KR101997655B1 - Method and Apparatus for Image Encoding/Decoding - Google Patents

Abstract

An embodiment of the present invention relates to a method and apparatus for image encoding / decoding using a high precision filter.
According to an embodiment of the present invention, a prediction block of a current block is generated and a residual block generated by subtracting the prediction block from the current block is generated to generate a frequency transform block using a different transform method according to a transform unit. After generating the frequency transform block, the video encoding step of quantizing and encoding the bitstream and extracting information about the quantized frequency transform block and the transform unit from the bitstream and inverse quantization of the quantized frequency transform block to restore the frequency transform block And restoring the residual block by inversely transforming the frequency transform block, and restoring the frequency transform block using another inverse transform method according to the extracted transform unit, and generating a prediction block from a reference block referred to by a motion vector. Adds the residual block to be recovered and the prediction block An image encoding / decoding method including an image decoding step of reconstructing a block is provided.

Description

Method and apparatus for image encoding / decoding {Method and Apparatus for Image Encoding / Decoding}

An embodiment of the present invention relates to a method and apparatus for image encoding / decoding. More specifically, the present invention relates to an image encoding / decoding method and apparatus for improving a compression ratio by transforming a residual block or inversely transforming a frequency coefficient block using a different transform method or another inverse transform method according to a transform unit.

The content described in this section merely provides background information on the embodiments of the present invention and does not necessarily constitute prior art.

Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) have formed a joint team of MPEG and VCEG, called Joint Collaborative Team on Video Coding (JCT-VC), over existing H.264 / AVC (Advanced video Coding). Standardization is underway with a new standard called Excellent and Excellent High Efficiency Video Coding (HEVC).

 HEVC uses coding tree block (CTB) as a video coding unit, where CTB is defined by various square shapes. CTB is often called Coding Unit.

1 shows the size and shape of a conventional coding unit (CU).

Referring to FIG. 1, the coding unit (CU) has a quad tree shape, and when the maximum coding unit (Largest Coding Unit) having a size of 64 × 64, the depth is set to 0 and the depth becomes 3. The coding is performed by finding the optimal prediction unit recursively until a coding unit (CU) of 8 × 8 size.

The basic unit of prediction is defined as a prediction unit, and one coding unit (CU) is divided into a plurality of blocks and used for prediction.

2 shows a type of a typical prediction unit (PU) and a prediction order of the prediction unit (PU) in one coding unit (CU).

Referring to FIG. 2, after performing a skip mode in a coding unit (CU) having a size of 2N × 2N, an inter 2N × 2N mode, an inter 2N × N mode, an inter N × 2N mode, and an inter N Prediction is performed in the order of xN mode, intra 2N × 2N mode, and intra N × N mode. However, except for the coding unit (CU) having a size of 8 × 8 (2N × 2N), all the coding units (CU) are applied to all the remaining prediction units (PU) except the inter N × N mode and intra N × N mode. Perform predictions for

Motion estimation used for inter-picture prediction in HEVC is performed in a prediction unit. Motion estimation is performed to reduce bitrate by removing temporal redundancy in video encoding. In particular, HEVC uses a large amount of motion estimation by using an inductive process of a coding unit (CU) and a prediction unit (PU) of various sizes, and thus has high coding efficiency.

Prediction of a motion vector is performed by referring to a past image on the basis of a time axis or by referring to both a past image and a future image. An image referred to to encode or decode a current frame is called a reference image. Since HEVC supports multiple reference frames, a coding efficiency may be greater than that of using only the previous frame as a reference picture by selecting a block of a frame having the most overlap with the current block as the reference picture. In addition, a skip mode included in the coding unit (CU), inter 2N × 2N mode, inter 2N × N mode, inter N × 2N mode, inter N × N mode, intra 2N × 2N mode, intra N × N Rate-distortion optimization is further used to select the most optimal mode among all possible coding modes, including modes, further improving the coding efficiency of HEVC.

According to the HEVC standard, which is designed to encode / decode video data, a transform unit of various sizes is used to reduce spatial correlation of residual signals in a block and increase compression ratio of energy after performing intra / intra prediction. A quantization method is used to increase the compression ratio with respect to the frequency coefficient, and a constant conversion method irrespective of the size of the transform becomes a factor of reducing the compression ratio of the image.

In order to solve this problem, an embodiment of the present invention has a main object of improving a compression ratio by transforming a residual block or inversely transforming a frequency coefficient block using another transform method or another inverse transform method according to a transform unit.

In order to achieve the above object, an embodiment of the present invention provides a device for encoding / decoding an image, generating a prediction block of a current block and converting a residual block generated by subtracting the prediction block from the current block. An image encoder for generating a frequency transform block and generating the frequency transform block by using a transform method according to a transform unit, and then quantizing and encoding the frequency transform block into a bitstream; And extracting information about the quantized frequency transform block and the transform unit from the bitstream, inversely quantizing the quantized frequency transform block, to restore the frequency transform block, and to inversely transform the frequency transform block to restore the residual block. An image reconstructing the frequency transform block using a different inverse transform method according to a transform unit, generating a prediction block from a reference block referred to by a motion vector, and adding the reconstructed residual block and the prediction block to reconstruct the current block. It provides a video encoding / decoding apparatus comprising a decoder.

In addition, an embodiment of the present invention to achieve another object of the present invention, an apparatus for encoding an image, comprising: a prediction unit for generating a prediction block of the current block; A subtraction unit for generating a residual block by subtracting the prediction block from the current block; A transformer for generating the frequency transform block by transforming the residual block and generating a frequency transform block using a different transform method according to a transform unit; A quantizer configured to quantize the frequency transform block to generate a quantized frequency transform block; And a bitstream generator which encodes the quantized frequency transform block into a bitstream.

The transform unit may convert the residual block into a frequency transform block by using a Discrete Sine Transform (DST) technique when converting the smallest transform unit among sizes available as a transform unit (TU). .

The transformation unit having the minimum size may be 4 × 4 size.

The transform unit may convert the residual block into a frequency transform block using a Discrete Cosine Transform (DCT) technique for transform units other than the minimum size transform unit.

The transform unit may use a different transform method according to a transform unit only when the prediction mode is inter prediction.

In addition, to achieve another object of the present invention, an embodiment of the present invention, an apparatus for decoding an image, comprising: a bitstream decoding unit for extracting information about a quantized frequency transform block and transform unit from the bitstream; An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block; An inverse transform unit for restoring the residual block by inversely transforming the frequency transform block, and restoring the frequency transform block using another inverse transform method according to the extracted transform unit; A prediction unit generating a prediction block from a reference block referenced by the motion vector; And an adder configured to reconstruct the current block by adding the reconstructed residual block and the prediction block.

The inverse transform unit may restore the frequency transform block using an inverse discrete sine transform (inverse DST) technique when the information identified from the extracted transform unit is a minimum transform unit among sizes available as the transform unit size. Can be.

The transformation unit having the minimum size may be 4 × 4 size.

The inverse transform unit may restore the frequency transform block to a transform unit other than the minimum size transform unit by using an inverse discrete cosine transform (inverse DCT) technique.

In addition, to achieve another object of the present invention, an embodiment of the present invention, in a method for encoding / decoding an image, generates a prediction block of a current block and subtracts the prediction block from the current block Generating a frequency transform block by transforming the block, and generating the frequency transform block using a different transform method according to a transform unit, and then quantizing and encoding the block into a bitstream; And extracting information about the quantized frequency transform block and the transform unit from the bitstream, inversely quantizing the quantized frequency transform block, to restore the frequency transform block, and to inversely transform the frequency transform block to restore the residual block. An image for reconstructing the frequency transform block by using an inverse transform method according to a transform unit, generating a prediction block from a reference block referred to by a motion vector, and adding the reconstructed residual block and the prediction block to reconstruct the current block. It provides a video encoding / decoding method comprising a decoding step.

In addition, to achieve another object of the present invention, an embodiment of the present invention, a method for encoding an image, comprising: a prediction step of generating a prediction block of the current block; Subtracting the prediction block from the current block to generate a residual block; Transforming the residual block to generate a frequency transform block, and generating the frequency transform block using a different transform method according to a transform unit; A quantization step of quantizing the frequency transform block to generate a quantized frequency transform block; And a bitstream generation step of encoding the quantized frequency transform block into a bitstream.

In addition, to achieve another object of the present invention, an embodiment of the present invention, a method for decoding an image, comprising: a bitstream decoding step of extracting information about a quantized frequency transform block and transform unit from the bitstream; An inverse quantization step of restoring a frequency transform block by inverse quantizing the quantized frequency transform block; An inverse transform step of restoring the residual block by inversely transforming the frequency transform block, and restoring the frequency transform block using another inverse transform method according to the extracted transform unit; A prediction step of generating a prediction block from a reference block referred to by the motion vector; And an addition step of reconstructing the current block by adding the reconstructed residual block and the prediction block.

As described above, according to the embodiment of the present invention, the compression ratio is improved by transforming the residual block or inversely transforming the frequency coefficient block by using another transform method or another inverse transform method according to the transform unit.

In particular, if the transform unit is the smallest transform unit such as 4x4 size, DST is used and DCT is used for the higher transform units to maximize the compression rate during encoding and to decode the bitstream by inverse transform in a similar manner when decoding. can do.

1 is a diagram illustrating a size and a quadtree shape of a coding unit (CU).
FIG. 2 is a diagram illustrating the types and procedures of prediction units (PUs) encoded in one coding unit (CU).
3 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.
4 illustrates a concept of an integer sine inverse transform according to an embodiment of the present invention.
5 is a block diagram schematically illustrating an image decoding apparatus according to an embodiment of the present invention.
6 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.
7 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.

A video encoding apparatus (Video Encoding Apparatus), a video decoding apparatus (Video Decoding Apparatus) to be described below is a personal computer (PC), notebook computer, personal digital assistant (PDA), portable multimedia player (PMP) It may be a user terminal such as a portable multimedia player (PSP), a PlayStation Portable (PSP), a wireless communication terminal, a smart phone, a TV, or a server terminal such as an application server or a service server. A communication device such as a communication modem for communicating with various devices or a wired / wireless communication network, a memory for storing various programs and data for encoding or decoding an image or inter or intra prediction for encoding or decoding, and executing a program And a microprocessor for controlling and the like. It can mean a variety of devices.

In addition, the image encoded in the bitstream by the video encoding apparatus is real-time or non-real-time through the wired or wireless communication network, such as the Internet, local area wireless communication network, wireless LAN network, WiBro network, mobile communication network, or the like, or a cable, universal serial bus (USB: Universal) It may be transmitted to an image decoding apparatus through various communication interfaces such as a serial bus, and may be decoded by the image decoding apparatus to restore and reproduce the image.

In general, a video may be composed of a series of pictures, and each picture may be divided into a predetermined area such as a frame or a block. When a region of an image is divided into blocks, the divided blocks may be classified into intra blocks and inter blocks according to an encoding method. An intra block refers to a block that is encoded by using an intra prediction coding method. An intra prediction coding is performed by using pixels of blocks previously encoded, decoded, and reconstructed in a current picture that performs current encoding. A prediction block is generated by predicting pixels of a block and a difference value with pixels of the current block is encoded. An inter block refers to a block that is encoded using inter prediction coding. Inter prediction coding generates a prediction block by predicting a current block within a current picture by referring to one or more past or future pictures, and then generates a current block. This is a method of encoding the difference value with. Here, a frame referred to for encoding or decoding the current picture is referred to as a reference frame.

3 is a block diagram schematically illustrating a video encoding apparatus according to an embodiment of the present invention.

The image encoding apparatus 300 according to an embodiment of the present invention generates a residual block between the current block and the prediction block by generating a prediction block from the current block of the image through the prediction unit 310, and then transforms and quantizes the image. As shown in FIG. 3, a predictor 310, a subtractor 320, a transformer 330, a quantizer 340, a bitstream generator 350, and an inverse quantizer are shown in FIG. 3. 360, an inverse transform unit 370, an adder 380, and a frame memory 390.

The input image to be encoded is input in a coding unit. In the present invention, the coding unit has a N × N block form and N has a size of 2 ⁿ . The coding unit is configured in the form of a quad tree and performs coding from the largest coding unit (Largest Coding Unit) to a specified depth.

Prediction is performed for each prediction unit within one coding unit (CU), where the prediction unit (PU) is in the form of an NxM block and N and M are 2 ⁿ , 2 ^m (n, m, respectively). Has a natural number). The type of prediction in one coding unit (CU) may have four inter prediction unit (PU) types and two intra prediction unit (PU) types, as shown in FIG. 2, and all prediction units (PUs). Coding is performed to transmit information on the type of prediction unit (PU) having the best compression efficiency to the image decoding apparatus. In this case, the criterion for the compression efficiency may be a rate distortion cost including a number of bits required for image information transmission and a difference value between an original block and a prediction block.

The prediction unit 310 generates a prediction block using another frame to predict the prediction unit (PU) of the current coding unit (CU) or has already coded pixel values (eg, left, bottom left, top, top right and pixels). Value) to generate a predictive block. That is, in the intra prediction mode, the prediction unit 310 determines the prediction mode by using coding unit (CU) information of the left, lower left, upper, and upper right that have already been encoded and reconstructed, and then uses the prediction mode. While generating the prediction block, in the inter prediction mode, the prediction unit 310 generates a motion vector through motion estimation in a previous frame that has already been encoded and reconstructed, and generates a prediction block in the motion compensation process using the motion vector. can do.

The subtractor 320 generates a residual signal by calculating a difference value between the original pixel value of each pixel of the current block and the predicted value generated by the predictor 310.

The converter 330 converts the residual signal generated by the subtractor 320 into the frequency domain. The transform unit 330 is transformed for each transform unit. The transform unit TU is in the form of an N × N block, and N has a size of 2 ⁿ . After the prediction is performed on the current prediction unit (PU) in the transform unit 330, the transformation is performed for each transformation unit (TU), and the size of the transformation unit (TU) is equal to or larger than that of the corresponding coding unit (CU). Small and independent of the size of the prediction unit (PU). The transformation unit (TU) may perform transformation by dividing the transformation unit into a quadtree form, such as a coding unit (CU), and sequentially perform transformations from the size of the coding unit (CU) to an arbitrarily designated depth. At this time, a split transform flag for each depth of the transform unit (TU) is transmitted to the image decoding apparatus as information on the size of the transform unit (TU) having the lowest rate-distortion value (RD cost).

Here, the transform unit 330 may convert the residual signal into the frequency domain by using a discrete cosine transform (DCT: Discrete Cosine Transform, hereinafter referred to as a 'DCT transform') that transforms an image signal on a time axis into a frequency axis. .

In particular, according to an embodiment of the present invention, when the prediction unit 310 is predicted by inter prediction and is the smallest size (for example, 4 × 4 size) among the sizes available as the size of the transform unit (TU), it is derived through the prediction. When the correlation coefficient is derived through the following Equation 1 or Equation 2, the residual signal has a value between -0.5 and 0.5 in probability. Therefore, in this case, the coding performance can be improved by using Discrete Sine Transform (DST), which shows a higher compression performance compared to the DCT transform. If the transform unit is larger than the minimum of the available sizes as the size of the transform unit (TU), the DCT transform may be used.

In Equations 1 and 2, B ^c represents a block having a size of 4x4 and B ^s represents a block in which the current block is moved by one pixel in the horizontal direction (right or left). B ^c _{x, y} represents the position (x, y) in the current 4x4 block, and B ^LS _{x, y} and B ^RS _{x, y} are in the 4x4 block shifted one pixel position left and right in the current 4x4 block, respectively. position (x, y) and B ' ^C , B' ^LS and B ' ^RS represent the average value of the pixels in the 4x4 block corresponding to B ^c _{x, y} , B ^LS _{x, y} , B ^RS _{x, y} respectively. Indicates. If the current 4x4 transform unit (TU) is the rightmost transform unit (TU) in the coding unit (CU), Equation 1 is used, otherwise Equation 2 is used to calculate the correlation coefficient.

The quantization unit 340 quantizes the frequency conversion block formed of the residual signal transformed into the frequency domain by the conversion unit 330. In this case, the coded block flag cbf is transmitted, which means that the coefficient is 0 for the luminance unit and the TU of the chrominance signal, in which the quantized transform coefficients are all zero. As a quantization method, various quantization techniques such as dead zone uniform threshold quantization (DZUTQ) or quantization weighted matrix (Quantization Weighted Matrix) may be used.

The bitstream generator 350 encodes the quantized frequency transform block including the frequency coefficient quantized by the quantizer 340 into a bitstream. In particular, in the embodiment of the present invention, when the prediction unit is predicted by inter prediction and the transformation unit (TU) has a minimum size, for example, a size of 4x4, among the available sizes as the size of the transformation unit (TU), the DST transformation is used. DCT transform is used for 8x8 TU, 16x16 TU, 32x32 TU, etc., so that no information indicating the type of transform is transmitted and the compression rate can be improved. As the encoding technique, an entropy encoding technique may be used, but various encoding techniques may be used without being limited thereto.

In addition, the bitstream generator 350 may include not only a bit string obtained by encoding the quantized frequency coefficients but also various pieces of information necessary to decode the encoded bit sequence in the encoded data. That is, the coded data includes a field including a coded block pattern (CBP), a delta quantization parameter, and a bit string in which the quantization frequency coefficients are encoded, and information necessary for prediction (eg, intra prediction). In the case of Intra prediction mode or a motion vector in the case of inter prediction, etc.) may include a field that contains a bit.

The inverse quantizer 360 inverse quantizes the transformed and quantized residual block (that is, the quantized frequency transform block) to restore the frequency coefficient block, and the inverse transformer 370 inverts the restored frequency coefficient block. Reconstruction the residual block by Inverse Transform. Here, inverse quantization and inverse transformation may be performed by inversely performing a transformation process performed by the transform unit 330 and a quantization process performed by the quantization unit 340, respectively. In particular, in the embodiment of the present invention, when the prediction unit is predicted by inter prediction and the transformation unit (TU) has a minimum size, for example, a size of 4x4, that is available as the size of the transformation unit (TU), the inverse DST transform is used. Inverse DCT conversion can be used for larger 8x8 TU, 16x16 TU, 32x32 TU, and the like. That is, the inverse quantization unit 360 and the inverse transform unit 370 are information about transform and quantization generated and transmitted from the transform unit 330 and the quantization unit 340 (for example, information on the transform and quantization type). Inverse quantization and inverse transformation may be performed using.

The adder 380 generates the reconstructed block by adding the prediction block generated by the predictor 310 and the residual block generated by the inverse transform unit 370.

The frame memory 390 is used as a reference block to store a block reconstructed by the adder 380 to generate a prediction block when performing intra or inter prediction.

Meanwhile, the discrete sine transform in the transform unit 330 may be calculated through Equation 3.

X in Equation 3 represents a target image to be discrete sine transformed, and Y represents a second image coefficient that is discrete sine transformed. N also represents the magnitude of the discrete sine transform.

4 is a diagram illustrating the concept of discrete sine transform according to an embodiment of the present invention.

If the target image is 4x4 in size, the discrete sine transform in Equation 1 may be simply performed according to the embodiment shown in FIG.

5 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.

As shown in FIG. 5, the image decoding apparatus 500 according to an exemplary embodiment of the present invention may include a bitstream decoder 510, an inverse quantizer 520, an inverse transform unit 530, an adder 540, and the like. The prediction unit 550 is configured.

The bitstream decoder 510 extracts the quantized frequency transform block by decoding the bitstream.

The bitstream decoder 510 may decode or extract not only the quantized frequency transform block but also information necessary for decoding by decoding the encoded data. Information necessary for decoding refers to information necessary for decoding the coded bit string in the encoded data (ie, the bitstream). For example, information about a block type, information about a motion vector, information about a transform and quantization type, and the like. It can be, and a variety of other information can be.

That is, the bitstream decoder 510 extracts a quantized frequency transform block including pixel information of the current block of the image by decoding the bitstream, which is data encoded by the image encoding apparatus 300, and then extracts the extracted quantized frequency transform block. Information necessary for the prediction is transmitted to the prediction unit 550.

The predictor 550 may predict the current block in the same manner as the predictor 310 of the image encoding apparatus 300 by using the information necessary for the prediction transmitted from the bitstream decoder 510.

The prediction unit 550 generates a prediction block using a reference block referenced by the motion vector of the current block.

The inverse quantizer 520 inverse quantizes the quantized frequency transform block extracted from the bitstream by the bitstream decoder 510.

The inverse transformer 530 of the image decoding apparatus 500 according to an embodiment of the present invention predicts the current block by inter prediction in the same manner as the inverse transformer 370 of the image encoding apparatus 300 described above with reference to FIG. 3. If the transform unit (TU) is 4x4, the inverse quantized frequency transform block is inversely transformed into the pixel region using the inverse DST transform, and the inverse quantized frequency transform block is transformed using the inverse DCT transform for the other frequency transform blocks. Inverse conversion can be made to the pixel region.

The adder 540 reconstructs the original pixel value of the current block by adding the residual signal inversely transformed and reconstructed by the inverse transformer 530 and the predicted pixel value predicted by the predictor 550.

The current block reconstructed by the adder 540 or the predictor 550 may be transferred to the frame memory 560, and used by the predictor 550 to predict another block.

The frame memory 560 stores the reconstructed image to enable generation of the prediction block in the image and the prediction block between the images.

An image encoding / decoding apparatus according to an embodiment of the present invention may be configured by connecting a bitstream output terminal of the image encoding apparatus 300 of FIG. 3 to a bitstream input terminal of the image decoding apparatus 500 of FIG. 4.

An image encoding / decoding apparatus according to an embodiment of the present invention generates a frequency transform block by generating a prediction block of a current block and subtracting the residual block generated by subtracting the prediction block from the current block, but converting the transform block according to a conversion unit. An image encoder for generating a frequency transform block using the method and quantizing the encoding block to encode the bitstream into a bitstream; Extract information about the quantized frequency transform block and the transform unit from the bitstream, inverse quantize the quantized frequency transform block, restore the frequency transform block, and inverse transform the frequency transform block to restore the residual block, Accordingly, an image decoder for reconstructing the frequency transform block using another inverse transform method, generating a prediction block from the reference block referenced by the motion vector, and adding the reconstructed residual block and the prediction block to reconstruct the current block.

The image encoder may be implemented by the image encoding apparatus 300 according to an embodiment of the present invention, and the image decoder may be implemented by the image decoding apparatus 500 according to an embodiment of the present invention.

6 is a flowchart illustrating an image encoding method according to an embodiment of the present invention.

That is, the image encoding apparatus 300 generates a residual signal by calculating a difference value between the original pixel value and the predicted pixel value of the current block by generating a predicted value using the motion vector value of the current block of the image (S610). A subtraction step (S620), when the residual signal generated through inter prediction is a 4x4 transform unit (TU), and the transform step of transforming to the frequency domain by using a DST transform and a DCT transform when the transform unit has a different size ( In operation S630, the quantization step of quantizing the residual signal transformed into the frequency domain (S640) and the bitstream generation step (S650) of determining an optimal quantized frequency transform residual signal and encoding the bitstream are encoded.

Here, the prediction step (S610) is the function of the prediction unit 310, the subtraction step (S620) is the function of the subtraction unit 320, the transform step (S630) is the function of the transformer 330, the quantization step ( S640 corresponds to the function of the quantization unit 340, and the bitstream generation step S650 corresponds to the function of the bitstream generation unit 350, and thus detailed description thereof will be omitted.

7 is a flowchart illustrating an image decoding method according to an embodiment of the present invention.

The video decoding apparatus 500 that receives and stores a bitstream of an image through a wired / wireless communication network or a cable may move a current block of the image to a motion vector value to reproduce the image according to a user's selection or an algorithm of another program being executed. A predictive value is generated by using and a decoded original pixel value image obtained by adding an inversely quantized inverse transformed residual signal only when a quantized frequency transform coefficient exists.

To this end, the image decoding apparatus 500 decodes the bitstream and extracts a quantized frequency transform block representing information about pixel values of the current block of the image (S710). Inverse quantization step (S720) of inverse quantization of the quantized frequency transform block, inverse DCT transform when the inverse quantized frequency transform block is inverse DST transform when the 4x4 transform unit (TU) predicted by inter prediction and larger than the 4x4 transform unit Inverse transform step (S730) to inverse transform into a pixel region, a prediction step of generating a prediction value using a motion vector value (S740), the residual signal of the current block reconstructed in step S730 and each pixel of the current block predicted in step S740 The decoded bitstream is decoded through the addition step S750 of reconstructing the original pixel value of the current block by adding the predicted pixel values of. If there is no quantized frequency transform residual signal, the result of the prediction step S740 becomes the current block.

Here, the bitstream decoding step S710 corresponds to the operation of the bitstream decoding unit 510, the inverse quantization step S720 corresponds to the operation of the inverse quantization unit 520, and the inverse transform step S730 is an inverse transform unit. The operation S530 corresponds to the operation of the predictor 550, and the operation S750 corresponds to the operation of the adder 540, and thus detailed description thereof will be omitted.

An image encoding / decoding method according to an embodiment of the present invention may be realized by combining the image encoding method according to an embodiment of the present invention and the image decoding method according to an embodiment of the present invention.

In the image encoding / decoding method according to an embodiment of the present invention, a prediction block of a current block is generated and a residual block generated by subtracting the prediction block from the current block is generated to generate a frequency transform block according to a conversion unit. An image encoding step of generating the frequency transform block using another transform method and then quantizing and encoding the bit transform stream into bitstreams, extracting information about the quantized frequency transform block and transform unit from the bitstream, and inverting the quantized frequency transform block. Reconstruct the frequency transform block by quantization and inverse transform the frequency transform block to recover the residual block, and restore the frequency transform block using a different inverse transform method according to the extracted transform unit and refer to the reference block referred to by the motion vector. The reconstructed residual block by generating a prediction block from the By adding the prediction block and a picture decoding step of reconstructing the current block.

The image encoding step may be implemented as an image encoding step according to an embodiment of the present invention, and the image decoding step may be implemented as an image decoding step according to an embodiment of the present invention.

Although all the components constituting the embodiment of the present invention may be implemented in one independent hardware, each of some or all of the components are selectively combined to some or all of the functions combined in one or a plurality of hardware. It may be implemented as a computer program having a program module for performing the operation. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program may be stored in a computer readable storage medium and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are not intended to limit the technical idea of the present invention but to describe the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent scope should be interpreted as being included in the scope of the present invention.

As described above, according to the embodiment of the present invention, it is a very useful invention to generate the effect of improving the compression ratio by transforming the residual block or inversely transforming the frequency coefficient block by using another transform method or another inverse transform method according to the transform unit. .

Claims (4)

In the apparatus for decoding an image,
A bitstream decoder for generating a quantized frequency transform block from the bitstream;
An inverse quantizer for restoring a frequency transform block by inversely quantizing the quantized frequency transform block;
An inverse transform unit for restoring the residual block by inversely transforming the frequency transform block, selecting an inverse transform method in consideration of information on the size of a transform unit, and restoring the residual block;
A prediction unit generating a prediction block using inter prediction or intra prediction; And
An adder which adds the residual block and the prediction block to reconstruct a current block;
Including,
The inverse transform unit,
Consider the DST type conversion method only when the information on the size of the conversion unit means that the information on the size of the conversion unit is 4 × 4 size.
The DST type conversion technique
Four inputs {X (0), X (1), X (2), X (3)} and the product of the coefficients corresponding to each of them are summed and the four outputs {X '(0), X' (1), X '(2), X' (3)}
The transformation that produces the four output values {X '(0), X' (1), X '(2), X' (3)}
The coefficient corresponding to the first input value X (0), the coefficient corresponding to the second input value X (1), and the third input value X (2), used to find the first output value X '(0). The signs of the coefficients are all positive,
The sign of the coefficient corresponding to the first input value X (0) and the coefficient corresponding to the second input value X (1), used to find the second output value X '(1), is both positive,
The sign of the coefficient corresponding to the second input value X (1) and the coefficient corresponding to the third input value X (2), used to find the third output value X '(2), is both negative (-),
The sign of the coefficient corresponding to the second input value X (1), used to obtain the fourth output value X '(3), is negative and the sign of the coefficient corresponding to the third input value X (2) is positive. The video decoding apparatus, characterized in that the conversion is (+). delete The method according to claim 1,
The transform unit is divided into quadtree forms from coding units,
And a bitstream decoder to decode a split transform flag for each depth of the transform unit. The method according to claim 3,
The bitstream decoder,
And a coded block flag corresponding to the transform unit and indicating that all of the quantization coefficients are zero.

Images