KR102669273B1 - Methods and apparatus for video transform encoding/decoding - Google Patents

Abstract

The video decoding method according to the present invention acquires context information on the decoding target block, uses the context information to derive an inverse transformation method for the decoding target block, and uses the derived inverse transformation method to perform inverse transformation on the decoding target block. Perform. According to the present invention, image compression performance can be improved while minimizing the amount of transmitted information.

Description

Video transformation encoding/decoding method and device {METHODS AND APPARATUS FOR VIDEO TRANSFORM ENCODING/DECODING}

The present invention relates to image processing, and more specifically to image encoding/decoding methods and devices.

Recently, as broadcasting services with HD (High Definition) resolution have expanded not only domestically but also globally, many users are becoming accustomed to high-resolution, high-definition video, and many organizations are accelerating the development of next-generation video devices. In addition, as interest in UHD (Ultra High Definition), which has a resolution more than four times that of HDTV, increases along with HDTV, compression technology for higher resolution and higher quality images is required.

Image compression technologies include inter prediction technology, which predicts pixel values included in the current picture from temporally previous and/or subsequent pictures, and intra prediction technology, which predicts pixel values included in the current picture using pixel information within the current picture. (intra) prediction technology, conversion technology that decomposes the original image signal into a signal in the frequency domain, inverse conversion technology that restores the signal in the frequency domain to the original image signal, assigning short codes to symbols with high frequency of occurrence and frequency of occurrence There is an entropy coding technique that assigns long codes to low-value symbols. Video data can be efficiently compressed using these video compression technologies.

The technical object of the present invention is to provide an image encoding method and device that can improve image compression performance while minimizing the amount of information transmitted.

Another technical object of the present invention is to provide an image decoding method and device that can improve image compression performance while minimizing the amount of information transmitted.

Another technical problem of the present invention is to provide a conversion method and device that can improve image compression performance while minimizing the amount of information transmitted.

Another technical problem of the present invention is to provide an inverse conversion method and device that can improve image compression performance while minimizing the amount of information transmitted.

Another technical problem of the present invention is to provide an entropy encoding method and device that can improve image compression performance while minimizing the amount of information transmitted.

Another technical problem of the present invention is to provide an entropy decoding method and device that can improve image compression performance while minimizing the amount of information transmitted.

A video decoding method is provided. The method uses at least one of information included in a decoding target block, a block already decoded before decoding the decoding target block, and pixel values of a pixel already restored before decoding the decoding target block, to decode the decoding target block. Obtaining context information for, deriving an inverse transformation method for the decoding target block using the context information, and performing inverse transformation on the decoding target block using the derived inverse transformation method, In the step of deriving an inverse transformation method for the decoding target block using the context information, the inverse transformation method is derived using at least one of a method of selecting one of a plurality of inverse transformation methods and a method of updating an initial inverse transformation equation. The plurality of inverse transformation methods are inverse transformation methods already determined before decoding the decoding target block, and the initial inverse transformation equation is an inverse transformation formula already determined before decoding the decoding target block.

According to the video encoding method according to the present invention, video compression performance can be improved while minimizing the amount of transmitted information.

According to the video decoding method according to the present invention, video compression performance can be improved while minimizing the amount of transmitted information.

According to the conversion method according to the present invention, image compression performance can be improved while minimizing the amount of transmitted information.

According to the inverse conversion method according to the present invention, image compression performance can be improved while minimizing the amount of transmitted information.

According to the entropy encoding method according to the present invention, image compression performance can be improved while minimizing the amount of transmitted information.

According to the entropy decoding method according to the present invention, image compression performance can be improved while minimizing the amount of transmitted information.

Figure 1 is a block diagram showing the configuration of a video encoding device according to an embodiment.
Figure 2 is a block diagram showing the configuration of a video decoding device according to an embodiment.
Figure 3 is a flowchart schematically showing a conversion method according to an embodiment of the present invention.
Figure 4 is a flowchart schematically showing an entropy encoding method according to an embodiment of the present invention.
Figure 5 is a flowchart schematically showing an entropy decoding method according to an embodiment of the present invention.
Figure 6 is a flowchart schematically showing an inverse transformation method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In describing the embodiments of the present specification, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present specification, the detailed description will be omitted.

When a component is said to be “connected” or “connected” to another component, it is understood that it may be directly connected or connected to the other component, but that other components may exist in between. It should be. In addition, the description of “including” a specific configuration in the present invention does not exclude configurations other than the configuration, and means that additional configurations may be included in the practice of the present invention or the scope of the technical idea of the present invention.

Terms such as first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, a first component may be named a second component, and similarly, the second component may also be named a first component without departing from the scope of the present invention.

Additionally, the components appearing in the embodiments of the present invention are shown independently to show different characteristic functions, and this does not mean that each component is comprised of separate hardware or a single software component. That is, each component is listed and included as a separate component for convenience of explanation, and at least two of each component can be combined to form one component, or one component can be divided into a plurality of components to perform a function, and each of these components can be divided into a plurality of components. Integrated embodiments and separate embodiments of the constituent parts are also included in the scope of the present invention as long as they do not deviate from the essence of the present invention.

Additionally, some components may not be essential components that perform essential functions in the present invention, but may simply be optional components to improve performance. The present invention can be implemented by including only essential components for implementing the essence of the present invention excluding components used only to improve performance, and a structure including only essential components excluding optional components used only to improve performance. is also included in the scope of rights of the present invention.

Figure 1 is a block diagram showing the configuration of a video encoding device according to an embodiment. Referring to FIG. 1, the image encoding device 100 includes a motion prediction unit 111, a motion compensation unit 112, an intra prediction unit 120, a switch 115, a subtractor 125, and a conversion unit 130. , a quantization unit 140, an entropy encoding unit 150, an inverse quantization unit 160, an inverse transform unit 170, an adder 175, a filter unit 180, and a reference image buffer 190.

The image encoding device 100 may perform encoding on an input image in intra mode or inter mode and output a bit stream. Intra prediction refers to prediction within a screen, and inter prediction refers to prediction between screens. In the case of intra mode, the switch 115 is switched to intra, and in the case of inter mode, the switch 115 is switched to inter. The image encoding apparatus 100 may generate a prediction block for an input block of an input image and then encode the difference between the input block and the prediction block.

In the case of intra mode, the intra prediction unit 120 may generate a prediction block by performing spatial prediction using pixel values of already encoded blocks surrounding the current block.

In the case of inter mode, the motion prediction unit 111 can obtain a motion vector by finding an area that best matches the input block in the reference image stored in the reference image buffer 190 during the motion prediction process. The motion compensation unit 112 may generate a prediction block by performing motion compensation using a motion vector and a reference image stored in the reference image buffer 190.

The subtractor 125 may generate a residual block by the difference between the input block and the generated prediction block. The transform unit 130 may perform transform on the remaining block and output a transform coefficient. Transformation methods may include fixed transformation methods, optimal transformation methods, and intra-prediction mode-specific transformation methods (MDDT: Mode Dependent Directional Transform). Details of each are described later.

The quantization unit 140 may quantize the input transform coefficient according to the quantization parameter and output a quantized coefficient.

The entropy encoding unit 150 may perform entropy encoding based on values calculated by the quantization unit 140 or encoding parameter values calculated during the encoding process, and output a bit stream.

When entropy coding is applied, a small number of bits are allocated to symbols with a high probability of occurrence and a large number of bits are allocated to symbols with a low probability of occurrence to represent symbols, so that the bits for the symbols to be encoded are expressed. The size of the column may be reduced. Therefore, the compression performance of video encoding can be improved through entropy coding.

For entropy coding, coding methods such as exponential gollomb, CAVLC (Context-Adaptive Variable Length Coding), and CABAC (Context-Adaptive Binary Arithmetic Coding) may be used. For example, the entropy encoding unit 150 may store a table for performing entropy encoding, such as a Variable Lenghth Coding/Code (VLC) table, and the entropy encoding unit 150 may store the stored variable length coding. Entropy encoding can be performed using a (VLC) table. In addition, as another example, the entropy encoding unit 150 binarizes symbols, converts them into bins, predicts the probability of occurrence of bins according to a context model, and performs arithmetic encoding of the bins. You can also use the CABAC entropy encoding method to generate a bitstream. Here, bin means the value (0 or 1) of each binary number when the symbol is expressed as a string of binary numbers through binarization.

The entropy encoding unit 150 may perform entropy encoding according to the entropy encoding model. The entropy encoding model may include, for example, a symbol's frequency of occurrence, probability of occurrence, or probability transition model. Additionally, in another embodiment, the entropy encoding model may include information about the binary encoding method. Here, the binary coding method may include variable length coding, arithmetic coding, etc.

The quantized coefficient may be inversely quantized in the inverse quantization unit 160 and inversely transformed in the inverse transformation unit 170. The inverse-quantized and inverse-transformed coefficients may be added to the prediction block through the adder 175 and a restored block may be generated.

The restored block passes through the filter unit 180, and the filter unit 180 applies at least one of a deblocking filter, Sample Adaptive Offset (SAO), and Adaptive Loop Filter (ALF) to the restored block or restored picture. can do. The restored block that has passed through the filter unit 180 may be stored in the reference image buffer 190.

Figure 2 is a block diagram showing the configuration of a video decoding device according to an embodiment. Referring to FIG. 2, the image decoding device 200 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, a motion compensation unit 250, and a filter unit. 260 and a reference image buffer 270.

The image decoding device 200 may receive a bitstream output from an encoder, perform decoding in intra mode or inter mode, and output a reconstructed image, that is, a reconstructed image. In the case of intra mode, the switch may be converted to intra, and in the case of inter mode, the switch may be converted to inter. The image decoding apparatus 200 may obtain a reconstructed residual block from an input bitstream, generate a prediction block, and then add the reconstructed residual block and the prediction block to generate a reconstructed block, that is, a reconstructed block.

The entropy decoding unit 210 may entropy decode the input bitstream according to a probability distribution to generate symbols including symbols in the form of quantized coefficients. The entropy decoding method is similar to the entropy encoding method described above.

When the entropy decoding method is applied, a small number of bits are allocated to symbols with a high probability of occurrence and a large number of bits are allocated to symbols with a low probability of occurrence to represent the symbols, thereby reducing the size of the bit string for each symbol. can be reduced. Therefore, the compression performance of video decoding can be improved through the entropy decoding method.

The entropy decoding unit 210, like the entropy encoding unit 150, may perform entropy decoding according to an entropy decoding model. The entropy decoding model may include, for example, a symbol's occurrence frequency, occurrence probability value, or probability transition model. Additionally, in another embodiment, the entropy decoding model may include information about the binary encoding method. Here, the binary coding method may include variable length coding, arithmetic coding, etc.

The quantized coefficients are inversely quantized in the inverse quantization unit 220 and inversely transformed in the inverse transformation unit 230. As a result of the inverse quantization/inverse transformation of the quantized coefficients, a restored residual block may be generated. When inverse transformation is performed, the fixed transformation method, optimal transformation method, intra-prediction mode dependent directional transformation (MDDT) method, etc. described above in the embodiment of FIG. 1 may be applied inversely. Details of each are described later.

In the case of intra mode, the intra prediction unit 240 may generate a prediction block by performing spatial prediction using pixel values of already encoded blocks surrounding the current block. In the inter mode, the motion compensation unit 250 may generate a prediction block by performing motion compensation using a motion vector and a reference image stored in the reference image buffer 270.

The restored residual block and the prediction block are added through an adder 255, and the added block passes through a filter unit 260. The filter unit 260 may apply at least one of a deblocking filter, SAO, and ALF to a restored block or a restored picture. The filter unit 260 outputs a reconstructed image, that is, a restored image. The restored image is stored in the reference image buffer 270 and can be used for inter-screen prediction.

As described above in the embodiments of Figures 1 and 2, transformation methods may include a fixed transformation method, an optimal transformation method, and a transformation method for each intra-prediction mode.

The fixed transformation method is a method of performing transformation or inverse transformation using a fixed transformation equation such as DCT (Discrete Cosine Transform). The fixed conversion method cannot reflect changes in the characteristics of the signal to be encoded, and has limitations in compression performance.

The optimal conversion method is a method of obtaining statistical values such as auto-correlation or covariance of the signal to be encoded and then performing conversion by determining a conversion equation according to the statistical values. An example of an optimal transformation method may include KLT (Karhunen-Loeve Transform). The optimal conversion method can perform signal adaptive conversion. However, in the optimal conversion method, since the decoder cannot know information about statistics or the transformation formula itself, there is a disadvantage that an additional amount of encoding bits is required to transmit the information about the statistics or the transformation formula itself to the decoder.

The intra-screen prediction mode-specific transformation (MDDT: Mode Dependent Directional Transform) method determines the optimal transformation equation in advance for a number of training data for each prediction mode within the screen, and then fixedly uses the pre-determined optimal transformation equation. This is how to perform the conversion. The optimal conversion equation may be a conversion equation using KLT, DCT, DST (Discrete Sine Transform), etc.

In the conversion method for each intra-prediction mode, signal characteristics according to the prediction mode can be reflected, and compression performance can be improved because additional encoding bits transmitted from the encoder to the decoder are not required. However, if the characteristics of the signal to be encoded are different from the training data, compression performance may be limited.

Therefore, in order to minimize the generation of additional encoding bits transmitted from the encoder to the decoder and improve image compression performance, the context information obtained from the encoding/decoding target block, the already encoded/decoded block, or the already restored pixel value is used to transform/decode A method for performing inverse transformation may be provided.

Figure 3 is a flowchart schematically showing a conversion method according to an embodiment of the present invention. Transformation according to the embodiment of FIG. 3 may be performed in an encoder, and in one embodiment, may be performed in a conversion unit of the encoder.

Referring to FIG. 3, the encoder acquires context information to be used to derive a transformation method for the encoding target block (S310). There may be various types of context information to be used in deriving the conversion method.

In one embodiment, the context information may include encoding parameters of a block to be encoded or encoding parameters of an already encoded block. Encoding parameters may include prediction mode information, motion or movement information, quantization information, block size information, or filter information.

Prediction mode information may include, for example, information indicating whether it is inter-screen prediction or intra-screen prediction, and/or information indicating in detail which prediction direction and/or prediction method is used. Motion or movement information includes, for example, information indicating whether a motion vector or movement (transformation) vector is encoded, vector size information, reference image index, and/or information about the number of reference images. It can be included. Quantization information may include information about quantization parameters, quantization step size, type of quantization weight matrix, and/or value of the quantization weight matrix, etc. Block size information may include information about the size of the block, whether the block is divided, the division type of the block, the location of the division boundary of the block, and/or the direction of the division boundary of the block, etc. Filter information includes information indicating whether to apply an interpolation filter, denoising filter, debanding filter, deblocking filter, etc., filter coefficient information, and/or filter strength. It may include information about.

The values of the above encoding parameters can be directly used as context information. Additionally, after the combination or frequency of occurrence of encoding parameter values is obtained, the combination or frequency of occurrence information may be used as context information.

In another embodiment, the context information may include transform coefficient information of an already encoded block.

For example, the transform coefficient information may include information indicating whether the transform coefficient exists. Information indicating the presence or absence of a conversion coefficient may include a Coded Block Pattern, Coded Block Flags, and Split Flags. Also, for example, the transformation coefficient information may also include information regarding the distribution of the transformation coefficient and/or the distribution type of the transformation coefficient. Information regarding the distribution of the transformation coefficient and/or the distribution type of the transformation coefficient includes information indicating that the transformation coefficient is mainly distributed in the low-frequency region, information indicating that the transformation coefficient is mainly present in a specific direction, and information indicating that the transformation coefficient is evenly distributed. There may be information that indicates . Transform coefficient information may also include information about the scan method and/or scan order for the transform coefficient.

In another embodiment, the context information may include pixel values that have already been restored. The area where the pixel values are obtained can be determined in various ways. For example, the pixel values may be pixel values of already restored pixels among pixels located spatially close to the encoding target block. As another example, the pixel values may be pixel values of pixels that have already been restored among the pixels used to predict the pixel value of the encoding target block.

The encoder may use all or part of the pixel values to derive feature values that can represent the pixel values, and then use the feature values as context information. The characteristic value may be, for example, a statistical value such as an average variance (deviation) value or correlation (correlation). Additionally, the feature value may be a value representing a feature of the image, such as gradient or edgeness, calculated from pixel values. Here, edginess may be a result obtained by applying a Sobel operator or a Laplacian operator.

The encoder is used to derive a transformation method of the encoding target block using at least one of the above-described encoding parameters of the encoding target block, encoding parameters of the already encoded block, transform coefficient information of the already encoded block, and already restored pixel value information. Contextual information can be obtained.

Referring again to FIG. 3, the encoder uses the acquired context information to derive a transformation method applied to the encoding target block (S320).

When determining a conversion method using context information, the encoder can use various methods or means. For example, a means of determining whether the context information itself is included in a specific range may be used. As another example, a means of determining whether the context information itself satisfies a specific condition may be used. As another example, a means of determining whether a difference between the same type of context information is included in a specific range may be used.

The encoder can use the determined result as is to derive a conversion method. Additionally, the encoder may apply the above determination means several times to measure the frequency of occurrence of context information included in a specific range and/or context information that satisfies a specific condition and obtain a frequency value. At this time, the encoder may use the frequency value to derive a conversion method.

When deriving a conversion method, the encoder may select one of a plurality of predetermined conversion methods.

At this time, the encoder can select one of a plurality of conversion methods with different methods themselves. Multiple transformation methods may include KLT, DCT, DST, Hadamard transform, wavelet transform, and directional transform.

The encoder may select one of a plurality of transformation equations within the same transformation method. For example, the plurality of transformation equations may be a plurality of transformation equations optimized for different training data in the KLT transformation method.

When deriving a conversion method, the encoder may update a predetermined initial conversion equation according to context information. Here, the predetermined initial conversion equation means, for example, the initial conversion equation for conversion methods such as KLT and DCT.

The initial conversion equation may be predetermined at various stages. For example, the initial conversion formula may be determined in advance at the sequence stage or at the GOP (Group of Picture) stage. GOP refers to a picture group, that is, a group of pictures. Additionally, as another example, the initial transformation equation may be predetermined at the picture, slice, coding unit, or transformation unit level.

The decision on whether to update the initial transformation equation and/or the actual update process for the initial transformation expression may be performed before transformation of the encoding target block. Additionally, the decision on whether to update the initial transformation equation and/or the actual update process for the initial transformation expression may be performed after transformation of the encoding target block.

When deriving a transformation method using the acquired context information, the encoder uses at least one of the methods of selecting one of the plurality of predetermined transformation methods described above and the method of updating the predetermined initial transformation equation according to the context information. You can.

Referring again to FIG. 3, the encoder performs transformation on the encoding target block of the input image using the derived transformation method (S330). Transformation can be performed in units of transformation blocks.

The encoder can derive the optimal transformation method and/or transformation equation on a block-by-block basis using context information about the encoding target block. Additionally, the encoder can perform conversion in a context-adaptive or signal-adaptive manner by performing conversion using the above conversion method and/or conversion equation. Therefore, the encoder can perform conversion as adaptively as possible according to changes in the characteristics of the signal to be encoded.

Additionally, in the embodiment of FIG. 3, the information used to derive the transformation method and/or transformation equation or to obtain context information is information that the decoder can obtain without any additional information transmitted from the encoder. Therefore, in the conversion method according to an embodiment of the present invention, the amount of additional bits transmitted from the encoder to the decoder may be minimized or not generated.

Transform coefficients generated as a result of transformation may be quantized according to a quantization parameter and provided to the entropy encoder. As described above in the embodiment of FIG. 1, the entropy encoding unit may perform entropy encoding according to the entropy encoding model. When the conversion method is derived in a context-adaptive manner, as in the embodiment of FIG. 3, the probability of occurrence or frequency of symbols, or the binary coding method, etc. may vary depending on context information during entropy encoding. Therefore, in one embodiment, when the transformation method is derived in a context-adaptive manner, the entropy encoding model may also be derived in a context-adaptive manner.

Figure 4 is a flowchart schematically showing an entropy encoding method according to an embodiment of the present invention. Entropy encoding according to the embodiment of FIG. 4 may be performed in an encoder, and in one embodiment, may be performed in an entropy encoding unit of the encoder.

Referring to FIG. 4, the encoder acquires context information used to derive an entropy encoding model (S410).

The context information may not be context information separately obtained in the entropy encoding process, and the context information obtained in the conversion process may be used as is in the entropy encoding process.

The encoder derives an entropy encoding model using the acquired context information (S420).

As described above in the embodiment of FIG. 3, when deriving a transformation method, the encoder may select one of a plurality of predetermined transformation methods. At this time, the encoder may transmit or signal an indicator indicating which of a plurality of predetermined conversion methods has been selected. The encoder can use the acquired context information to derive an entropy encoding model of the indicator. In one embodiment, the encoder may update the entropy encoding model for a plurality of predetermined transformation methods according to the intra-prediction mode value of the neighboring block(s) of the encoding target block. The method of deriving the entropy encoding model is not limited to the above embodiment, and may be determined differently depending on the type of acquired context information, conversion method derivation method, etc.

The derived entropy encoding model may include, for example, a symbol occurrence frequency, occurrence probability value, or probability transition model. Also, as another example, the entropy encoding model may include information about the binary encoding method. Here, the binary coding method may include variable length coding, arithmetic coding, etc.

Referring again to FIG. 4, the encoder performs entropy encoding using the derived entropy encoding model (S430).

In the embodiment of FIG. 4, the information used to derive the entropy encoding model is information that the decoder can obtain without any additional information transmitted from the encoder. Therefore, when the entropy encoding method according to an embodiment of the present invention is used, the amount of additional bits transmitted from the encoder to the decoder can be minimized. In addition, the encoder performs entropy coding using the acquired context information, so that coding can be performed as adaptively as possible according to changes in the characteristics of the signal to be encoded.

According to the embodiment of the present invention described above in FIGS. 3 and 4, conversion and/or entropy coding can be performed as adaptively as possible according to changes in the characteristics of the signal to be encoded, without generating additional encoded bits transmitted from the encoder to the decoder. there is. Therefore, the compression performance of video encoding is improved.

Similar to the above-described transformation coding method, the inverse transformation method can be adaptively derived in the decoder according to context information. When a context-adaptive inverse transformation method is used in the decoder, information about the probability of occurrence of symbols used during entropy decoding may vary depending on the context information. Therefore, even in entropy decoding, an entropy decoding model adaptively derived according to context information can be used.

Figure 5 is a flowchart schematically showing an entropy decoding method according to an embodiment of the present invention. Entropy decoding according to the embodiment of FIG. 5 may be performed in a decoder, and in one embodiment, may be performed in an entropy decoding unit of the decoder.

Referring to FIG. 5, the decoder acquires context information used to derive an entropy decoding model (S510). There may be various types of context information used to derive the entropy decoding model.

In one embodiment, the context information may include encoding parameters of a block to be decoded and/or encoding parameters of a block that has already been decoded. Encoding parameters may include prediction mode information, motion or movement information, quantization information, block size information, or filter information.

Prediction mode information may include, for example, information indicating whether it is inter-screen prediction or intra-screen prediction, and/or information indicating in detail which prediction direction and/or prediction method is used. Motion or movement information includes, for example, information indicating whether a motion vector or movement (transformation) vector is encoded, vector size information, reference image index, and/or information about the number of reference images. It can be included. Quantization information may include information about quantization parameters, quantization step size, type of quantization weight matrix, and/or value of the quantization weight matrix, etc. Block size information may include information about the size of the block, whether the block is divided, the division type of the block, the location of the division boundary of the block, and/or the direction of the division boundary of the block, etc. Filter information includes information indicating whether to apply an interpolation filter, denoising filter, debanding filter, deblocking filter, etc., filter coefficient information, and/or filter strength. It may include information about.

The values of the above encoding parameters can be directly used as context information. Additionally, after the combination or frequency of occurrence of encoding parameter values is obtained, the combination or frequency of occurrence information may be used as context information.

In another embodiment, the context information may include transform coefficient information of an already decoded block.

For example, the transform coefficient information may include information indicating whether the transform coefficient exists. Information indicating the presence or absence of a conversion coefficient may include a Coded Block Pattern, Coded Block Flags, and Split Flags. Also, for example, the transformation coefficient information may also include information regarding the distribution of the transformation coefficient and/or the distribution type of the transformation coefficient. Information regarding the distribution of the transformation coefficient and/or the distribution type of the transformation coefficient includes information indicating that the transformation coefficient is mainly distributed in the low-frequency region, information indicating that the transformation coefficient is mainly present in a specific direction, and information indicating that the transformation coefficient is evenly distributed. There may be information that indicates . Transform coefficient information may also include information about the scan method and/or scan order for the transform coefficient.

In another embodiment, the context information may include pixel values that have already been restored. The area where the pixel values are obtained can be determined in various ways. For example, the pixel values may be pixel values of already restored pixels among pixels located spatially close to the decoding target block. As another example, the pixel values may be pixel values of pixels that have already been restored among the pixels used to predict the pixel value of the decoding target block.

The decoder may use all or part of the pixel values to derive feature values that can represent the pixel values, and then use the feature values as context information. The characteristic value may be, for example, a statistical value such as an average variance or deviation value or correlation. Additionally, the feature value may be a value representing a feature of the image, such as gradient or edgeness, calculated from pixel values.

The decoder generates context information to be used for deriving an entropy decoding model using at least one of the above-described encoding parameters of the decoding target block, encoding parameters of an already decoded block, transform coefficient information of an already decoded block, and already restored pixel value information. It can be obtained.

The decoder derives an entropy decoding model using the acquired context information (S520).

As will be described later in the embodiment of FIG. 6, the decoder according to an embodiment of the present invention may select one of a plurality of predetermined inverse transformation methods when deriving an inverse transformation method. At this time, the decoder can parse the bitstream to obtain an indicator for selecting a transformation method, and then select one of a plurality of predetermined inverse transformation methods according to the indicator value. The decoder can use previously acquired context information to derive an entropy decoding model required for parsing the indicator. In one embodiment, the decoder may update the entropy decoding model for a plurality of predetermined inverse transformation methods according to the intra-prediction mode value of the block(s) surrounding the decoding target block. The method of deriving the entropy decoding model is not limited to the above embodiment, and may be determined differently depending on the type of acquired context information, the method of deriving the inverse transformation method, etc.

The derived entropy decoding model may include, for example, a symbol occurrence frequency, an occurrence probability value, or a probability transition model. Also, as another example, the entropy decoding model may include information about the binary encoding method. Here, the binary coding method may include variable length coding, arithmetic coding, etc.

Referring again to FIG. 5, the decoder performs entropy decoding using the derived entropy decoding model (S530).

In the embodiment of FIG. 5, the information used to derive the entropy decoding model is information that the decoder can obtain without any additional information transmitted from the encoder. Therefore, when the entropy decoding method according to an embodiment of the present invention is used, the amount of additional bits transmitted from the encoder to the decoder can be minimized. In addition, the decoder performs entropy decoding using the acquired context information, so that decoding can be performed as adaptively as possible according to changes in the characteristics of the signal to be decoded.

Values generated as a result of entropy decoding may be provided to the quantization unit of the decoder, and the quantization unit may perform inverse quantization according to the quantization parameter. Transform coefficients generated as a result of inverse quantization may be provided to the inverse transform unit of the decoder.

Figure 6 is a flowchart schematically showing an inverse transformation method according to an embodiment of the present invention. Inverse transformation according to the embodiment of FIG. 6 may be performed in a decoder, and in one embodiment, may be performed in an inverse transformation unit of the decoder.

Referring to FIG. 6, the decoder acquires context information to be used in deriving the inverse transformation method of the decoding target block (S610).

The context information may not be context information separately obtained during the inverse transformation process, and the context information obtained during the entropy decoding process may be used as is in the inverse transformation process. If context information is not used in the entropy decoding process, the decoder may separately obtain context information in the inverse transformation process. At this time, there may be various types of context information to be used in deriving the inverse transformation method. For example, the context information may be the types of information described above in the embodiment of FIG. 5.

Referring again to FIG. 6, the decoder uses the acquired context information to derive an inverse transformation method applied to the decoding target block (S620).

When determining the inverse transformation method using context information, the decoder can use various methods or means. For example, a means of determining whether the context information itself is included in a specific range may be used. As another example, a means of determining whether the context information itself satisfies a specific condition may be used. As another example, a means of determining whether a difference between the same type of context information is included in a specific range may be used.

The decoder can use the determined result as is to derive an inverse transformation method. In addition, the decoder may apply the above-mentioned determination means several times to measure the frequency of occurrence of context information included in a specific range and/or context information satisfying a specific condition to obtain a frequency value. At this time, the decoder may use the frequency value to derive an inverse transformation method.

When deriving an inverse transformation method, the decoder may select one of a plurality of predetermined inverse transformation methods.

At this time, the decoder can select one of a plurality of inverse transformation methods with different methods themselves. A plurality of inverse transformation methods may include KLT, DCT, DST, Hadamard transform, wavelet transform, and directional transform.

The decoder may select one of a plurality of inverse transformation equations within the same inverse transformation method. For example, the plurality of inverse transformation equations may be a plurality of inverse transformation equations optimized for different training data in the KLT transformation method.

When deriving an inverse transformation method, the decoder may update a predetermined initial inverse transformation equation according to context information. Here, the predetermined initial inversion equation means, for example, the initial inversion equation for an inversion method such as KLT, DCT, etc.

The initial inverse transformation equation may be determined in advance at various stages. For example, the initial inverse transformation equation may be predetermined at the sequence stage or at the GOP (Group of Picture) stage. Additionally, as another example, the initial inverse transformation equation may be predetermined at the picture, slice, coding unit, or transformation unit stage.

The decision on whether to update the initial inverse transformation equation and/or the actual update process for the initial inverse transformation equation may be performed before the inverse transformation of the decoding target block. Additionally, the decision on whether to update the initial inverse transformation equation and/or the actual update process for the initial inverse transformation equation may be performed after the inverse transformation of the decoding target block.

In deriving an inverse transformation method using the acquired context information, the decoder uses at least one of the methods of selecting one of the plurality of predetermined inverse transformation methods described above and the method of updating the predetermined initial inverse transformation equation according to the context information. Available.

Referring again to FIG. 6, the decoder performs inverse transformation on the decoding target block using the derived inverse transformation method (S630). Inverse transformation can be performed on transform coefficients in units of transform blocks.

In the embodiment of FIG. 6, the decoder may derive the optimal inverse transformation method and/or inverse transformation equation on a block-by-block basis using context information about the decoding target block. Additionally, the decoder may perform inverse transformation in a context-adaptive or signal-adaptive manner by performing inverse transformation using the inverse transformation method and/or the inverse transformation equation. Therefore, the decoder can perform inverse transformation as adaptively as possible according to changes in the characteristics of the signal to be decoded.

Additionally, in the embodiment of FIG. 6, the information used to derive the inverse transformation method and/or the inverse transformation equation or to obtain context information is information that the decoder can obtain without any additional information transmitted from the encoder. Therefore, in the inverse transformation method according to an embodiment of the present invention, the amount of additional bits transmitted from the encoder to the decoder may be minimized or may not be generated.

According to the embodiment of the present invention described above in FIGS. 5 and 6, inverse transformation and/or entropy decoding can be performed as adaptively as possible according to changes in the characteristics of the signal to be decoded, without generating additional encoded bits transmitted from the encoder to the decoder. there is. Therefore, the compression performance of video encoding and decoding is improved.

In the above-described embodiments, the methods are described based on flowcharts as a series of steps or blocks, but the present invention is not limited to the order of steps, and some steps may occur in a different order or simultaneously with other steps as described above. there is. Additionally, those of ordinary skill in the art will recognize that the steps shown in the flowchart are not exclusive and that other steps may be included or one or more steps in the flowchart may be deleted without affecting the scope of the present invention. You will understand.

The above-described embodiments include examples of various aspects. Although it is not possible to describe all possible combinations for representing the various aspects, those skilled in the art will recognize that other combinations are possible. Accordingly, the present invention is intended to include all other substitutions, modifications and changes falling within the scope of the following claims.

Claims (20)

Receiving a bitstream about video information;
parsing the bitstream to derive quantization parameter information for determining a quantization parameter for a current block;
determining a transformation method for the current block among a plurality of transformation methods using the quantization parameter information;
deriving a residual signal by performing transformation on the current block according to a transformation method for the current block; and
Generating a restored block for the current block using the residual signal,
An image decoding method in which the transformation method for the current block is determined using both the quantization parameter information and size information regarding the size of the current block. According to paragraph 1,
An image decoding method wherein the plurality of transformation methods include Hadamard transform, discrete cosine transform, and discrete sine transform. According to paragraph 2,
A video decoding method in which the transformation method for the current block is determined by further using prediction mode information indicating whether the current block is an inter-screen prediction block or an intra-screen prediction block. Receiving a bitstream about video information;
parsing the bitstream to derive quantization parameter information for determining a quantization parameter for a current block;
determining a transformation method for the current block among a plurality of transformation methods using the quantization parameter information;
deriving a residual signal by performing transformation on the current block according to a transformation method for the current block; and
Generating a restored block for the current block using the residual signal,
The step of determining a transformation method for the current block includes determining whether to apply transformation for each intra-prediction mode using the quantization parameter information. According to paragraph 4,
If it is decided to apply transformation for each prediction mode within the screen,
A video decoding method in which the transformation method for the current block is determined further using an intra-screen prediction mode. According to clause 5,
A video decoding method in which, when it is determined not to apply transformation for each intra-screen prediction mode, the transformation method for the current block is determined regardless of the intra-screen prediction mode. An entropy decoding unit that receives a bitstream related to image information and parses the bitstream to derive quantization parameters for the current block;
Determining a transformation method for the current block among a plurality of transformation methods based on the quantization parameter for the current block, and performing transformation on the current block according to the transformation method for the current block to derive a residual signal Inverse conversion unit; and
An adder that generates a restored block for the current block using the residual signal,
The inverse transform unit determines a transformation method for the current block using quantization parameter information that determines a quantization parameter for the current block,
The inverse transform unit determines a transformation method for the current block using both the quantization parameter information and size information regarding the size of the current block. In clause 7,
The plurality of transformation methods include Hadamard transform, discrete cosine transform, and discrete sine transform. determining a transformation method for the current block among a plurality of transformation methods;
deriving a residual signal by performing transformation on the current block according to a transformation method for the current block; and
Generating a restored block for the current block using the residual signal,
Quantization parameter information for the current block is generated,
The quantization parameter information indicates a transformation method for the current block among the plurality of transformation methods,
The transformation method for the current block is an image encoding method that corresponds to the quantization parameter information and size information regarding the size of the current block. According to clause 9,
An image encoding method wherein the plurality of transformation methods include Hadamard transform, discrete cosine transform, and discrete sine transform. According to clause 10,
The transformation method for the current block is an image encoding method corresponding to prediction mode information indicating whether the current block is an inter-screen prediction block or an intra-screen prediction block. determining a transformation method for the current block among a plurality of transformation methods;
deriving a residual signal by performing transformation on the current block according to a transformation method for the current block; and
Generating a restored block for the current block using the residual signal,
Quantization parameter information for the current block is generated,
The quantization parameter information indicates a transformation method for the current block among the plurality of transformation methods,
The step of determining a transformation method for the current block includes determining whether to apply transformation for each intra-screen prediction mode,
Whether to apply transformation for each intra-screen prediction mode is determined by an image encoding method corresponding to the quantization parameter information. According to clause 12,
If it is decided to apply transformation for each prediction mode within the screen,
The transformation method for the current block is an image encoding method corresponding to an intra-screen prediction mode. According to clause 13,
An image encoding method in which, when it is determined not to apply transformation for each intra-prediction mode, the transformation method for the current block is determined regardless of the intra-prediction mode. In the computer-readable recording medium storing a bitstream for video decoding, the bitstream includes:
Quantization parameter information that determines the quantization parameters for the current block
Including,
A transformation method for the current block is determined among a plurality of transformation methods using the quantization parameter information,
A residual signal is derived by performing transformation on the current block according to the transformation method for the current block,
A restored block for the current block is generated using the residual signal,
A computer-readable recording medium in which the conversion method for the current block is determined using both the quantization parameter information and size information regarding the size of the current block. According to clause 15,
A computer-readable recording medium, wherein the plurality of transformation methods include Hadamard transform, discrete cosine transform, and discrete sine transform. According to clause 16,
The conversion method for the current block is determined by further using prediction mode information indicating whether the current block is an inter-prediction block or an intra-prediction block. In the computer-readable recording medium storing a bitstream for video decoding, the bitstream includes:
Quantization parameter information that determines the quantization parameters for the current block
Including,
A transformation method for the current block is determined among a plurality of transformation methods using the quantization parameter information,
A residual signal is derived by performing transformation on the current block according to the transformation method for the current block,
A restored block for the current block is generated using the residual signal,
A computer-readable recording medium in which, in determining a transformation method for the current block, whether to apply transformation for each intra-prediction mode is determined using the quantization parameter information. According to clause 18,
If it is decided to apply transformation for each prediction mode within the screen,
A computer-readable recording medium wherein the conversion method for the current block is determined further using an intra-prediction mode. According to clause 19,
A computer-readable recording medium in which, when it is determined not to apply transformation for each intra-prediction mode, a transformation method for the current block is determined regardless of the intra-prediction mode.

Images