KR101341998B1 - Inter Prediction Apparatus of Considering Rotational Motion and Video Encoding/Decoding Method and Apparatus Using Same - Google Patents

Abstract

The present invention relates to an inter prediction apparatus considering rotational motion and an image encoding / decoding method and apparatus using the same.

The present invention provides an apparatus for encoding an image, wherein the motion vector and the rotation information of the current block are determined by estimating the rotational movement of the current block of the image to compensate for the rotational movement of the current block, thereby predicting the current block to generate a prediction block. Prediction unit; A subtraction unit for generating a residual block by subtracting the prediction block from the current block; A transformer for converting the residual block into the frequency domain; A quantizer for quantizing the transformed residual block; And an encoder configured to output a bitstream by encoding the quantized residual block, the motion vector, and the rotation information.

According to the present invention, by efficiently estimating and compensating for the rotational motion of an image, the prediction accuracy of the image for encoding and decoding can be improved, thereby improving the encoding and decoding efficiency.

Image, sign, decoding, motion, estimation, compensation, prediction, rotation, straight line, cost

Description

Inter Prediction Apparatus of Considering Rotational Motion and Video Encoding / Decoding Method and Apparatus Using Same}

The present invention relates to an inter prediction apparatus considering rotational motion and an image encoding / decoding method and apparatus using the same. More specifically, the present invention relates to an apparatus for predicting the current block by efficiently estimating and compensating the motion of the current block when predicting the current block of the image, and an apparatus and method for encoding or decoding an image using the same.

Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) have developed superior video compression techniques superior to the existing MPEG-4 Part 2 and H.263 standards. The new standard is called H.264 / AVC (Advanced Video Coding) and was jointly released as MPEG-4 Part 10 AVC and ITU-T Recommendation H.264.

In such H.264 / AVC (hereinafter abbreviated as 'H.264'), integer discrete cosine transform (DCT) is called, and motion estimation and compensation of transform block size (Variable Block Size) It consists of Motion Estimation and Compensation, Quantization and Entropy Coding.

Temporal redundancy, which has the most data redundancy in a video, provides high data compression rate by performing motion estimation and motion compensation using the data of the reference picture, and encoding residual signals between the image and the original image compensated by the estimated motion vector. Bring.

Meanwhile, the movement occurring in the video may occur in various ways such as a linear motion, a rotation motion, a zoom, a pan, and the like. In the conventional motion estimation and compensation method, motion estimation is performed considering only linear motion. However, since the linear motion estimation alone does not accurately match in the rotational motion region, it is difficult to properly estimate the motion and thus does not provide sufficient compression performance.

In order to solve the above problems, the present invention, in encoding or decoding an image, improves the accuracy of prediction by estimating and compensating not only the linear movement but also the rotational movement of the current block of the image, thereby improving the encoding and decoding efficiency of the image. The main purpose is to improve.

In order to achieve the above object, the present invention provides a device for encoding an image, comprising: determining the motion vector and rotation information of the current block by estimating the rotational movement of the current block of the image to compensate for the rotational movement of the current block; A predictor configured to predict a prediction block and generate a prediction block; A subtraction unit for generating a residual block by subtracting the prediction block from the current block; A transformer for converting the residual block into the frequency domain; A quantizer for quantizing the transformed residual block; And an encoder configured to output a bitstream by encoding the quantized residual block, the motion vector, and the rotation information.

In addition, according to another object of the present invention, in the video encoding method, by predicting the rotational motion of the current block of the image to determine the motion vector and rotation information of the current block to predict the current block by compensating the rotational movement of the current block A prediction step of generating a prediction block; Subtracting the prediction block from the current block to generate a residual block; Transforming the residual block into a frequency domain; A quantization step of quantizing the transformed residual block; And an encoding step of outputting the bitstream by encoding the quantized residual block, the motion vector, and the rotation information.

According to still another object of the present invention, there is provided an apparatus for decoding an image, comprising: a decoder configured to decode a bitstream to extract a residual block, a motion vector, and rotation information; An inverse quantizer for inversely quantizing the residual block; An inverse transform unit for inversely transforming an inverse quantized residual block; A prediction unit generating a prediction block by predicting the current block by compensating for the rotational movement of the current block of the image using the motion vector and the rotation information; And an adder configured to reconstruct the current block by adding the inversely transformed residual block and the predictive block.

According to still another object of the present invention, there is provided a method of decoding an image, comprising: a decoding step of decoding a bitstream to extract a residual block, a motion vector, and rotation information; An inverse quantization step of inverse quantizing the residual block; Inverse transforming the inversely quantized residual block; A prediction step of predicting the current block by generating a prediction block by compensating for the rotational movement of the current block of the image using the motion vector and the rotation information; And an addition step of reconstructing the current block by adding the inverse transformed residual block and the prediction block.

According to still another object of the present invention, there is provided an apparatus for inter prediction of a current block of an image, the apparatus comprising: a reference picture storage unit for storing a reference picture; Determines the predictive motion vector to set the search region in the reference picture, and calculates the coding cost by rotating the reference block for one or more search position points in the search region by one or more rotation angles, but at the search position point having the minimum coding cost. A motion estimator for determining a motion vector and a rotation angle with respect to a reference block of as a motion vector and a rotation angle of the current block; And a motion compensator configured to rotate the reference block indicated by the motion vector in the reference picture by the rotation angle indicated by the rotation angle, and output a prediction block of the current block.

According to still another object of the present invention, there is provided a method of encoding an image, comprising: a motion estimation step of estimating a motion including a rotational motion of a current block by using one or more reference pictures; A motion compensation step of generating a prediction block by compensating for the motion of the current block based on the estimated motion; And an encoding step of generating and encoding a residual block by subtracting the current block and the prediction block, wherein the motion estimation step provides a motion vector and rotation information for motion compensation of the current block. do.

According to still another object of the present invention, there is provided a method of encoding an image, comprising: a motion estimation step of estimating a motion including a rotational motion of a current block by using one or more reference pictures; A motion compensation step of generating a prediction block by compensating for the motion of the current block based on the estimated motion; And an encoding step of generating and encoding a residual block by subtracting the current block and the prediction block, wherein the motion estimation step comprises generating a motion vector and reference picture information for motion compensation of the current block. to provide.

According to still another object of the present invention, there is provided a method of decoding an image, comprising: a decoding step of decoding a bitstream to generate a residual block of a current block and extracting motion vectors and rotation information; A motion compensation step of generating a prediction block of the current block by performing motion compensation including a rotational motion of the current block based on the motion vector and the rotation information; And a reconstruction step of reconstructing the current block by adding the prediction block and the residual block.

According to still another object of the present invention, there is provided a method of decoding an image, comprising: a decoding step of decoding a bitstream to generate a residual block of a current block and extracting a motion vector and reference picture information; A motion compensation step of generating a prediction block of the current block by performing motion compensation including rotational motion of the current block based on the motion vector and the reference picture information; And a reconstruction step of reconstructing the current block by adding the prediction block and the residual block.

As described above, according to the present invention, the prediction accuracy of the image for encoding and decoding can be improved by efficiently estimating and compensating the rotational motion of the image, thereby improving the encoding and decoding efficiency.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. It should be noted that, in adding reference numerals to the constituent elements of the drawings, the same constituent elements are denoted by the same reference symbols as possible even if they are shown in different drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

In addition, in describing the component of this invention, terms, such as 1st, 2nd, A, B, (a), (b), can be used. These terms are intended to distinguish the constituent elements from other constituent elements, and the terms do not limit the nature, order or order of the constituent elements. If a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected to or connected to that other component, but there may be another configuration between each component. It is to be understood that the elements may be "connected", "coupled" or "connected".

1 is an exemplary diagram illustrating a process of searching for a macro block in a reference block.

There are various techniques for encoding an image. Typically, an image may be divided into pixels and encoded in pixels, or an image may be divided into blocks and encoded in blocks. In the case of encoding in block units, one picture to be encoded may be divided into a macroblock, a subblock constituting the macroblock, and the like. Prediction is performed on a block basis and encoding is performed.

Since an image is usually composed of 30 pictures in one second, it is very difficult to distinguish the difference between a picture and a neighboring picture with a human eye because the difference between pictures is small. Therefore, if an image is output as 30 pictures in one second, humans recognize that each picture is continuous.

Therefore, if the previous picture and the current picture are similar, the unknown pixel value of the current picture can be predicted from the pixel value of the already-known picture constituting the previous picture. Such a prediction technique is called inter prediction. Such inter prediction is based on a motion prediction technique. Motion prediction is performed in a manner of referring to a previous picture or referring to both a previous picture and a future picture based on the time axis. A picture referred to to encode or decode a current picture is called a reference picture.

Referring to FIG. 1, the most similar to the current block 112 in the search area 122 of the reference picture 120 to predict the motion by estimating and compensating the motion of the current block 112 of the current picture 110. Find reference block 124. The motion vector represents a difference in relative position between the found reference block and the current block.

In FIG. 1, the reference block 124 most similar to the current block 112 is located at a point in the reference picture 120 where the block corresponding to the current block 112 is moved diagonally, so that the current block is referred to by the current block 112. It can be seen that the reference block 124 of the picture 120 moved in a straight line diagonally.

2 is a block diagram schematically illustrating an electronic configuration of an image encoding apparatus according to a first embodiment of the present invention.

The image encoding apparatus 200 according to the first embodiment of the present invention includes a predictor 210, a subtractor 220, a transformer 230, a quantizer 240, an encoder 250, and an inverse quantizer ( 260, an inverse transform unit 270, an adder 280, and a deblocking filter unit 280.

The image encoding apparatus 200 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), a PlayStation Portable ), A mobile communication terminal, and the like, and may be a communication device such as a communication modem for performing communication with various devices or wired / wireless communication networks, a memory for storing various programs for encoding an image and data, a program And a microprocessor for calculating and controlling the microprocessor.

The predictor 210 predicts a current block of an image and generates a predicted block. That is, the prediction unit 210 predicts each pixel value of the current block to be encoded in the image by using inter prediction, thereby converting the predicted pixel value to the pixel value of each pixel. Create a predicted block having To this end, when the current block is input, the prediction unit 210 determines the motion vector and the rotation information of the current block by estimating the motion of the current block including the rotation motion, and determines the motion of the current block by using the determined motion vector and the rotation information. By compensating, the prediction block is generated by predicting the current block.

In an embodiment, the predictor 210 generates one or more rotational reference pictures in which the reference picture is rotated by one or more rotational angles, wherein the one or more rotational reference pictures are divided into one or more blocks, and are divided within one palmar picture. Each block is rotated at the same angle. As described above, a block that is divided and rotated in the rotation reference picture is hereinafter referred to as a reference block. The prediction unit 210 performs motion estimation by using one or more rotational reference pictures, and rotates the motion information and the motion vector including the reference picture information, which is information about a reference picture having a minimum encoding cost, and the motion information of the current block. Can be determined as a vector. The reference picture information includes information about the rotation angle of the reference picture, that is, the rotation angle of blocks divided in the reference picture.

In another embodiment, the prediction unit 210 determines a predictive motion vector to set a search region in the reference picture, and estimates a motion estimation including a rotation motion estimation for a reference block at one or more search position points in the search region. The cost, ie, the motion estimation cost, may be calculated to determine the motion vector and rotation information for the reference block at the search location point having the minimum coding cost as the motion vector and rotation information of the current block. In this case, the rotation information may include a flag for identifying that the current block is a block that has performed the rotation motion estimation and a rotation angle of the reference block having the minimum encoding cost.

The subtractor 220 subtracts the prediction block from the current block to generate a residual block. That is, the subtractor 220 generates a residual block having a residual signal by calculating a difference value between an original pixel value of each pixel of the current block and a predicted pixel value of each pixel of the prediction block. do.

The transform unit 230 transforms the residual block into the frequency domain. That is, the transformer 230 converts the residual block into a frequency domain to generate a residual block having a frequency coefficient. Here, the transform unit 230 transforms the residual block, and various transforms for transforming the image signal of the time axis to the frequency axis, such as a Hadamard transform, a discrete cosine transform based transform, or the like. The residual signal can be converted into a frequency domain using a technique, and the residual signal converted into the frequency domain becomes a frequency coefficient.

The quantization unit 240 quantizes the residual block transformed by the transformer 230. That is, the quantization unit 240 quantizes the frequency coefficients of the residual block to generate quantization frequency coefficients. Here, the quantization unit 240 uses Dead Zone Uniform Threshold Quantization (DZUTQ), Quantization Weighted Matrix, or the quantization method to quantize the frequency coefficient of the residual block. Etc. can be used.

The encoder 250 generates a bitstream by encoding the frequency coefficients of the residual block quantized by the quantizer 240. In addition, when the motion vector and the rotation information are transmitted from the predictor 210, the encoder 250 may code the quantized frequency coefficients of the residual block to output a bitstream. As such an encoding technique, entropy encoding technology may be used, but various other encoding techniques may be used without being limited thereto.

The inverse quantizer 260 inverse quantizes the residual block quantized by the quantizer 230. That is, the inverse quantizer 260 inversely quantizes the quantized frequency coefficients of the quantized residual block to generate frequency coefficients.

The inverse transform unit 270 inversely transforms the residual block inversely quantized by the inverse quantizer 260. That is, the inverse transform unit 270 inversely transforms the frequency coefficients of the inverse quantized residual block to restore the residual block having the pixel signal on the time axis.

The adder 280 reconstructs the current block by adding the prediction block predicted by the predictor 210 to the residual block inversely transformed by the inverse transformer 270.

The deblocking filter 290 deblocking filtering the current block reconstructed by the adder 280. Here, the deblocking filtering refers to an operation of reducing block distortion generated by encoding an image in block units, and applying a deblocking filter to a block boundary and a macroblock boundary, or applying or applying a deblocking filter only to a macroblock boundary. You can optionally use one of the methods that does not use filters.

In this way, the current block reconstructed by the adder 280 and deblocked filtered by the deblocking filter 290 is input to the predictor 210 and stored as a reference picture used when predicting the next picture.

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

When the current picture of the original image to be encoded is input, the image encoding apparatus 200 divides the image into macroblocks or subblock units of the macroblocks, and determines an optimal block mode among various block modes (or encoding modes). Predict and encode a current block to be encoded according to a mode.

In this case, when the inter mode is determined as the block mode and inter prediction is performed, the image encoding apparatus 200 predicts the motion of the current block to be encoded using another previously encoded and decoded picture. That is, the image encoding apparatus 200 determines a motion vector and rotation information of the current block by estimating a motion including the rotation motion of the current block (S310), and determines the motion of the current block by using the determined motion vector and the rotation information. By compensating, the current block is predicted to generate a prediction block (S320).

When the prediction block is generated, the image encoding apparatus 200 generates a residual block by subtracting the prediction block from the current block (S330), transforms and quantizes the residual block (S340), and quantizes the residual block, motion vector, and rotation information. The bitstream is generated by encoding the signal (S350).

As described above, the image encoded in the bitstream by the image encoding apparatus 200 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, The image decoding apparatus may be transmitted to an image decoding apparatus to be described later through various communication interfaces such as a universal serial bus (USB), and may be decoded by the image decoding apparatus to restore and reproduce the image.

4 is a block diagram schematically illustrating an electronic configuration of an image decoding apparatus according to a second embodiment of the present invention.

The image decoding apparatus 400 according to the second embodiment of the present invention includes a decoder 410, an inverse quantizer 420, an inverse transformer 430, a predictor 440, an adder 450, and deblocking. It may be configured to include a filter unit 460.

The video decoding apparatus 400 may be a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP), or a PlayStation Portable (PSP). ), 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 decoding an image, and executing a program. Means a variety of devices including a microprocessor for operation and control.

The decoder 410 decodes the bitstream to extract the residual block, the motion vector, and the rotation information. That is, the decoder 410 decodes a bitstream, which is an image encoded by the image encoding apparatus 200, and extracts a residual block, motion vector, and rotation information including pixel information of a current block of the image.

The inverse quantizer 420 inverse quantizes the residual block extracted by the decoder 410, and the inverse transformer 430 inversely transforms the inverse quantized residual block. Here, the inverse quantization and the inverse transform may be inversely transformed with the inverse quantization in the same manner as the inverse quantization and inverse transformation described above with reference to FIG. 2.

The prediction unit 440 generates a prediction block by predicting the current block by compensating for the motion including the rotation motion of the current block of the image by using the motion vector and the rotation information extracted by the decoder 410.

As an example, the prediction unit 440 divides the reference picture into one or more blocks and rotates the reference picture by the rotation angle indicated by the reference picture information to generate a rotation reference picture, and the reference indicated by the motion vector in the rotation reference picture. Searching for the block may compensate for the movement including the rotational movement of the current block.

As another embodiment, the predictor 440 may search for a reference block indicated by the motion vector in the reference picture and rotate the reference block by the rotation angle indicated by the rotation information to compensate for the motion including the rotational movement of the current block. Can be.

The adder 450 reconstructs the current block by adding the predicted block predicted by the predictor 440 to the residual block inversely transformed by the inverse transformer 430.

The deblocking filter 460 deblocks the current block reconstructed by the adder 450 to reduce block distortion by reducing an error that may occur in the quantization process.

The current block reconstructed by the adder 450 is combined in picture units and output as a reconstructed image, and the current block deblocked and filtered by the deblocking filter 460 when the prediction unit 440 predicts the next picture. It is stored as a reference picture to be used.

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

The image decoding apparatus 400 that receives and stores a bitstream of an image through a wired or wireless communication network or a cable, decodes and restores the image in order to reproduce the image according to a user's selection or an algorithm of another program being executed. To this end, the image decoding apparatus 400 extracts the residual block by decoding the bitstream, and extracts a motion vector and rotation information from the bitstream at step S510. When the residual block, the motion vector, and the rotation information are extracted, the image decoding apparatus 400 inverse quantizes and inversely transforms the extracted residual block (S520).

In addition, the image decoding apparatus 400 generates a prediction block by predicting the current block by compensating for the rotational movement of the current block of the image using the motion vector and the rotation information extracted in step S510 (S530).

The image decoding apparatus 400 reconstructs the current block by adding the residual block inversely transformed in operation S520 and the prediction block generated in operation S530 (S540). The image encoding apparatus 400 stores the reconstructed current block in picture units and outputs a reconstructed image.

6 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for encoding according to a third embodiment of the present invention.

The inter prediction apparatus for encoding according to the third embodiment of the present invention may be implemented by the prediction unit 210 of the image encoding apparatus 200 described above with reference to FIG. 2. Hereinafter, the inter prediction apparatus for encoding according to the third embodiment of the present invention will be abbreviated as a predictor 210.

The predictor 210 according to the third embodiment of the present invention may include a motion estimator 610, a motion compensator 620, and a reference picture storage 630.

The motion estimator 610 determines a predicted motion vector, sets a search area in a reference picture stored in the reference picture storage 630, and sets a reference block for one or more search position points in the search area by one or more rotation angles. The coding cost is calculated by rotating, but the motion vector and the rotation angle for the reference block at the search location point having the minimum coding cost are determined as the motion vector and the rotation angle of the current block.

That is, the motion estimator 610 determines the predicted motion vector, sets a predetermined region around the reference block indicated by the predicted motion vector in the reference picture as the search region, and sets one or more points in the search region to the search position. Set to point. The motion estimation unit 610 rotates the reference block at each search position point by one or more rotation angles, calculates the encoding cost of the rotated reference blocks, and compares each encoding position point by comparing the encoding cost according to each rotation angle. Determine the rotation angle of the reference block with the minimum coding cost in. The motion estimation unit 610 compares the minimum coding cost at each search location point with each other to determine the search location point having the minimum coding cost, and determines the rotation angle of the reference block at the search location point as the final rotation angle. The position of the current block relative to the reference block is determined as a motion vector.

Here, the coding cost may be calculated based on the difference between the current block and the reference block, and may be a sum of squared difference (SSD) and a sum of absolute difference (SAD). have. In addition, the encoding cost may be a rate-distortion cost calculated by assuming that the reference block is a prediction block of the current block.

In addition, the motion estimator 610 may determine a predicted motion vector for motion estimation using the motion vector of the neighboring block of the current block. The motion of the neighboring block that performs linear motion compensation among one or more neighboring blocks of the current block The determination may be performed using a vector or using an average value of a motion vector of a neighboring block regardless of whether linear motion compensation or rotational motion compensation is performed. That is, the motion vector of the neighboring block that performs the linear motion compensation instead of the neighboring block that performs the rotational motion compensation among the neighboring blocks may be determined as the motion vector of the current block, and the motion vector obtained by averaging the motion vector of the neighboring block is the current block. Can be determined as. Since the performance of the rotational motion estimation varies depending on the rotation center point in the region where the rotational motion of the image is located, the optimal motion vector determined by estimating the rotational motion and the linear motion only determined by estimating the rotational motion may be inconsistent. Therefore, the size of the difference vector between the motion vector of the current block and the predictive motion vector can be reduced by using the motion vector of the neighboring block determined by estimating only the linear motion as the predictive motion vector.

In addition, the motion estimator 610 may determine the predicted motion vector for calculating the difference vector using the motion vector of the neighboring block of the current block. Whether the current block performs the rotational motion compensation or the linear motion compensation is performed. According to the prediction motion vector can be determined differently. That is, when the current block performs the rotational motion compensation, the prediction motion vector for the difference vector may be determined using only the motion vectors of the neighboring blocks that perform the rotational motion compensation among the neighboring blocks, and the current block performs the linear motion compensation. In one case, a prediction motion vector for a difference vector may be determined using only motion vectors of neighboring blocks which have performed linear motion compensation among neighboring blocks.

Referring to FIG. 12, for example, a predictive motion vector for motion estimation may be performed by performing motion estimation using average values of A, B, and C blocks as predictive motion vectors, regardless of whether or not rotation motion compensation is performed. The motion vector of the used C block can be used as a prediction vector. In the predicted motion vector for the difference vector, when the current block performs the rotational motion compensation, the difference vector is obtained by subtracting the average value of the A and B blocks from the motion vector of the current block, and the current block performs the linear motion compensation. In this case, the difference vector is obtained by subtracting the motion vector of the C block from the motion vector of the current block.

In addition, when setting the rotation angle, the motion estimation unit 610 may set one or more rotation angles within a preset rotation angle range. In addition, the motion estimator 610 may generate a flag for identifying that the current block is a block in which the rotation motion estimation is performed, and output the flag and the rotation angle as rotation information.

In addition, the motion estimator 610 determines a predicted motion vector to set a search region in the reference picture, calculates a linear motion estimation cost for one or more search position points in the search region, and rotates about one or more search position points. Determine whether to estimate motion to estimate rotational motion for one or more search location points, calculate cost of rotational motion estimation, and determine the minimum of linear motion estimates and rotational motion estimates for at least one search location point. The coding cost may be determined, and a motion vector and a rotation angle of the reference block at the search location point having the minimum coding cost among the minimum coding costs for each search location point may be determined as the motion vector and the rotation angle of the current block.

The motion compensator 620 generates and outputs a prediction block of the current block by rotating the reference block indicated by the motion vector in the reference picture stored in the reference picture storage 630 by the rotation angle indicated by the rotation angle.

The reference picture storage unit 630 stores the reconstructed current block output from the deblocking filter unit 290 on a picture basis and stores the picture as a reference picture.

7 is a block diagram schematically illustrating a motion estimation unit according to a third embodiment of the present invention.

The motion estimator 610 of the predictor 210 described above with reference to FIG. 6 includes a predicted motion vector determiner 710, a search position setter 720, a rotation angle setter 730, and a center point setter 740. , A block rotation unit 750, an encoding cost calculator 760, and a motion vector determiner 770.

The prediction motion vector determiner 710 determines the prediction motion vector using the motion vector of the neighboring block. The search position setting unit 720 sets a search region in the reference picture using the determined predicted motion vector, and sets one or more search position points within the set search region. The rotation angle setting unit 730 sets one or more rotation angles within a preset rotation angle range.

The center point setting unit 740 sets a center point for a reference block at one or more search position points. That is, the center point setting unit 740 sets a center point that is the center of rotation of the reference block. The center point may be a search location point, and various positions in the search block, such as a point that becomes the center of the reference block at the search location point, may be Can be the center point. According to the setting of the center point, the efficiency of the motion estimation may vary.

The block rotating unit 750 rotates a reference block for one or more search position points at one or more rotation angles set based on each center point.

The encoding cost calculator 760 calculates an encoding cost for each reference block rotated by the block rotation unit 750 and calculates an encoding cost for each reference block rotated at each rotation angle at one search position point.

The motion vector determiner 770 determines a reference block having a minimum encoding cost among encoding costs for a reference block rotated by one or more rotation angles at one or more search position points calculated by the encoding cost calculator 760, The rotation angle of the determined reference block is determined as the rotation angle for the rotation motion estimation, and the relative position of the current block with respect to the reference block is determined as the motion vector.

8 is a flowchart illustrating an inter prediction method for encoding according to a third embodiment of the present invention.

The prediction unit 210 according to the third embodiment of the present invention determines the prediction motion vector using the motion vector of the neighboring block (S810). The predictor 210 sets a search region in the reference picture using the determined predicted motion vector and sets one or more search position points in the set search region (S820).

The prediction unit 210 calculates a linear motion estimation cost at each search position point (S830). That is, the prediction unit 210 calculates the linear motion estimation cost by calculating the search block at each search location point and the SSD or SAD of the current block.

The predictor 210 determines whether to perform rotation motion estimation at each search position point (S840). Here, the predictor 210 may determine whether to perform rotation motion estimation at each search position point using the value of the quantization frequency coefficient. That is, the prediction unit 210 generates a reference block corresponding to each search position point as a prediction block, generates a difference between the current block and the prediction block as a residual block, and then transforms and quantizes the residual block to one of the quantized frequency coefficients. It is determined whether the ideal is not '0', and when at least one is not '0', it is determined to perform rotation motion estimation at each search position point, and when the quantization frequency coefficients are all '0', the corresponding search is performed. Determine not to perform rotational motion estimation at the location point. This is because the prediction is accurate when the quantization frequency coefficients are all '0' as a result of encoding without performing the rotation motion estimation, and thus it is not necessary to perform the rotation motion estimation.

As a result of checking in step S840, when performing rotational motion estimation at each search position point, at least one rotation angle at each search position point is set, a center point at each search position point is set, and at each search position point, By calculating the encoding cost by rotating the corresponding reference block by one or more rotation angles, the rotational motion estimation cost at each search position point is calculated (S850).

As a result of checking in step S840, when the rotation motion estimation is not performed at each search position point, only the linear motion estimation cost calculated in step S830 remains at the search position point.

The prediction unit 210 compares the linear motion estimation cost and the rotational motion estimation cost (calculated only when the rotational motion estimation is performed) at one or more searched location points, which are calculated in steps S830 and S850, and are each search position. The minimum coding cost at the point is determined (S860).

When the minimum coding cost at each search location point is determined, the prediction unit 210 determines the motion vector and the rotation angle of the reference block at the search location point having the minimum coding cost among the minimum coding costs at each search location point. It is determined as the motion vector and the rotation angle (S870).

An example of determining the minimum coding cost at each search position point assuming that two search position points are set in the search area, the rotation angle range is ± 30 °, and the magnitude of the rotation angle is 10 °. When it is determined that the rotational motion estimation is not performed for the first search location point, the linear motion estimation cost for the first search location point is determined as the minimum coding cost. In the case of the second search position point, when it is determined to perform the rotation motion estimation, the reference block for the second search position point is rotated by six set rotation angles (± 10 °, ± 20 °, ± 30) and each rotation After calculating the six rotational motion estimation costs by calculating the encoding cost at the angle, the smallest of the linear motion estimation cost and the six rotational motion estimation costs at the second search position point (the rotation of the rotation angle of + 10 °) Assuming that the motion estimation cost is the minimum coding cost at the second search location point, determine the minimum coding cost. Again, the minimum coding is performed between the linear motion estimation cost for the first search location point, which is the minimum coding cost at the first search location point, and the rotation motion estimation cost with a rotation angle of + 10 °, which is the minimum coding cost at the second search location point. Determine a reference block having a cost (assuming that the minimum coding cost at the second search point is the minimum coding cost), that is, the rotation angle + 10 ° of the reference block at the second search location point as the rotation angle of the current block, The difference of the relative position between the reference block and the current block at the second search position point is determined as the motion vector of the current block. If the linear motion estimation cost without the rotation motion estimation is determined as the minimum coding cost, only the motion vector is output and the rotation angle is not output.

In operation S870, when the motion vector and the rotation angle of the current block are determined, the prediction unit 210 searches for the reference block in the reference picture using the motion vector, and predicts by compensating the rotational movement by rotating the retrieved reference block by the rotation angle. A block is generated (S880).

9 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for decoding according to a third embodiment of the present invention.

The inter prediction apparatus for decoding according to the third embodiment of the present invention may be implemented by the prediction unit 440 of the image decoding apparatus 400 described above with reference to FIG. 4. Hereinafter, the inter prediction apparatus for decoding according to the third embodiment of the present invention is abbreviated as a predictor 440.

The predictor 440 according to the third embodiment of the present invention includes a motion compensator 910 and a reference picture storage 920.

When the motion compensator 910 receives the motion vector and the rotation information from the decoder 410, the motion compensator 910 checks whether or not the current block is a block on which the rotation motion estimation has been performed using a flag included in the rotation information. In the case of the performed block, the prediction block is generated by rotating the reference block indicated by the motion vector in the reference picture stored in the reference picture storage unit 920 by the rotation angle included in the rotation information.

The reference picture storage unit 920 stores the deblocked current block output from the deblocking filter unit 460 in picture units and stores the picture as a reference picture.

FIG. 10 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for encoding according to a fourth embodiment of the present invention.

Inter prediction apparatus for encoding according to a fourth embodiment of the present invention Through FIG. 2, the prediction unit 210 of the image encoding apparatus 200 may be implemented. Hereinafter, the inter prediction apparatus for encoding according to the fourth embodiment of the present invention is abbreviated as a predictor 210.

The predictor 210 according to the fourth embodiment of the present invention may include a motion estimator 1010, a motion compensator 1020, a reference picture storage unit 1030, and a reference picture rotation unit 1040. .

The motion estimator 1010 determines a reference block having the lowest coding cost among the reference blocks of the one or more rotational reference pictures, and identifies a reference picture including a motion vector for the determined reference block and the determined reference block. The information is determined, and the motion vector, the reference picture information, and the like are output to the encoder 250. That is, the motion estimation unit 1010 calculates and compares encoding cost for each reference block among reference blocks of one or more reference pictures rotated by the reference picture rotation unit 1040 to determine a reference block having the minimum encoding cost. do.

Here, the motion estimator 1010 may determine the predicted motion vector using the motion vector of the neighboring block that has performed linear motion compensation among one or more neighboring blocks of the current block. In addition, the motion estimator 1010 may set a search region using a predictive motion vector in one or more reference pictures, and determine a reference block having a minimum encoding cost in the search region. Here, the coding cost may be calculated based on the difference between the current block and the reference block, and may be a sum of squared difference (SSD) and a sum of absolute difference (SAD). have. In addition, the encoding cost may be a rate-distortion cost calculated by assuming that the reference block is a prediction block of the current block. In addition, the rotation angle may be determined within a preset rotation angle range.

The motion compensator 1020 performs motion using the motion vector determined by the motion estimator 1010 in the rotated reference picture in which the reference picture is rotated by the angle identified by the reference picture information determined by the motion estimator 1010. By compensating, a prediction block of the current block is generated and output to the subtractor 220 and the adder 280.

The reference picture rotating unit 1040 divides the reference picture stored in the reference picture storage unit 1030 into a plurality of blocks and rotates the divided blocks at one or more preset rotation angles to generate one or more rotating reference pictures. That is, a plurality of rotational reference pictures are generated according to the rotational angle. For example, when the rotation angles are ± 10 ° and ± 20 °, four rotational reference pictures are generated. The divided blocks of each rotating reference picture are rotated by the corresponding rotation angle.

The reference picture storage unit 1030 stores the deblocked current block output from the deblocking filter unit 290 in picture units and stores the picture as a reference picture.

11 is a flowchart illustrating an inter prediction method for encoding according to a fourth embodiment of the present invention.

The prediction unit 210 according to the fourth embodiment of the present invention divides the reference picture into one or more blocks (S1110), and generates one or more rotating reference pictures by rotating the blocks divided from the reference picture by one or more rotation angles. (S1120). The prediction unit 210 calculates and compares encoding costs of reference blocks of one or more rotational reference pictures by estimating the motion of the current block using the generated one or more rotational reference pictures to determine a reference block having the minimum encoding cost ( In operation S1140, reference picture information, a motion vector, and a rotation angle of the reference block having the minimum coding cost are determined. The prediction unit 210 outputs the reference block indicated by the motion vector in the reference picture according to the determined reference picture information as the prediction block of the current block (S1150).

Embodiments of the present invention described above may be realized as another embodiment to be described later. According to an image encoding method according to a fifth embodiment of the present invention, an image encoding apparatus estimates a motion including a rotation movement of a current block by using one or more reference pictures, and based on the estimated motion, calculates a motion of the current block. A compensation block may be generated by compensating, and a residual block may be generated and encoded by subtracting the current block and the prediction block. Here, in estimating the motion, the image encoding apparatus may generate a motion vector and rotation information for motion compensation of the current block. The rotation information may indicate a rotation angle of the reference block indicated by the motion vector.

In addition, according to the image decoding method according to the fifth embodiment of the present invention, the image encoding apparatus decodes a bitstream to generate a residual block of the current block, extract motion vector and rotation information, and extract motion vector and rotation information. Based on the motion compensation including the rotational motion of the current block, the prediction block of the current block is generated, and the prediction block and the residual block may be added to restore the current block. Here, the rotation information may indicate the rotation angle of the reference block indicated by the motion vector, as described above.

Further, according to the image encoding method according to the sixth embodiment of the present invention, the image encoding apparatus estimates a motion including a rotational movement of the current block by using one or more reference pictures, and based on the estimated motion, The prediction block may be generated by compensating for the motion, and the encoding may be performed to generate and encode a residual block by subtracting the current block and the prediction block. Here, in estimating motion, the image encoding apparatus may generate a motion vector and reference picture information for motion compensation of the current block. The reference picture information includes information indicating whether to use a reference picture returned at a predetermined angle. It may include a reference picture index indicating one of one or more reference pictures having different rotation angles.

Also, according to the video encoding method according to the sixth embodiment of the present invention, the video encoding apparatus decodes the bitstream to generate a residual block of the current block, extract motion vector and reference picture information, and extract motion vector and reference picture information. The prediction block of the current block may be generated by performing motion compensation including the rotational movement of the current block, and the current block may be reconstructed by adding the prediction block and the residual block. Here, the reference picture information may include information indicating whether to use the reference picture rotated at a predetermined angle as described above, and includes a reference picture index indicating one of one or more reference pictures having different rotation angles. can do.

As described above, according to the present invention, the current block can be predicted by estimating and compensating not only the linear motion but also the rotational motion of the image, thereby improving the accuracy of prediction and the image encoding and decoding efficiency through the prediction.

Further, according to the present invention, it is possible to determine whether to estimate the rotational motion by using the coefficient value of the quantization frequency coefficient at each search position point in the search region of the reference picture, and to estimate the rotational motion only when the rotational motion estimation is necessary. In addition, the computational complexity can be reduced by lowering the complexity of the calculation according to the rotational motion estimation and compensation.

While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. Codes and code segments constituting the computer program may be easily inferred by those skilled in the art. Such a computer program can be stored in a computer-readable storage medium, readable and executed by a computer, thereby realizing an embodiment 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.

It is also to be understood that the terms such as " comprises, "" comprising," or "having ", as used herein, mean that a component can be implanted unless specifically stated to the contrary. But should be construed as including other elements. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

The foregoing description is merely illustrative of the technical idea of the present invention, and various changes and modifications may be made by those skilled in the art without departing from the essential characteristics of the present invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of 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, the present invention is applied to a method and apparatus for predicting a current block by efficiently estimating and compensating the motion of a current block when predicting a current block of an image, and an apparatus and method for encoding or decoding an image using the same. Thus, by effectively estimating and compensating for the rotational motion of the image, the prediction accuracy of the image for encoding and decoding can be improved, thereby generating an effect of improving the encoding and decoding efficiency.

1 is an exemplary diagram illustrating a process of searching for a macro block in a reference block;

2 is a block diagram schematically illustrating an electronic configuration of a video encoding apparatus according to a first embodiment of the present invention;

3 is a flowchart illustrating a video encoding method according to a first embodiment of the present invention;

4 is a block diagram schematically illustrating an electronic configuration of an image decoding apparatus according to a second embodiment of the present invention;

5 is a flowchart illustrating an image decoding method according to a second embodiment of the present invention;

6 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for encoding according to a third embodiment of the present invention;

7 is a block diagram schematically illustrating a motion estimation unit according to a third embodiment of the present invention;

8 is a flowchart illustrating an inter prediction method for encoding according to a third embodiment of the present invention;

9 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for decoding according to a third embodiment of the present invention;

10 is a block diagram schematically illustrating an electronic configuration of an inter prediction apparatus for encoding according to a fourth embodiment of the present invention;

11 is a flowchart illustrating an inter prediction method for encoding according to a fourth embodiment of the present invention;

12 is an exemplary view for explaining a process of determining a predicted motion vector according to the present invention.

Claims (22)

In the apparatus for encoding a video, A predictor configured to predict a rotational movement of the current block from the reference picture to generate a predicted block; A subtractor for subtracting the prediction block from the current block to generate a residual block; A transformer for converting the residual block into a frequency domain; A quantizer for quantizing the transformed residual block; And An encoder which outputs a bitstream by encoding the quantized residual block and rotation motion information about the rotation motion, The rotation motion information includes information indicating that the current block is a block for predicting rotation motion and information indicating a rotation angle of the reference picture. The method of claim 1, wherein the encoding unit, And encoding the identification information of the reference picture referred to for predicting the rotational motion to output the bitstream. The method of claim 2, The prediction unit further predicts the linear motion of the current block to generate the prediction block, The encoder encodes the motion vector information corresponding to the linear motion, and outputs the bitstream. The method of claim 1, wherein the prediction unit, And a linear motion of the current block is predicted from the rotated reference picture which rotates the reference area of the reference picture according to the rotational angle. 5. The method of claim 4, And the motion vector information includes a difference vector corresponding to a difference between a predicted motion vector determined from a motion vector of a neighboring block and the predicted linear motion. In the video encoding method, A prediction step of predicting rotational movement of the current block from the reference picture to generate a prediction block; Subtracting the prediction block from the current block to generate a residual block; Transforming the residual block into a frequency domain; A quantization step of quantizing the transformed residual block; And An encoding step of outputting a bitstream by encoding the quantized residual block and rotation motion information about the rotation motion Including, The rotation motion information includes information indicating that the current block is a block for predicting rotation motion and information indicating a rotation angle of the reference picture. In the apparatus for decoding an image, Residual blocks and rotational motion information of the current block are extracted by decoding the bitstream, wherein the rotational motion information indicates information indicating that the current block is a block for predicting rotational motion and a rotation angle of a reference picture referenced by the current block. A decoder including information indicating; An inverse quantizer for inversely quantizing the residual block; An inverse transform unit for inversely transforming the inverse quantized residual block; A prediction unit generating a prediction block of the current block from the reference picture using the rotation motion information; And An adder configured to add the inverse transformed residual block and the prediction block to restore the current block; Video decoding apparatus comprising a. The method of claim 7, wherein And the decoder further extracts identification information of a reference picture referenced to predict the rotational motion. The method of claim 7, wherein The decoder further extracts motion vector information corresponding to the linear motion of the current block, And the prediction unit generates the prediction block by predicting the linear motion and the rotational motion of the current block according to the motion vector information and the rotation motion information. In the method of decoding an image, Residual blocks and rotational motion information of the current block are extracted by decoding the bitstream, wherein the rotational motion information indicates information indicating that the current block is a block for predicting rotational motion and a rotation angle of a reference picture referenced by the current block. A decoding step including indicating information; An inverse quantization step of inversely quantizing the residual block; Inverse transforming the inverse quantized residual block; A prediction step of generating a prediction block of the current block from the reference picture using the rotation motion information; And An addition step of reconstructing the current block by adding the inverse transformed residual block and the prediction block Image decoding method comprising a. An apparatus for inter prediction of a current block of an image, A reference picture storage unit for storing a reference picture; Determine a prediction motion vector to set a search region in the reference picture, and calculate a coding cost by rotating a reference block for one or more search position points in the search region by one or more rotation angles, wherein the search position has a minimum coding cost A motion estimator for determining a motion vector and a rotation angle with respect to a reference block at a point as the motion vector and the rotation angle of the current block; And A motion compensator configured to rotate the reference block indicated by the motion vector in the reference picture by a rotation angle indicated by the rotation angle to output a prediction block of the current block Inter prediction apparatus comprising a. The method of claim 11, wherein the motion estimation unit, A predicted motion vector determiner configured to determine the predicted motion vector; A search position setting unit configured to set the search region using the predicted motion vector and to set the at least one search position point in the search region; A rotation angle setting unit configured to set the one or more rotation angles; A center point setting unit for setting a center point of a reference block with respect to the at least one search position point; A block rotation unit configured to rotate the reference block for the one or more search position points at the one or more rotation angles with respect to the center point; An encoding cost calculator configured to calculate an encoding cost for the reference block rotated by the one or more rotation angles for each of the one or more search position points; And A motion vector determiner that determines a motion vector and a rotation angle with respect to a reference block at the search location point having the minimum coding cost as the motion vector and the rotation angle Inter prediction apparatus comprising a. In the method of encoding an image, A motion estimation step of estimating a motion including a rotational motion of the current block using one or more reference pictures; A motion compensation step of generating a prediction block by compensating for the motion of the current block based on the estimated motion; And An encoding step of generating and encoding a residual block by subtracting the current block and the prediction block Wherein the motion estimation step generates motion vectors and rotation information for motion compensation of the current block. The method of claim 13, wherein the rotation information, And a rotation angle of the reference block indicated by the motion vector. In the method of encoding an image, A motion estimation step of estimating a motion including a rotational motion of the current block using one or more reference pictures; A motion compensation step of generating a prediction block by compensating for the motion of the current block based on the estimated motion; And An encoding step of generating and encoding a residual block by subtracting the current block and the prediction block Wherein the motion estimation step generates motion vectors and reference picture information for motion compensation of the current block. The method of claim 15, wherein the reference picture information, And information indicating whether or not to use a reference picture returned at a predetermined angle. The method of claim 15, wherein the reference picture information, And a reference picture index indicating one of one or more reference pictures having different rotation angles. In the method of decoding an image, A decoding step of decoding the bitstream to generate a residual block of the current block and extracting motion vectors and rotation information; A motion compensation step of generating a prediction block of the current block by performing motion compensation including a rotational motion of the current block based on the motion vector and the rotation information; And Restoring the current block by adding the prediction block and the residual block; Image decoding method comprising a. The method of claim 18, wherein the rotation information, And a rotation angle of the reference block indicated by the motion vector. In the method of decoding an image, A decoding step of decoding the bitstream to generate a residual block of the current block and extracting a motion vector and reference picture information; A motion compensation step of generating a prediction block of the current block by performing motion compensation including a rotational motion of the current block based on the motion vector and the reference picture information; And Restoring the current block by adding the prediction block and the residual block; Image decoding method comprising a. The method of claim 20, wherein the reference picture information, And information indicating whether to use the reference picture rotated at a predetermined angle. The method of claim 20, wherein the reference picture information, And a reference picture index indicating one of one or more reference pictures having different rotation angles.

Images