by the encoder side, or a decoding unit or decoding sub-unit of the image to be reconstructed by the decoder side. Is. No restrictions are imposed.

In the non-merge mode of the inter-prediction technology, the encoder side performs subtraction on the pixel value corresponding to the current image block and the pixel value corresponding to the matching image block to obtain the residual, and the residual is Entropy coding is performed on the values obtained after transformation and quantization, and finally the bitstream obtained via entropy coding and the motion offset obtained via exploration (ie, motion vector difference). ) Are written in the code string. Correspondingly, during predictive compensation (or motion compensation), the decoder side first obtains an entropy-coded bit stream and then performs entropy coding to perform the corresponding residuals and the corresponding motion. Get the vector difference, then get the motion vector value based on the obtained motion vector difference and the motion vector value of the adjacent block, and get the corresponding matching image block from the reference image based on the motion vector value. , The pixel value corresponding to the matching image block and the pixel value corresponding to the residual are added to obtain the value of the current decoded image block.

In the non-skip mode of the merge mode of the inter-prediction technology, the encoder side performs subtraction on the pixel value corresponding to the current image block and the pixel value corresponding to the matching image block to obtain the residual. Entropy coding is performed on the values obtained after the residuals are transformed and quantized, and finally the bitstream obtained via entropy coding is written to the code string. Correspondingly, during predictive compensation (or motion compensation), the decoder side first obtains the entropy-coded bitstream and then performs entropy-coding to obtain the corresponding residuals. Based on the motion vector value of the adjacent block, the corresponding matching image block is obtained from the reference image, the pixel value corresponding to the matching image block and the pixel value corresponding to the residual are added, and the current image block decoded is obtained. Get the value of.

In the skip mode of the merge mode of the inter-prediction technique, the residuals and motion vector differences do not need to be acquired to save bit rate. In both the coding and decoding processes, the corresponding matching image block is directly acquired based on the motion information of the adjacent block and used as the value of the current image block.

As shown in Figure 1A, in merge mode, adjacent blocks are marked as different index values, which are the adjacent blocks whose motion vector information is used by the current block as the predicted motion information for the current block. Used to determine. As shown in Figure 1B, in non-merge mode, the motion vector predictor list is constructed by detecting adjacent blocks, and the motion vector predictor is now from the motion vector predictor list for the current block. Selected as the motion vector predictor for the block of. It can be seen that in interframe coding and decoding, the motion vector information of the adjacent block is used directly as the motion vector predictor of the current block. This method of directly acquiring the motion vector predictor of the inter-prediction inevitably causes a deviation of the predicted motion accuracy.

FIG. 2 is a schematic block diagram of a video codec device or electronic device 50. The device or electronic device may be incorporated into the codec according to the embodiment of the present invention. FIG. 3 is a schematic device diagram of a video codec system according to an embodiment of the present invention. The units of FIGS. 2 and 3 will be described below.

The electronic device 50 may be, for example, a mobile terminal or a user device in a wireless communication system. It should be understood that embodiments of the invention may be performed by any electronic device or device that may need to encode or decode, or encode and decode, the video image.

The device 50 may include a housing built into the device and configured to protect the device. The device 50 may further include a display 32 in the form of a liquid crystal display. In another embodiment of the invention, the display may be any suitable display technique applied to an image or video display. The device 50 may further include a keypad 34. In another embodiment of the invention, any suitable data or user interface mechanism may be used. For example, the user interface may be implemented as a virtual keyboard or data entry system to act as part of a touch-sensitive display. The device may include a microphone 36 or any suitable audio input, which may be a digital or analog signal input. The device 50 may further include the following audio output devices: In this embodiment of the invention, the audio output device may be any one of a headset 38, a speaker, an analog audio output connection or a digital audio output connection. The device 50 may also include a battery 40. In another embodiment of the invention, the device may be powered by any suitable mobile energy device, such as a solar cell, fuel cell or clock mechanism generator. The device may further include an infrared port 42 configured to perform short-range line-of-sight communication with another device. In another embodiment, the device 50 may further include any suitable short-range communication solution, such as a Bluetooth® wireless connection or a USB / Liveline wired connection.

The device 50 may include a controller 56 or a processor configured to control the device 50. The controller 56 may be connected to the memory 58. In the present embodiment of the present invention, the memory may store image data and audio data and / or may store instructions executed on the controller 56. The controller 56 may also be connected to a codec circuit 54 adapted to perform audio and / or video data coding and decoding, or the controller 56 may be a codec circuit for performing auxiliary coding and decoding. Connected to 54.

The device 50 may further include a card reader 48 or smart card 46 configured to provide user information and adapted to provide information for network authentication and authorized user authentication.

The device 50 may further include a wireless interface circuit 52. The radio interface circuit is connected to a controller and is adapted to generate a radio communication signal used, for example, to communicate with a cellular communication network, a radio communication system or a radio local area network. The device 50 may further include an antenna 44. The antenna is connected to the radio interface circuit 52 to transmit the radio frequency signal generated by the radio interface circuit 52 to other (s) devices and to receive the radio frequency signal from the other (s) devices. ..

In some embodiments of the invention, the device 50 comprises a camera capable of recording or detecting a single frame. The codec 54 or controller receives and processes these single frames. In some embodiments of the invention, the device may receive video image data to be processed from another device prior to transmission and / or storage. In some embodiments of the invention, the device 50 may receive an image for encoding / decoding over a wireless or wired connection.

FIG. 3 is a schematic block diagram of another video codec system 10 according to an embodiment of the present invention. As shown in FIG. 3, the video codec system 10 includes a source device 12 and a destination device 14. The source device 12 produces coded video data. Therefore, the source device 12 may be referred to as a video coding device or a video coding device. The destination device 14 may decode the coded video data generated by the source device 12. Therefore, the destination device 14 may be referred to as a video decoding device or a video decoding device. The source device 12 and the destination device 14 may be examples of a video codec device or a video codec device. The source device 12 and the destination device 14 are desktop computers, mobile computing devices, notebooks (such as smartphones, televisions, cameras, display devices, digital media players, video game machines, in-vehicle computers or other similar devices, etc. It may include a wide range of devices, including, for example, laptop) computers, tablet computers, settop boxes, and handsets.

The destination device 14 may receive encoded video data from the source device 12 via the channel 16. Channel 16 may include one or more media and / or devices capable of transmitting encoded video data from the source device 12 to the destination device 14. In one example, channel 16 may include one or more communication media that allow the source device 12 to directly transmit the encoded video data to the destination device 14 in real time. In this example, the source device 12 may modulate the encoded video data according to a communication standard (eg, a wireless communication protocol) and may transmit the modulated video data to the destination device 14. One or more communication media may include radio and / or wired communication media such as radio frequency (RF) spectra or one or more physical transmission lines. The one or more communication media may form part of a packet-based network (eg, a local area network, a wide area network or a global network (eg, the Internet)). One or more communication media may include a router, switch, base station or another device that facilitates communication from the source device 12 to the destination device 14.

In another example, channel 16 may include a storage medium for storing the coded video data produced by the source device 12. In this example, the destination device 14 can access the storage medium via disk access or card access. The storage medium may include various local access data storage media such as Blu-ray, DVD, CD-ROM, flash memory or other suitable digital storage media configured to store encoded video data.

In another example, channel 16 may include another intermediate storage device that stores the coded video data generated by the file server or source device 12. In this example, the destination device 14 may access the encoded video data stored in the file server or another intermediate storage device via streaming transmission or download. The file server may be of a server type capable of storing encoded video data and transmitting the encoded video data to the destination device 14. Examples of file servers include web servers (eg, apply to websites), file transfer protocol (FTP) servers, network attached storage (NAS) devices and local disk drives.

The destination device 14 may access the encoded video data via a standard data connection (eg, an internet connection). Examples of data connection types are wireless channels (eg, Wi-Fi connections), wired connections (eg, DSL or cable modems) or, adapted to access encoded video data stored on a file server. Includes a combination of both. The transmission of the coded video data from the file server may be a streaming transmission, a download transmission, or a combination thereof.

The technique of the present invention is not limited to wireless application scenarios. For example, this technology can be used for over-the-air television broadcasting, cable television transmission, satellite television transmission, streaming video transmission (eg, over the Internet), and the encoding of video data stored on data storage media. It may be used to support video coding and decoding in a variety of multimedia applications, including conversion, decoding of video data stored on data storage media and other uses. In some examples, the video codec system 10 may be configured to support unidirectional or two-way video transmission, thereby applications such as video streaming transmission, video playback, video broadcasting and / or video telephone communication. To support.

In the example in FIG. 3, the source device 12 includes a video source 18, a video coder 20, and an output interface 22. In some examples, the output interface 22 may include a modulator / demodulator (modem) and / or a transmitter. The video source 18 has a video capture device (eg, a video camera), a video archive containing previously captured video data, a video input interface configured to receive video data from a video content provider, and / or video data. It may include a computer graphic system configured to generate or a combination of the video data sources described above.

The video coder 20 can encode the video data from the video source 18. In some examples, the source device 12 sends the encoded video data directly to the destination device 14 by using the output interface 22. The encoded video data may also be stored on a storage medium or file server for subsequent access by the destination device 14 for decoding and / or playback.

In the example in FIG. 3, the destination device 14 includes an input interface 28, a video decoder 30, and a display device 32. In some examples, the input interface 28 includes a receiver and / or a modem. The input interface 28 may receive encoded video data via channel 16. The display device 32 may be integrated into the destination device 14 or may be located outside the destination device 14. Normally, the display device 32 displays the decoded video data. The display device 32 may include various display devices such as liquid crystal displays (LCDs), plasma displays, organic light emitting diode (OLED) displays or other types of display devices.

The video coder 20 and the video decoder 30 may operate according to a video compression standard (eg, High Efficiency Video Coding H.265 standard) and may be HEVC test model (HM) compliant. The H.265 standard text description ITU-TH.265 (V3) (04/2015) was promulgated on April 29, 2015, from http://handle.itu.int/11.1002/1000/12455. It can be downloaded and the entire contents of the file are incorporated herein by reference.

Alternatively, the video coder 20 and the video decoder 30 may operate according to different ownership or industry standards. Standards are ITU-T H.261, ISO / IEC MPEG-1 Visual, ITU-T H.262, ISO / IEC MPEG-2 Visual, ITU-T H.263, ISO / IEC MPEG-4 Visual or ITU- Includes T H.264 (also known as ISO / IEC MPEG-4 AVC) and includes scalable video coding (SVC) and multiview video coding (MVC) extensions. It should be understood that the techniques of the invention are not limited to any particular codec standard or technique.

In addition, FIG. 3 is merely an example, and the techniques of the invention apply to video codec applications (eg, one-sided video coding or video decoding) that do not necessarily include any data communication between the encoding and decoding devices. You can do it. In another example, the data is retrieved from local memory, the data is transmitted via network streaming transmission, or the data is manipulated in a similar manner. The coding device may encode the data and store the data in the memory, and / or the decoding device may retrieve the data from the memory and decode the data. In many examples, coding and decoding are performed by multiple devices that do not communicate with each other, but only encode the data into and / or retrieve the data from the memory and decode the data.

The video coder 20 and video decoder 30 are each one or more microprocessors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), discrete logic, hardware or these. It may be implemented as any one of a plurality of suitable circuits, such as any combination. If the technology is implemented partially or entirely by using the software, the device may store the software's instructions in a suitable non-temporary computer-readable storage medium, with one or more processors. By using it, instructions may be executed in hardware to perform the techniques of the present invention. Any one of the above items (including hardware, software, hardware and software combinations, etc.) may be considered as one or more processors. Each of the video coder 20 and the video decoder 30 may be contained in one or more codecs or decoders, and either the video coder 20 and the video decoder 30 may be combined in a separate device / codec / decoder. It may be integrated as part of a decoder (CODEC).

In the present invention, it may be shown that the video coder 20 typically uses a signal to transmit information to another device (eg, a video decoder 30). The term "transmit by using a signal" may usually refer to a syntax element and / or may refer to the transmission of encoded video data. Transmission may occur in real time or near real time. Alternatively, this communication may occur over a period of time, for example, when the syntactic elements are stored in a computer-readable storage medium by using the coded binary data during coding. The syntax element may be retrieved by the decoder at any time after it has been stored on the medium.

As shown in FIG. 4, embodiments of the present invention provide a method for image coding via inter-prediction. The specific procedure is as follows.

Step 401: The predictive motion information of the first coded block is determined, where the predictive motion information includes a motion vector predictor and the predictive reference image information, and the predictive reference image information represents the predictive reference image block. used.

The first coded block is the block to be coded that should be processed now, the motion vector predictor in the predicted motion information includes a forward motion vector predictor and a backward motion vector predictor, and the predicted reference image information is , Includes reference frame index information for forward predictive reference image blocks and reference frame index information for backward predictive reference image blocks.

Further, the predictive reference image block of the first coded block is acquired based on the motion vector predictor of the first coded block or the predictive reference image information.

Optionally, the predicted motion information of the first coded block is a forward motion vector predictor, a backward motion vector predictor, a Picture Order Count (POC) corresponding to the forward predictive reference image block, and a backward predictive reference image block. May include POCs corresponding to, or only POCs corresponding to forward motion vector predictors and forward predictive reference image blocks, or contain only POCs corresponding to backward motion vector predictors and backward predictive reference image blocks. It's fine.

Specifically, for the method for determining the predicted motion information of the first coded block, refer to the method for determining the predicted motion information of the current coded block in the prior art.

When the inter-prediction merge mode is used, the candidate predictive motion information list is constructed based on the motion information of the adjacent block of the first coded block, and the candidate predictive motion information is the first from the candidate predictive motion information list. It is selected as the predicted motion information of the coded block. The candidate motion information list includes motion vector predictors, reference frame index information of reference image blocks, and the like. It should be noted that the motion vector predictor in this mode is the motion vector value of the adjacent block.

When the non-merge mode of interprediction is used, the motion vector predictor list is constructed based on the motion information of the adjacent blocks of the first coded block, and the motion vector is the first code from the motion vector predictor list. Selected as the motion vector predictor of the conversion block. Note that the motion vector predictor in this mode may be the motion vector value of the adjacent block or the sum of the motion vector values of the selected adjacent block and the motion vector difference of the first coded block. It should be. The motion vector difference is the difference between the motion vector obtained by performing motion estimation on the first coded block and the motion vector value of the selected adjacent block.

Step 402: Based on the predicted motion information, perform bidirectional prediction on the first coded block to get the initial coded predictive block of the first coded block, and based on the initial coded predictive block, Acquires the first coded prediction block of the first coded block.

Furthermore, if the POC corresponding to the predictive reference image block is not equal to the POC of the first coded block, bidirectional predictions, including forward and backward predictions, will be relative to the first coded block based on the predicted motion information. Will be executed. If the POC corresponding to the predictive reference image block is less than the POC of the first coded block, forward prediction is performed for the first coded block based on the predicted motion information, or the predictive reference image block If the corresponding POC is greater than the POC of the first coded block, a backward prediction is performed for the first coded block based on the predicted motion information.

Specifically, the step of executing bidirectional prediction for the first coded block based on the predicted motion information and acquiring the forward initial coded prediction block of the first coded block involves the following process. include.

S421. Based on the predicted motion information, forward prediction is executed for the first coded block, specifically, forward prediction is executed for the first coded block based on the forward predicted motion information, and the first 1 Get the forward initial coded prediction block of the coded block.

In a possible implementation, forward prediction is performed on the first coded block based on the forward predicted reference image block in the predicted motion information to obtain the forward initial coded predictive block of the first coded block.

In another possible implementation, a forward prediction is performed on the first coded block based on the forward motion vector predictor in the predicted motion information to obtain the forward initial coded predictive block of the first coded block. ..

S422. Based on the predicted motion information, the backward prediction is executed for the first coded block, specifically, the backward prediction is executed for the first coded block based on the backward predicted motion information. 1 Get the backward initial coded prediction block of the coded block.

In a possible implementation, forward prediction is performed on the first coded block based on the backward predictive reference image block in the predicted motion information to obtain the forward initial coded predictive block of the first coded block.

In another possible implementation, a forward prediction is performed on the first coded block based on the backward motion vector predictor in the predicted motion information to obtain the forward initial coded predictor block of the first coded block. ..

The initial coding prediction block includes a front initial coding prediction block and a rear initial coding prediction block.

Further, the step of acquiring the first coded prediction block of the first coded block based on the initial coded prediction block includes the following three implementations.

In the first implementation, weighted addition is performed on the forward initial coded prediction block and the backward initial coded predictive block to obtain the first coded predictive block of the first coded block.

In this implementation, optionally, the average value of the forward initial coded prediction block and the backward initial coded predictive block is used as the first coded predictive block of the first coded block.

In the second implementation, the forward initial coding prediction block is used as the first coding prediction block of the first coding block.

In the third implementation, the backward initial coding prediction block is used as the first coding prediction block of the first coding block.

Step 403: Perform motion estimation with first accuracy on the predictive reference image block based on the motion vector predictor to obtain at least one second coded predictor block.

In some embodiments, this step is to perform a first-precision motion search on the predictive reference image block, where the search position for the motion search is the motion vector predictor and the first. Determined by using precision. In a feasible practice, the motion search position is around the position represented by the motion vector predictor and within the first accuracy range. For example, if the position indicated by the motion vector predictor is (1, 1), the first accuracy is 1/2 pixel accuracy and the motion search position is (1.5, 1), (1.5, 1.5). , (1.5, -0.5), (1, 1), (1, 1.5), (1, -0.5), (-0.5, 1), (-0.5, 1.5) and (-0.5, -0.5). ..

Specifically, based on the motion vector predictor, the steps to perform a first-precision motion estimate for a predictive reference image block and obtain at least one second coded predictor block are as follows: including:

S431. Based on the forward motion vector predictor, a motion search with the first accuracy is performed on the forward predictive reference image block, and each forward coded predictive block found is used as the forward second coded predictive block. Get at least one second coded prediction block.

S432. Based on the backward motion vector predictor, a motion search with the first accuracy is performed on the backward predictive reference image block, and each found backward coded predictive block is used as the backward second coded predictive block. At least one second coded prediction block is acquired, where the second coded predictive block includes a forward second coded predictive block and a backward second coded predictive block.

The first precision includes integer pixel precision, 1/2 pixel precision, 1/4 pixel precision or 1/8 pixel precision. This is not limited.

Step 404: Calculate the difference between the 1st coded prediction block and each 2nd coded prediction block, and predict the motion vector between the 1st coded block and the 2nd coded prediction block with the smallest difference. Use the child as the target motion vector predictor for the first coded block.

Specifically, the difference between the first coded prediction block and each second coded prediction block is calculated, and the movement between the first coded block and the second coded prediction block having the smallest difference. The step of using the vector predictor as the target motion vector predictor of the first coded block involves the following process:

S441. The target forward motion vector prediction between the first coded block and the second forward coded prediction block with the smallest difference by comparing the differences between the forward second coded prediction block and the first coded predictive block. Use the child as a target motion vector predictor.

S442. The target backward motion vector prediction between the first coded block and the second backward coded prediction block with the smallest difference by comparing the differences between the backward second coded prediction block and the first coded predictive block. The child is used as a target motion vector predictor, where the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor.

Specifically, when the difference between the second coded prediction block and the first coded prediction block is compared, the sum of the absolute values of the pixel differences in the two image blocks is the same as the second coded prediction block. It may be used as the value of the difference from the first coded prediction block. Optionally, the sum of the squared values of the pixel differences in the two image blocks may be used as the value of the difference between the second coded prediction block and the first coded prediction block. The difference comparison method is not particularly limited.

Step 405: Perform bidirectional prediction on the first coded block based on the target motion vector predictor to get the third coded predictor block of the first coded block.

Specifically, the step of executing bidirectional prediction for the first coded block based on the target motion vector predictor and acquiring the third coded prediction block of the first coded block is as follows. Including process:

S451. Based on the target forward motion vector predictor, forward prediction is performed on the first coded block to obtain the forward third coded prediction block of the first coded block.

S452. Based on the target backward motion vector predictor, perform backward prediction on the 1st coded block to get the 3rd backward coded predictive block of the 1st coded block, where the 1st coded block The third coded prediction block of the above includes a front third coded prediction block and a rear third coded prediction block.

Step 406: Obtain the target coding prediction block of the first coding block based on the third coding prediction block, and encode the first coding block based on the target coding prediction block.

Specifically, the step of acquiring the target coding prediction block of the first coding block based on the third coding prediction block includes the following three implementations.

In the first implementation, weighted addition is performed on the forward third coded prediction block and the rear third coded predictive block to obtain the target coded predictive block of the first coded block.

In this implementation, optionally, the average value of the forward third coded predictive block and the backward third coded predictive block is used as the target coded predictive block of the first coded block.

In the second implementation, the forward third coded predictive block is used as the target coded predictive block of the first coded block.

In the third implementation, the rear third coded predictive block is used as the target coded predictive block of the first coded block.

In some embodiments, the flag bits may be encoded in the upper layer information of the code sequence such as slice level, picture level or sequence level, and the flag bits are applied to the image coding method shown in FIG. It should be understood that it is used to indicate whether or not.

As shown in FIG. 5, embodiments of the present invention provide a method for image decoding via inter-prediction. The specific procedure is as follows.

Step 501: Determine the predicted motion information of the first decoded block, where the predicted motion information includes a motion vector predictor and the predicted reference image information, and the predicted reference image information is used to represent the predicted reference image block. Will be done.

The first decoding block is the block to be processed now that needs to be constructed based on the analyzed code string information, and the motion vector predictors in the predicted motion information are the forward motion vector predictor and the backward motion. Includes a vector predictor, the predictive reference image information includes reference frame index information for the forward predictive reference image block and reference frame index information for the backward predictive reference image block.

Further, the predictive reference image block of the first coded block is acquired based on the motion vector predictor or the predictive reference image information of the first decoded block.

Optionally, the predicted motion information of the first decoded block can be a forward motion vector predictor, a backward motion vector predictor, a Picture Order Count (POC) corresponding to the forward predictive reference image block, and a backward predictive reference image block. It may contain the corresponding POC, or may contain only the POC corresponding to the forward motion vector predictor and the forward predictive reference image block, or may contain only the POC corresponding to the backward motion vector predictor and the backward predictive reference image block. good.

Specifically, for the method for determining the predicted motion information of the first decoded block, refer to the method for determining the predicted motion information of the current decoded block in the prior art.

When the inter-prediction merge mode is used, the candidate prediction motion information list is constructed based on the motion information of the adjacent block of the first decoding block, and the candidate prediction motion information is first decoded from the candidate prediction motion information list. Selected as the predicted motion information for the block. The candidate motion information list includes motion vector predictors, reference frame index information of reference image blocks, and the like. It should be noted that the motion vector predictor in this mode is the motion vector value of the adjacent block.

When the non-merge mode of interprediction is used, the motion vector predictor list is constructed based on the motion information of the adjacent blocks of the first decoded block, and the motion vector is from the motion vector predictor list to the first decoded block. Selected as the motion vector predictor of. Note that the motion vector predictor in this mode may be the motion vector value of the adjacent block or the sum of the motion vector values of the selected adjacent block and the motion vector difference of the first decoded block. Should be. The motion vector difference may be obtained directly via decoding.

Step 502: Perform bidirectional prediction on the first decoding block based on the predicted motion information to obtain the initial decoding prediction block of the first decoding block, and based on the initial decoding prediction block, the first decoding block. Get the first decryption prediction block of.

Further, if the POC corresponding to the predictive reference image block is not equal to the POC of the first decoded block, bidirectional prediction including forward and backward predictions is performed for the first decoded block based on the predicted motion information. NS. If the POC corresponding to the predictive reference image block is less than the POC of the first decoded block, forward prediction is performed for the first decoded block based on the predicted motion information, or it corresponds to the predicted reference image block. If the POC is greater than the POC of the first decoded block, a backward prediction is performed for the first decoded block based on the predicted motion information.

Specifically, the step of executing bidirectional prediction for the first decoding block based on the predicted motion information and acquiring the forward initial decoding prediction block of the first decoding block includes the following process.

S521. Based on the predicted motion information, forward prediction is executed for the first decoding block, specifically, forward prediction is executed for the first decoding block based on the forward predicted motion information, and the first decoding is performed. Get the forward initial decryption prediction block of the block.

In a possible implementation, forward prediction is performed on the first decoding block based on the forward prediction reference image block in the predicted motion information to obtain the forward initial decoding prediction block of the first decoding block.

In another possible implementation, forward prediction is performed on the first decoding block based on the forward motion vector predictor in the predicted motion information to obtain the forward initial decoding prediction block of the first decoding block.

S522. Based on the predicted motion information, the backward prediction is executed for the first decoding block, specifically, the backward prediction is executed for the first decoding block based on the backward predicted motion information, and the first decoding is performed. Get the backward initial decryption prediction block of the block.

In a possible implementation, forward prediction is performed on the first decoding block based on the backward prediction reference image block in the predicted motion information to obtain the forward initial decoding prediction block of the first decoding block.

In another possible implementation, forward prediction is performed on the first decoding block based on the backward motion vector predictor in the predicted motion information to obtain the forward initial decoding prediction block of the first decoding block.

The initial decoding prediction block includes a front initial decoding prediction block and a rear initial decoding prediction block.

Further, the step of acquiring the first decoding prediction block of the first decoding block based on the initial decoding prediction block includes the following three implementations.

In the first implementation, weighted addition is performed on the forward initial decoding prediction block and the backward initial decoding prediction block to acquire the first decoding prediction block of the first decoding block.

In this implementation, optionally, the average value of the forward initial decoding prediction block and the backward initial decoding prediction block is used as the first decoding prediction block of the first decoding block.

In the second implementation, the forward initial decoding prediction block is used as the first decoding prediction block of the first decoding block.

In the third implementation, the backward initial decoding prediction block is used as the first decoding prediction block of the first decoding block.

Step 503: Based on the motion vector predictor, perform motion estimation with first accuracy on the predictive reference image block to obtain at least one second decoded predictor block.

In some embodiments, this step is to perform a first-precision motion search on the predictive reference image block, where the search position for the motion search is the motion vector predictor and the first. Determined by using precision. In a feasible practice, the motion search position is around the position represented by the motion vector predictor and within the first accuracy range. For example, if the position indicated by the motion vector predictor is (1, 1), the first accuracy is 1/2 pixel accuracy and the motion search position is (1.5, 1), (1.5, 1.5). , (1.5, -0.5), (1, 1), (1, 1.5), (1, -0.5), (-0.5, 1), (-0.5, 1.5) and (-0.5, -0.5). ..

Specifically, based on the motion vector predictor, the steps to perform a first-precision motion estimate for a predictive reference image block and obtain at least one second decode predictor block are as follows: include:

S531. Based on the forward motion vector predictor, perform a first precision motion search on the forward predictive reference image block and use each found forward decode predictor block as the forward second decode predictor block at least one. Get two second decryption prediction blocks.

S532. Based on the backward motion vector predictor, perform a first precision motion search on the backward predictive reference image block and use each found backward decode predictor block as the backward second decode predictor block, at least one. Two second decoding prediction blocks are acquired, where the second decoding prediction block includes a front second decoding prediction block and a rear second decoding prediction block.

The first precision includes integer pixel precision, 1/2 pixel precision, 1/4 pixel precision or 1/8 pixel precision. This is not limited.

Step 504: Calculate the difference between the first decryption predictor block and each second decode predictor block, and obtain the motion vector predictor between the first decode predictor block and the second decode predictor block having the smallest difference. 1 Used as a target motion vector predictor for the decoding block.

Specifically, the difference between the first decoding prediction block and each second decoding prediction block is calculated, and the motion vector predictor between the first decoding block and the second decoding prediction block having the smallest difference is calculated. , The steps used as the target motion vector predictor in the first decoding block include the following processes:

S541. The target forward motion vector predictor between the first decoding block and the second forward decoding prediction block with the smallest difference is compared by comparing the difference between the forward second decoding prediction block and the first decoding prediction block. Used as a motion vector predictor.

S542. The target backward motion vector predictor between the first and second decoding prediction blocks with the smallest difference is targeted by comparing the differences between the backward second decoding prediction block and the first decoding prediction block. Used as a motion vector predictor, where the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor.

Specifically, when the differences between the second decoding prediction block and the first decoding prediction block are compared, the sum of the absolute values of the pixel differences in the two image blocks is the second decoding prediction block and the first decoding. It may be used as the value of the difference from the prediction block. Optionally, the sum of the squared values of the pixel differences in the two image blocks may be used as the value of the difference between the second decoding prediction block and the first decoding prediction block. The difference comparison method is not particularly limited.

Step 505: Based on the target motion vector predictor, perform bidirectional prediction on the first decoding block to get the third decoding prediction block of the first decoding block.

Specifically, the step of performing bidirectional prediction on the first decoding block based on the target motion vector predictor to obtain the third decoding prediction block of the first decoding block includes the following process: :

S551. Based on the target forward motion vector predictor, the forward prediction is executed for the first decoding block, and the forward third decoding prediction block of the first decoding block is acquired.

S552. Based on the target backward motion vector predictor, a backward prediction is performed on the first decoding block to obtain the backward third decoding prediction block of the first decoding block, where the third decoding of the first decoding block is obtained. The prediction block includes a front third decoding prediction block and a rear third decoding prediction block.

Step 506: Acquire the target decoding prediction block of the first decoding block based on the third decoding prediction block, and decode the first decoding block based on the target decoding prediction block.

Specifically, the step of acquiring the target decoding prediction block of the first decoding block based on the third decoding prediction block includes the following three implementations.

In the first implementation, weighted addition is performed on the front third decoding prediction block and the rear third decoding prediction block to acquire the target decoding prediction block of the first decoding block.

In this implementation, optionally, the average value of the front third decoding prediction block and the rear third decoding prediction block is used as the target decoding prediction block of the first decoding block.

In the second implementation, the forward third decoding prediction block is used as the target decoding prediction block of the first decoding block.

In the third implementation, the rear third decoding prediction block is used as the target decoding prediction block of the first decoding block.

In some embodiments, corresponding to the encoder side, the flag bits may be analyzed first in the upper layer information of the code sequence such as strip level, image level or sequence level, and the flag bits are shown in FIG. It should be understood that it is used to indicate whether the indicated image decoding method applies.

In the following, the coding method in FIG. 4 and the decoding method in FIG. 5 will be described in detail by using some specific embodiments.

Embodiment 1

As shown in FIGS. 6A and 6B, the current decoding block is the first decoding block, and the predicted motion information of the current decoding block is acquired. The forward and backward motion vector predictors of the current decoding block are (-10, 4) and (5, 6), respectively, and the POC corresponding to the current decoding block is 4, which corresponds to the forward prediction reference image block. Assume that the POC is 2 and the POC corresponding to the backward predictive reference image block is 6.

Forward and backward predictions are performed separately for the current decoding block, and the initial forward decoding prediction block (Forward Prediction Block, FPB) and the initial backward decoding prediction block (Backward Prediction Block, BPB) of the current decoding block are executed. Obtain and assume that the initial forward decoding prediction block and the initial backward decoding prediction block are FPB 1 and BPB 1, respectively. It is assumed that weighting is performed on FPB 1 and BPB 1 to obtain the first Decoding Prediction Block (DPB) of the current decoding block, and that the first decoding prediction block is DPB 1.

(-10, 4) and (5, 6) are used as reference inputs for the forward and backward motion vector predictors, and the first precision motion search is for the forward and backward predictive reference image blocks. Executed separately. In this case, the first accuracy is 1/2 pixel accuracy within one pixel range. The first decryption prediction block DPB 1 is used as a reference. The difference between the first decoding prediction block DPB 1 and each of the corresponding new forward and backward decoding prediction blocks found through motion search corresponds to the new decoding prediction block with the smallest difference from DPB 1. Forward and backward motion vector predictors are acquired and compared to act as target motion vector predictors. The forward and backward motion vector predictors are assumed to be (-11, 4) and (6, 6), respectively.

The target motion vector predictor is updated to (-11, 4) and (6, 6). In addition, based on the target motion vector predictor, forward and backward predictions are performed on the first decoding block, and weighting is performed on the new forward and backward decoding prediction blocks obtained. , Get the target decryption prediction block. The target decryption prediction block is assumed to be DPB 2, and the decoding prediction block of the current decoding block is updated to DPB 2.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Embodiment 2

As shown in FIGS. 6A and 6B, the current coded block is the first coded block, and the predicted motion information of the current coded block is acquired. The forward and backward motion vector predictors of the current coded block are (-10, 4) and (5, 6), respectively, and the POC corresponding to the current coded block is 4, which is the forward predictive reference image block. Assume that the corresponding POC is 2 and the corresponding POC for the backward predictive reference image block is 6.

Forward and backward predictions are performed separately for the current coded block to get the initial forward and backward coded predictive blocks and the initial forward coded predictive blocks and the initial forward coded predictive blocks of the current coded block. It is assumed that the initial backward coded prediction blocks are FPB 1 and BPB 1, respectively. It is assumed that weighting is performed on FPB 1 and BPB 1 to get the first coded predictive block of the current coded block, and that the first coded predictive block is DPB 1.

(-10, 4) and (5, 6) are used as reference inputs for the forward and backward motion vector predictors, and the first precision motion search is for the forward and backward predictive reference image blocks. Executed separately. In this case, the first accuracy is 1/2 pixel accuracy within one pixel range. The first coded prediction block DPB 1 is used as a reference. The difference between the first coded predictive block DPB 1 and each of the corresponding new forward and backward coded predictive blocks found via motion search is the smallest from DPB 1, the new coded predictive block. The forward and backward motion vector predictors corresponding to are obtained and compared to serve as the target motion vector predictors. The forward and backward motion vector predictors are assumed to be (-11, 4) and (6, 6), respectively.

The target motion vector predictor is updated to (-11, 4) and (6, 6). In addition, forward and backward predictions are performed on the first coded block based on the target motion vector predictor, and weighted addition is performed on the new forward and backward coded predictive blocks obtained. Executed to get the target coded prediction block. Assuming that the target coded predictive block is DPB 2, the coded predictive block of the current coded block is updated to DPB 2.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Embodiment 3

As shown in FIG. 7, the current decoding block is the first decoding block, and the predicted motion information of the current decoding block is acquired. It is assumed that the forward motion vector predictor of the current decoded block is (-21, 18), the POC corresponding to the current decoded block is 4, and the POC corresponding to the forward predicted reference image block is 2.

It is assumed that forward prediction is performed on the current decoding block to get the initial forward decoding prediction block of the current decoding block and that the initial forward decoding prediction block is FPB 1. In this case, FPB 1 is used as the first decoding prediction block of the current decoding block, and the first decoding prediction block is shown as DPB 1.

(-21, 18) is used as the reference input for the forward motion vector predictor, and the first precision motion search is performed on the forward predictive reference image block. In this case, the first accuracy is 1 pixel accuracy within the range of 5 pixels. The first decryption prediction block DPB 1 is used as a reference. The difference between the first decoding prediction block DPB 1 and the corresponding new backward decoding prediction block found via motion search is the forward motion vector prediction corresponding to the new decoding prediction block with the smallest difference from DPB 1. The children are obtained and compared to act as the target motion vector predictor. The forward motion vector predictor is assumed to be (-19, 19).

The target motion vector predictor is updated at (-19, 19). In addition, forward prediction is performed on the first decoding block based on the target motion vector predictor, and the new forward decoding prediction block obtained is used as the target decoding prediction block. The target decryption prediction block is assumed to be DPB 2, and the decoding prediction block of the current decoding block is updated to DPB 2.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Embodiment 4

As shown in FIG. 7, the current coded block is the first coded block, and the predicted motion information of the current coded block is acquired. Assume that the forward motion vector predictor of the current coded block is (-21, 18), the POC corresponding to the current coded block is 4, and the POC corresponding to the forward predictive reference image block is 2. do.

It is assumed that forward prediction is performed on the current coded block to get the initial forward coded predictive block of the current coded block and that the initial forward coded predictive block is FPB 1. In this case, FPB 1 is used as the first coded predictive block of the current coded block, and the first coded predictive block is shown as DPB 1.

(-21, 18) is used as the reference input for the forward motion vector predictor, and the first precision motion search is performed on the forward predictive reference image block. In this case, the first accuracy is 1 pixel accuracy within the range of 5 pixels. The first coded prediction block DPB 1 is used as a reference. The difference between the first coded predictive block DPB 1 and the corresponding new backward coded predictive block found via motion search is the forward corresponding to the new coded predictive block with the smallest difference from DPB 1. Motion vector predictors are acquired and compared to act as target motion vector predictors. The forward motion vector predictor is assumed to be (-19, 19).

The target motion vector predictor is updated at (-19, 19). In addition, a forward prediction is performed on the first coded block based on the target motion vector predictor, and the new forward coded predictive block obtained is used as the target coded predictive block. Assuming that the target coded predictive block is DPB 2, the coded predictive block of the current coded block is updated to DPB 2.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Embodiment 5

As shown in FIGS. 8A, 8B, 8C and 8D, the current coded block is the first coded block, and the predicted motion information of the current coded block is acquired. The forward and backward motion vector predictors of the current coded block are (-6, 12) and (8, 4), respectively, and the POC corresponding to the current coded block is 8, which is the forward predictive reference image block. Assume that the corresponding POC is 4 and the corresponding POC for the backward predictive reference image block is 12.

Forward and backward predictions are performed separately for the current coded block to get the initial forward and backward coded predictive blocks and the initial forward coded predictive blocks and the initial forward coded predictive blocks of the current coded block. It is assumed that the initial backward coded prediction blocks are FPB 1 and BPB 1, respectively. It is assumed that weighting is performed on FPB 1 and BPB 1 to get the first coded predictive block of the current coded block, and that the first coded predictive block is DPB 1.

(-6, 12) and (8, 4) are used as reference inputs for the forward and backward motion vector predictors, and the first precision motion search is for the forward and backward predictive reference image blocks. Executed separately. The first coded prediction block DPB 1 is used as a reference. The difference between the first coded predictive block DPB 1 and each of the corresponding new forward and backward coded predictive blocks found via motion search is the smallest from DPB 1, the new coded predictive block. The forward and backward motion vector predictors corresponding to are obtained and compared to serve as the target motion vector predictors. The forward and backward motion vector predictors are assumed to be (-11, 4) and (6, 6), respectively.

The target motion vector predictor is updated to (-11, 4) and (6, 6). In addition, forward and backward predictions are performed on the first coded block based on the target motion vector predictor, and weighted addition is performed on the new forward and backward coded predictive blocks obtained. Executed to get the target coded prediction block. Assuming that the target coded predictive block is DPB 2, the coded predictive block of the current coded block is updated to DPB 2.

Then (-11, 4) and (6, 6) are used as reference inputs for the forward and backward motion vector predictors, and the first precision motion search is in the forward and backward predictive reference image blocks. On the other hand, it is executed separately. The coded prediction block DPB 2 of the current coded block is used as a reference. The difference between the first coded predictive block DPB 2 and each of the corresponding new forward and backward coded predictive blocks found via motion search is the smallest from DPB 2, the new coded predictive block. The corresponding forward and backward motion vector predictors are obtained and compared to serve as the new target motion vector predictors. The forward and backward motion vector predictors are assumed to be (-7, 11) and (6, 5), respectively.

The target motion vector predictor is then updated to (-7, 11) and (6, 5). In addition, forward and backward predictions are performed on the first coded block based on the latest target motion vector predictor, and the new forward and backward coded predictive blocks obtained are weighted. Addition is performed to get the target coded prediction block. Assuming that the target coded predictive block is DPB 3, the coded predictive block of the current coded block is updated to DPB 3.

Further, the target motion vector predictor may then be updated according to the method described above, so that the number of iterations is not limited.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

Embodiment 6

In the present embodiment, the method of acquiring the motion vector predictor in the non-merge mode coding process of the inter-prediction is different from that of the above-described embodiment. Embodiment 6 describes in detail the decryption process in the non-merge mode of interprediction. Similarly, the coding and decoding processes are similar in the non-merged mode of interprediction, and details are not described again here.

As shown in FIGS. 9A and 9B, the current decoding block is the first decoding block, and the predicted motion information of the current decoding block is acquired. The forward and backward motion vector values of the current decoded block are (-10, 4) and (5, 6), respectively, and the forward and backward motion vector differences are (-2, 1) and (1, 1), respectively. Assume that the POC corresponding to the current decryption block is 4, the POC corresponding to the forward predictive reference image block is 2, and the POC corresponding to the backward predictive reference image block is 6.

Forward and backward predictions are executed separately for the current decoding block to obtain the initial forward decoding prediction block (-FPB) and the initial backward decoding prediction block (BPB) of the current decoding block, and the initial forward decoding prediction. It is assumed that the block and the initial backward decoding prediction block are FPB 1 and BPB 1, respectively. It is assumed that weighting is performed on FPB 1 and BPB 1 to obtain the first decoding prediction block (DPB) of the current decoding block, and that the first decoding prediction block is DPB 1.

Forward and backward motion vector predictors (-10, 4) + (-2, 1) = (-12, 5) and (5, 6) + (1, 1) = (6, 7) are forward and backward motion Used as a vector reference input, the first precision motion search is performed separately for the forward predictive reference image block and the backward predictive reference image block. In this case, the first accuracy is 1/4 pixel accuracy within one pixel range. The first decryption prediction block DPB 1 is used as a reference. The difference between the first decoding prediction block DPB 1 and each of the corresponding new forward and backward decoding prediction blocks found through motion search corresponds to the new decoding prediction block with the smallest difference from DPB 1. Forward and backward motion vectors are acquired and compared to act as a target motion vector predictor. The forward and backward motion vectors are assumed to be (-11, 4) and (6, 6), respectively.

The target motion vector predictor is updated to (-11, 4) and (6, 6). In addition, based on the target motion vector predictor, forward and backward predictions are performed on the first decoding block, and weighting is performed on the new forward and backward decoding prediction blocks obtained. , Get the target decryption prediction block. The target decryption prediction block is assumed to be DPB 2, and the decoding prediction block of the current decoding block is updated to DPB 2.

If a motion search of first precision is performed on the forward predictive reference image block and the backward predictive reference image block, the first precision may be any particular precision, eg, integer pixel precision, 1 /. It should be noted that it may be 2-pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy.

In conclusion, the coding and decoding methods described above may be applied to all bidirectional prediction processes in video image coding and decoding, with updated motion vector predictors and coding and decoding prediction blocks for all videos. It may be applied to image coding and decoding processes. The codec system obtains the current coded and decoded predicted blocks based on the predicted motion information of the current coded and decoded blocks and performs a motion search based on the motion vector predictor and the predicted reference image block. Gets a new coding and decoding prediction block and updates the motion vector predictor and the coding and decoding prediction block of the current coding and decoding block through a differential comparison with the current coding and decoding prediction block. , Thereby improving the motion accuracy.

According to the aforementioned embodiments, as shown in FIG. 10, embodiments of the present invention provide an apparatus 1000 for image coding via inter-prediction. As shown in FIG. 10, the apparatus 1000 includes a decision unit 1001, a processing unit 1002, and a coding unit 1003.

The determination unit 1001 is configured to determine the predicted motion information of the first coded block, where the predicted motion information includes a motion vector predictor and the predicted reference image information, and the predicted reference image information is a predicted reference. Used to represent an image block.

The processing unit 1002 performs bidirectional prediction on the first coded block based on the predicted motion information, acquires the initial coded predictive block of the first coded block, and is based on the initial coded predictive block. Then, the first coded prediction block of the first coded block is acquired, the motion search of the first accuracy is performed on the predicted reference image block, and at least one second coded predicted block is acquired. Here, the search position of the motion search is determined by using the motion vector predictor and the first precision, calculating the difference between the first coded prediction block and each second coded prediction block. The motion vector predictor between the first coded block and the second coded predictor with the smallest difference is used as the target motion vector predictor for the first coded block and is based on the target motion vector predictor. , A bidirectional prediction is performed on the first coded block to obtain the third coded prediction block of the first coded block.

The coding unit 1003 acquires the target coding prediction block of the first coding block based on the third coding prediction block, and encodes the first coding block based on the target coding prediction block. It is composed of.

Optionally, the predictive motion information includes forward predictive motion information and backward predictive motion information, the motion vector predictor includes forward motion vector predictors and backward motion vector predictors, and the predictive reference image block is a forward predictive reference image. Includes blocks and backward prediction reference image blocks.

Optionally, the initial coded prediction block includes a forward initial coded predictive block and a backward initial coded predictive block, performing bidirectional predictions on the first coded block based on the predicted motion information. When acquiring the initial coded prediction block of one coded block, the processing unit 1002 specifically
Based on the forward prediction motion information, the forward prediction is executed for the first coded block to obtain the forward initial coded prediction block of the first coded block.
Backward Prediction Based on the motion information, it is configured to perform backward prediction on the first coded block and acquire the backward initial coded predictive block of the first coded block.

Optionally, when acquiring the first coded predictive block of the first coded block based on the initial coded predictive block, the processing unit 1002 specifically,
Perform weighting addition on the forward initial coded prediction block and the backward initial coded predictive block to get the first coded predictive block of the first coded block, or
The forward initial coding prediction block is used as the first coding prediction block of the first coding block, or
The backward initial coding prediction block is configured to be used as the first coding prediction block of the first coding block.

Optionally, the second coded prediction block includes a forward second coded predictive block and a backward second coded predictive block, with first precision movement relative to the predictive reference image block based on the motion vector predictor. When performing a search to obtain at least one second coded prediction block, the processing unit 1002 specifically,
Based on the forward motion vector predictor, a motion search with the first accuracy is performed on the forward predictive reference image block, and each forward coded predictive block found is used as the forward second coded predictive block. Get at least one second coded prediction block and
Based on the backward motion vector predictor, a motion search with the first accuracy is performed on the backward predictive reference image block, and each found backward coded predictive block is used as the backward second coded predictive block. It is configured to get at least one second coded prediction block.

The first precision includes integer pixel precision, 1/2 pixel precision, 1/4 pixel precision or 1/8 pixel precision.

Optionally, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, calculates the difference between the first coded predictor block and each second coded predictor block, and minimizes. When the motion vector predictor between the first coded block and the second coded predictor with the difference is used as the target motion vector predictor of the first coded block, the processing unit 1002 specifically teeth,
The target forward motion vector prediction between the first coded block and the second forward coded prediction block with the smallest difference by comparing the differences between the forward second coded prediction block and the first coded predictive block. Use the child as a target motion vector predictor, or
The target backward motion vector prediction between the first coded block and the second backward coded predictive block with the smallest difference by comparing the differences between the backward second coded predictive block and the first coded predictive block. The child is configured to be used as a target motion vector predictor.

Optionally, the third coded predictor block of the first coded block includes a forward third coded predictor block and a backward third coded predictor block, and is based on the target motion vector predictor to the first coded block. On the other hand, when performing bidirectional prediction to obtain the third coded prediction block of the first coded block, the processing unit 1002 specifically,
Based on the target forward motion vector predictor, perform forward prediction on the first coded block to get the forward third coded predictive block of the first coded block.
Based on the target backward motion vector predictor, it is configured to perform backward prediction on the first coded block to get the backward third coded predictive block of the first coded block.

Optionally, when acquiring the target coding prediction block of the first coding block based on the third coding prediction block, the processing unit 1002 specifically,
Perform weighting addition on the forward 3rd coded prediction block and the backward 3rd coded predictive block to get the target coded predictive block of the 1st coded block, or
The forward third coded predictive block is used as the target coded predictive block of the first coded block, or
The rear third coded predictive block is configured to be used as the target coded predictive block of the first coded block.

It should be noted that the functional implementation and dialogue scheme of each unit of the apparatus 1000 in the present embodiment of the present invention is further referred to in the embodiments of the related methods. Details will not be explained here again.

According to the same concept of the invention, as shown in FIG. 11, embodiments of the present invention further provide a encoder 1100. As shown in FIG. 11, the encoder 1100 includes a processor 1101 and a memory 1102. The program code used to execute the solution of the present invention is stored in memory 1102 and is used to instruct processor 1101 to perform the coding method shown in FIG.

The code corresponding to the method shown in Figure 4 may also be incorporated into the chip by performing design programming for the processor, and as a result, when the chip is executed, the method shown in Figure 4 is executed. Can be done.

According to the aforementioned embodiments, as shown in FIG. 12, embodiments of the present invention provide device 1200 for image decoding via inter-prediction. As shown in FIG. 12, the apparatus 1200 includes a decision unit 1201, a processing unit 1202 and a decoding unit 1203.

The determination unit 1201 is configured to determine the predicted motion information of the first decoding block, where the predicted motion information includes a motion vector predictor and the predicted reference image information, and the predicted reference image information is the predicted reference image. Used to represent a block.

The processing unit 1202 performs bidirectional prediction on the first decoding block based on the predicted motion information, acquires the initial decoding prediction block of the first decoding block, and based on the initial decoding prediction block, the first Obtain the first decoding prediction block of the decoding block, perform the first precision motion search on the prediction reference image block, obtain at least one second decoding prediction block, and search for the motion search here. The position is determined by using the motion vector predictor and the first precision, the difference between the first decode predictor block and each second decode predictor block is calculated, and the first decode block with the smallest difference. Use the motion vector predictor between It is configured to execute and acquire the third decryption prediction block of the first decryption block.

The decoding unit 1203 is configured to acquire the target decoding prediction block of the first decoding block based on the third decoding prediction block and decode the first decoding block based on the target decoding prediction block.

Optionally, the predictive motion information includes forward predictive motion information and backward predictive motion information, the motion vector predictor includes forward motion vector predictors and backward motion vector predictors, and the predictive reference image block is a forward predictive reference image. Includes blocks and backward prediction reference image blocks.

Optionally, the initial decoding prediction block includes a forward initial decoding prediction block and a backward initial decoding prediction block, and based on the predicted motion information, performs bidirectional prediction on the first decoding block to perform bidirectional prediction on the first decoding block. When acquiring the initial decryption prediction block, the processing unit 1202 specifically
Based on the forward prediction motion information, the forward prediction is executed for the first decoding block, and the forward initial decoding prediction block of the first decoding block is acquired.
Based on the backward prediction motion information, the backward prediction is executed for the first decoding block, and the backward initial decoding prediction block of the first decoding block is acquired.

Optionally, when acquiring the first decoding prediction block of the first decoding block based on the initial decoding prediction block, the processing unit 1202 specifically,
Perform weighting addition on the forward initial decoding prediction block and the backward initial decoding prediction block to obtain the first decoding prediction block of the first decoding block, or
The forward initial decoding prediction block is used as the first decoding prediction block of the first decoding block, or
The backward initial decoding prediction block is configured to be used as the first decoding prediction block of the first decoding block.

Optionally, the second decoding prediction block includes an anterior second decoding prediction block and a rear second decoding prediction block, and performs a first-precision motion search for the predictive reference image block based on the motion vector predictor. Then, when acquiring at least one second decryption prediction block, the processing unit 1202 specifically
Based on the forward motion vector predictor, perform a first precision motion search on the forward predictive reference image block and use each found forward decode predictor block as the forward second decode predictor block at least one. Get two second decryption prediction blocks,
Based on the backward motion vector predictor, perform a first precision motion search on the backward predictive reference image block and use each found backward decode predictor block as the backward second decode predictor block, at least one. It is configured to get two second decryption prediction blocks.

The first precision includes integer pixel precision, 1/2 pixel precision, 1/4 pixel precision or 1/8 pixel precision.

Optionally, the target motion vector predictor includes a target forward motion vector predictor and a target backward motion vector predictor, calculates the difference between the first decode predictor block and each second decode predictor, and minimizes the difference. When the motion vector predictor between the first decoding block and the second decoding prediction block having is used as the target motion vector predictor of the first decoding block, the processing unit 1202 specifically:
The target forward motion vector predictor between the first decoding block and the second forward decoding prediction block with the smallest difference is compared by comparing the difference between the forward second decoding prediction block and the first decoding prediction block. Used as a motion vector predictor or
The target backward motion vector predictor between the first and second decoding prediction blocks with the smallest difference is targeted by comparing the differences between the backward second decoding prediction block and the first decoding prediction block. It is configured to be used as a motion vector predictor.

Optionally, the third decoding prediction block of the first decoding block includes a forward third decoding prediction block and a backward third decoding prediction block, and is bidirectionally predicted with respect to the first decoding block based on the target motion vector predictor. When the processing unit 1202 specifically obtains the third decoding prediction block of the first decoding block by executing
Based on the target forward motion vector predictor, perform forward prediction on the first decoding block to get the forward third decoding prediction block of the first decoding block.
Based on the target backward motion vector predictor, it is configured to perform backward prediction on the first decoding block and obtain the backward third decoding prediction block of the first decoding block.

Optionally, when acquiring the target decoding prediction block of the first decoding block based on the third decoding prediction block, the processing unit 1202 specifically,
Perform weighting addition on the forward 3rd decryption prediction block and the backward 3rd decoding prediction block to get the target decryption prediction block of the 1st decoding block, or
The forward third decoding prediction block is used as the target decoding prediction block of the first decoding block, or
The backward third decoding prediction block is configured to be used as the target decoding prediction block of the first decoding block.

It should be noted that the functional implementation and dialogue scheme of each unit of the apparatus 1200 in the present embodiment of the present invention is further referred to in the embodiments of the related methods. Details will not be explained here again.

It should be understood that the division of units in equipment 1200 and 1000 is merely a logical functional division. In practice, all or part of the unit may be integrated into one physical entity, or the units may be physically separated. For example, the above-mentioned units may be separately arranged processing elements, may be integrated into the controller chip for implementation, or may be stored in the controller's memory elements in the form of program code, as described above. May be called by a processing element of the controller to perform the function of the unit. In addition, the units may be integrated or implemented separately. The processing element may be an integrated circuit chip and has signal processing capability. In the implementation process, the steps in the method described above or the units described above can be performed by using hardware integrated logic circuits in the processing elements, or by using instructions in the form of software. be able to. The processing element may be a general-purpose processor such as a central processing unit (CPU), or one or more application-specific integrated circuits (CPU). ASIC), one or more microprocessors (English: digital signal processor, abbreviated as DSP) or one or more field-programmable gate array (English: field-programmable gate array, abbreviated as FPGA), etc. May be configured as one or more integrated circuits that perform.

According to the same invention concept, embodiments of the present invention further provide the decoder 1300. As shown in FIG. 13, the decoder 1300 includes a processor 1301 and a memory 1302. The program code used to execute the solutions of the present invention is stored in memory 1302 and is used to instruct processor 1301 to perform the decoding method shown in FIG.

The code corresponding to the method shown in Figure 5 may also be incorporated into the chip by performing design programming for the processor, and as a result, when the chip is executed, the method shown in Figure 5 Can be executed.

The processor in the encoder 1100 and the decoder 1300 in embodiments of the invention may be one or more integrated circuits configured to control program execution in a CPU, DSP, ASIC or solution of the invention. That can be understood. One or more memory contained in a computer system is read-only memory (ROM) or another type of static storage that can store static information and static instructions. , Random access memory (RAM for short) or another type of dynamic storage capable of storing information and instructions, or may be magnetic disk memory. These memories may be connected to the processor by using a bus or may be connected to the processor by using a dedicated connection cable.

One of ordinary skill in the art can understand that all or part of the steps in the aforementioned method of the embodiment may be performed by a program instructing the processor. The above-mentioned program may be stored in a computer-readable storage medium. Storage media are random access memory, read-only memory, flash memory, hard disk, solid state drive, magnetic tape, floppy disk, optical disc, or any of these. It is a non-transitory (English: non-transitory) medium such as a combination of.

The present invention will be described with reference to the respective flowcharts and block diagrams of the methods and devices in the embodiments of the present invention. It should be understood that computer program instructions may be used to implement each process and each block in flowcharts and block diagrams, and a combination of processes and blocks in flowcharts and block diagrams. These computer program instructions may be provided to a general purpose computer, a dedicated computer, an embedded processor or the processor of any other programmable data processing device to generate a machine, and as a result, any other programmable. Instructions executed by the computer or processor of a data processing device generate a device for performing a particular function in one or more processes in a flowchart, or in one or more blocks in a block diagram. ..

The above description is merely an exemplary practice of the invention, but is not intended to limit the scope of protection of the invention. Any modifications or substitutions readily conceived by one of ordinary skill in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be in accordance with the scope of claims.

32 display
34 keypad
42 infrared port
52 Wireless interface
54 codec
56 controller
58 memory
12 Source device
14 Destination device
18 Video source
20 video coder
22 Output interface
28 Input interface
30 Video decoder
32 Display device
1000 equipment
1001 decision unit
1002 processing unit
1003 coding unit
1100 encoder
1101 processor
1102 memory
1200 equipment
1201 decision unit
1202 processing unit
1203 decryption unit
1300 decoder
1301 processor
1302 memory

Claims (16)

A method for image coding via inter prediction,
A step of determining the predicted motion information of the first coded block, wherein the predicted motion information includes a motion vector predictor and a predicted reference image information, and the predicted reference image information represents a predicted reference image block. Used, steps and
A step of executing bidirectional prediction on the first coded block based on the predicted motion information and acquiring a front or back coded prediction block of the first coded block.
A step of executing a motion search with the first accuracy for the prediction reference image block to acquire at least one second coded prediction block, wherein the search position of the motion search is the motion vector predictor. And determined by using the first precision, the first precision includes a step and a step, including an integer pixel precision, a 1/2 pixel precision, a 1/4 pixel precision or a 1/8 pixel precision.
The difference is calculated based on the forward or backward coding prediction block and each second coding prediction block, and the motion vector predictor corresponding to the second coding prediction block having the smallest difference is obtained by the first coding. The steps to use as the block's target motion vector predictor, and
A step of executing bidirectional prediction on the first coded block based on the target motion vector predictor to obtain a third coded prediction block of the first coded block.
Based on the third coded prediction block to obtain a target coding prediction blocks of the first encoded block, on the basis of the target encoding prediction block, and a step that turn into code the first coded block How to include. The predicted motion information includes forward predicted motion information and backward predicted motion information, the motion vector predictor includes a forward motion vector predictor and a backward motion vector predictor, and the predictive reference image block is a forward predictive reference image. The method of claim 1, comprising a block and a backward predictive reference image block. The third coded prediction block of the first coded block includes a forward third coded predictive block and a backward third coded predictive block, and is based on the target motion vector predictor. The step of executing bidirectional prediction for the first coded block to obtain the third coded prediction block of the first coded block is
A step of executing forward prediction for the first coded block based on the target forward motion vector predictor to acquire the forward third coded predictive block of the first coded block.
A claim comprising performing a backward prediction on the first coded block based on the target backward motion vector predictor to obtain the rear third coded predictive block of the first coded block. The method described in Item 1. The step of acquiring the target coding prediction block of the first coding block based on the third coding prediction block is
A step of performing weighting addition to the front third coded prediction block and the rear third coded prediction block to obtain the target coded prediction block of the first coded block, or
A step of using the forward third coding prediction block as the target coding prediction block of the first coding block, or
The method of claim 3, comprising the step of using the rear third coded predictor block as the target coded predictor block of the first coded block. A method for image decoding via inter-prediction,
A step of determining the predicted motion information of the first decoded block, wherein the predicted motion information includes a motion vector predictor and a predicted reference image information, and the predicted reference image information is used to represent a predicted reference image block. To be done, steps and
Based on the predicted motion information, a step of executing bidirectional prediction on the first decoding block and acquiring a forward or backward decoding prediction block of the first decoding block, and
A step of executing a motion search with the first accuracy for the prediction reference image block to acquire at least one second decoding prediction block, wherein the search position of the motion search is the motion vector predictor and the motion vector predictor. Determined by using the first precision, the first precision includes a step and a step, including an integer pixel precision, a 1/2 pixel precision, a 1/4 pixel precision or a 1/8 pixel precision.
The difference is calculated based on the forward or backward decoding prediction block and each second decoding prediction block, and the motion vector predictor corresponding to the second decoding prediction block having the minimum difference is the target motion of the first decoding block. Steps to use as vector predictors and
Based on the target motion vector predictor, a step of executing bidirectional prediction on the first decoding block to acquire a third decoding prediction block of the first decoding block, and
A method including a step of acquiring a target decoding prediction block of the first decoding block based on the third decoding prediction block and decoding the first decoding block based on the target decoding prediction block. The predicted motion information includes forward predicted motion information and backward predicted motion information, the motion vector predictor includes a forward motion vector predictor and a backward motion vector predictor, and the predictive reference image block is a forward predictive reference image. 5. The method of claim 5, comprising blocks and backward predictive reference image blocks. The third decoding prediction block of the first decoding block includes a front third decoding prediction block and a rear third decoding prediction block, and is bidirectional to the first decoding block based on the target motion vector predictor. The step of executing the prediction and acquiring the third decoding prediction block of the first decoding block is
A step of executing forward prediction for the first decoding block based on the target forward motion vector predictor to acquire the forward third decoding prediction block of the first decoding block.
5. The fifth aspect of the present invention includes a step of executing backward prediction for the first decoding block based on the target backward motion vector predictor to acquire the rear third decoding prediction block of the first decoding block. The method described. The step of acquiring the target decoding prediction block of the first decoding block based on the third decoding prediction block is
A step of performing weighting addition to the front third decoding prediction block and the rear third decoding prediction block to acquire the target decoding prediction block of the first decoding block, or
A step of using the forward third decoding prediction block as the target decoding prediction block of the first decoding block, or
The method of claim 7, wherein the rear third decoding prediction block is used as the target decoding prediction block of the first decoding block. An apparatus for image coding via inter prediction,
A determination unit configured to determine the predicted motion information of the first coded block, wherein the predicted motion information includes a motion vector predictor and a predictive reference image information, and the predictive reference image information is a predictive reference. The decision unit used to represent the image block, and
Based on the predicted motion information, bidirectional prediction is executed for the first coded block, the front or back coded prediction block of the first coded block is acquired, and the prediction reference image block is used. A first precision motion search is performed to obtain at least one second coded prediction block, and the search position for the motion search is determined by using the motion vector predictor and the first accuracy. The first accuracy includes an integer pixel accuracy, a 1/2 pixel accuracy, a 1/4 pixel accuracy or a 1/8 pixel accuracy, and the front or back coding prediction block and each second coding prediction block. The difference is calculated based on, and the motion vector predictor corresponding to the second coded predictor having the smallest difference is used as the target motion vector predictor of the first coded block, and the target motion vector predictor is used. Based on, a processing unit configured to perform bidirectional prediction on the first coded block and acquire the third coded prediction block of the first coded block.
Based on the third coded prediction block to obtain a target coding prediction blocks of the first encoded block, on the basis of the target encoding prediction block, constructed so that turn into code the first coded block A device that includes a coding unit. The predicted motion information includes forward predicted motion information and backward predicted motion information, the motion vector predictor includes a forward motion vector predictor and a backward motion vector predictor, and the predictive reference image block is a forward predictive reference image. 9. The apparatus of claim 9, comprising a block and a backward predictive reference image block. The third coded prediction block of the first coded block includes a forward third coded predictive block and a backward third coded predictive block, and is based on the target motion vector predictor. When bidirectional prediction is performed on the first coded block to obtain the third coded prediction block of the first coded block, the processing unit receives the third coded prediction block.
Based on the target forward motion vector predictor, forward prediction is performed on the first coded block to obtain the forward third coded prediction block of the first coded block.
Based on the target backward motion vector predictor, it is configured to perform backward prediction on the first coded block to obtain the rear third coded predictive block of the first coded block. The device according to claim 9. When acquiring the target coding prediction block of the first coding block based on the third coding prediction block, the processing unit receives the target coding prediction block.
Weighting addition is performed on the front third coded prediction block and the rear third coded prediction block to obtain the target coded prediction block of the first coded block, or
The forward third coding prediction block is used as the target coding prediction block of the first coding block, or
11. The apparatus of claim 11, wherein the rear third coded predictor block is configured to be used as the target coded predictor block of the first coded block. A device for image decoding via inter-prediction,
A determination unit configured to determine the predicted motion information of the first decoded block, wherein the predicted motion information includes a motion vector predictor and a predicted reference image information, and the predicted reference image information is a predicted reference image. The decision unit used to represent the block, and
Based on the predicted motion information, bidirectional prediction is executed for the first decoding block, the front or rear decoding prediction block of the first decoding block is acquired, and the first prediction reference image block is obtained. An accurate motion search is performed to obtain at least one second decoding prediction block, and the search position for the motion search is determined by using the motion vector predictor and the first accuracy, said first. The accuracy of 1 includes integer pixel accuracy, 1/2 pixel accuracy, 1/4 pixel accuracy or 1/8 pixel accuracy, and the difference is calculated based on the front or rear decoding prediction block and each second decoding prediction block. The motion vector predictor corresponding to the second decoding predictor having the smallest difference is used as the target motion vector predictor of the first decoding block, and the first decoding is based on the target motion vector predictor. A processing unit configured to perform bidirectional prediction on the block and acquire the third decoding prediction block of the first decoding block.
A decoding unit configured to acquire the target decoding prediction block of the first decoding block based on the third decoding prediction block and to decode the first decoding block based on the target decoding prediction block. Equipment including. The predicted motion information includes forward predicted motion information and backward predicted motion information, the motion vector predictor includes a forward motion vector predictor and a backward motion vector predictor, and the predictive reference image block is a forward predictive reference image. 13. The apparatus of claim 13, comprising a block and a backward predictive reference image block. The third decoding prediction block of the first decoding block includes a front third decoding prediction block and a rear third decoding prediction block, and is bidirectional to the first decoding block based on the target motion vector predictor. When the prediction is executed and the third decoding prediction block of the first decoding block is acquired, the processing unit receives the third decoding prediction block.
Based on the target forward motion vector predictor, forward prediction is executed for the first decoding block to acquire the forward third decoding prediction block of the first decoding block.
13. Claim 13 configured to perform backward prediction on the first decoding block based on the target backward motion vector predictor to obtain the rear third decoding prediction block of the first decoding block. The device described in. When acquiring the target decoding prediction block of the first decoding block based on the third decoding prediction block, the processing unit receives the target decoding prediction block.
Weighting addition is performed on the front third decoding prediction block and the rear third decoding prediction block to acquire the target decoding prediction block of the first decoding block, or
The forward third decoding prediction block is used as the target decoding prediction block of the first decoding block, or
The apparatus according to claim 15, wherein the rear third decoding prediction block is configured to be used as the target decoding prediction block of the first decoding block.

Images