KR20110050914A - Method and apparatus for directional intra-prediction - Google Patents

Abstract

PURPOSE: A directional intra-prediction apparatus and prediction method thereof are provided to minimize the increase of overhead and to increase the accuracy and efficiency of the intra-prediction in video encoding. CONSTITUTION: A pixel row value generation unit(610) predicts pixel value summation by using prediction mode 0 or prediction mode 1 of N by N. The pixel row value generating unit generates N by 1 matrixes by using the pixel value summation. A pixel column value generation unit(620) predicts pixel value summation of N string by using a prediction mode 0 or prediction mode 1. The pixel column value generation unit generates N by 1 matrix by using pixel value summation. A whole pixel value prediction unit(630) predicts the summation of the whole pixel by using prediction mode 0 or prediction mode 1. A reference matrix generation unit(640) generates reference matrix Q by using the entire pixel value summation.

Description

Directional intra prediction device and its prediction method {Method and Apparatus for Directional Intra-prediction}

An embodiment of the present invention relates to a directional intra prediction device and a prediction method thereof. More specifically, when the mode to be executed is prediction mode 0 or prediction mode 1, the prediction block is determined by using neighboring pixels of the block to be encoded, thereby reducing the overhead in generating the bitstream compared to the H.264 standard. The present invention relates to a directional intra prediction apparatus and a prediction method thereof capable of increasing the accuracy and efficiency of intra prediction in video encoding without greatly increasing.

As information and communication technology including the Internet is developed, not only text and voice but also video communication are increasing. Conventional text-based communication methods are not enough to satisfy various needs of consumers, and accordingly, multimedia services that can accommodate various types of information such as text, video, and music are increasing. The multimedia data has a huge amount and requires a large storage medium and a wide bandwidth in transmission. Therefore, in order to transmit multimedia data including text, video, and audio, it is essential to use a compression coding technique.

The basic principle of compressing data is to eliminate redundancy in the data. Spatial overlap, such as the same color or object repeating in an image, temporal overlap, such as when there is almost no change in adjacent frames in a movie frame, or the same note over and over in audio, or high frequency of human vision and perception Data can be compressed by removing the psychological duplication taking into account the insensitive to.

As such a video compression method, interest in H.264 / AVC, which has further improved compression efficiency compared to MPEG-4 (Moving Picture Experts Group-4), has recently increased.

H.264 is a digital video codec standard with a very high data compression ratio, also called MPEG-4 Part 10 or Advanced Video Coding (AVC). This standard is based on the Video Coding Experts Group (VCEG) of the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) and the International Standardization Organization / International Electrotechnical Commission (ISO / IEC). This is the result of MPEG jointly forming and standardizing a Joint Video Team.

Various methods have been proposed to improve compression efficiency in compression encoding, and typical methods include a method using temporal prediction and a method using spatial prediction.

As shown in FIG. 1, the temporal prediction is performed by referring to a reference block 122 of another frame 120 that is adjacent in time when predicting the current block 112 of the current frame 110. That's the way. That is, in inter-prediction of the current block 112 of the current frame 110, the adjacent reference frame 120 is searched for in time, and the reference block (the most similar to the current block 112 in the reference frame 120) 122). Here, the reference block 122 is a block that can best predict the current block 112, and the block having the smallest sum of absolute difference (SAD) with the current block 112 may be the reference block 122. . The reference block 122 becomes a prediction block of the current block 112, and generates a residual block by subtracting the reference block 122 from the current block 112. The generated residual block is encoded and inserted into the bitstream. In this case, the relative difference between the position of the current block 112 in the current frame 110 and the position of the reference block 122 in the reference frame 120 is called a motion vector 130, and the motion vector 130 is also a residual block. Is encoded as follows. Temporal prediction is also referred to as inter prediction or inter prediction.

Spatial prediction is to obtain the prediction pixel value of the target block by using the reconstructed pixel value of the reference block adjacent to the target block in one frame, and directional intra-prediction (hereinafter referred to simply as intra prediction) It is also called intra prediction. H.264 specifies encoding / decoding using intra prediction.

Intra prediction is a method of predicting values of a current subblock by copying in a predetermined direction by using adjacent pixels in up and left directions for one sub-block, and encoding only the difference. In the intra prediction technique according to the H.264 standard, the prediction block for the current block is generated based on another block having the previous coding order. A value obtained by subtracting the current block and the prediction block is coded. The video encoder according to H.264 selects, for each block, a prediction mode in which the difference between the current block and the prediction block is minimal among the prediction modes.

Intra prediction according to the H.264 standard is illustrated in FIG. 2 in consideration of the position of adjacent pixels and the direction of the prediction used to generate predicted pixel values of 4 x 4 luma blocks and 8 x 8 luma blocks. Nine prediction modes as defined. The nine prediction modes are vertical prediction mode (prediction mode 0), horizontal prediction mode (prediction mode 1), DC prediction mode (prediction mode 2), Diagonal_Down_Left prediction mode (prediction mode 3), Diagontal_Down_Right prediction mode (depending on the prediction direction). Prediction mode 4), Vertical_Right prediction mode (prediction mode 5), Horizontal_Down prediction mode (prediction mode 6), Vertical_Left prediction mode (prediction mode 7), and Horizontal_Up prediction mode (prediction mode 8). Here, the DC prediction mode uses an average value of eight adjacent pixels.

In addition, four prediction modes are used for intra prediction processing for a 16 × 16 luma block, a vertical prediction mode (prediction mode 0), a horizontal prediction mode (prediction mode 1), a DC prediction mode (prediction mode 2), and a plane prediction mode. (Prediction mode 3) is that. The same four prediction modes are also used for intra prediction processing on 8 x 8 chroma blocks.

FIG. 3 shows an example of labeling for explaining the nine prediction modes of FIG. 2. In this case, a prediction block (region including a to p) for the current block is generated using the samples A to M that are decoded in advance. If E, F, G, and H cannot be decoded in advance, E, F, G, and H can be virtually generated by copying D to their positions.

FIG. 4 is a diagram for describing nine prediction modes of FIG. 2 using FIG. 3. Referring to the figure, in the prediction mode 0, the prediction block predicts the pixel value with the same pixel value for each vertical line. That is, the pixels of the prediction block predict the pixel value from the nearest pixels of the reference block located above the prediction block, and the reconstructed pixel values of the adjacent pixel A are converted into the first column pixels a, pixel e, pixel i and Set to the predicted pixel value for pixel m. Further, in the same way, second column pixel b, pixel f, pixel j and pixel n are predicted from the reconstructed pixel values of adjacent pixel B, and third column pixel c, pixel g, pixel k and pixel o are Predicted from the reconstructed pixel values, fourth column pixel d, pixel h, pixel l and pixel p predicts from the reconstructed pixel values of adjacent pixel D. As a result, as shown in Fig. 5A, a prediction block is generated in which the prediction pixel values of each column are the pixel values of pixel A, pixel B, pixel C and pixel D.

In addition, in the prediction mode 1, the prediction block predicts the pixel value with the same pixel value for each horizontal line. That is, the pixels of the prediction block predict the pixel value from the nearest pixels of the reference block located to the left of the prediction block, and the reconstructed pixel value of the adjacent pixel I is determined by the first row of pixels a, pixel b, pixel c and Set to the predicted pixel value for pixel d. Also, in the same way, the second row pixels e, pixel f, pixel g and pixel h are predicted from the reconstructed pixel values of adjacent pixel J, and the third row pixel i, pixel j, pixel k and pixel l are Predicted from the reconstructed pixel values, the fourth row pixel m, pixel n, pixel o and pixel p predicts from the reconstructed pixel values of adjacent pixel D. As a result, as shown in Fig. 5B, a prediction block is generated in which the prediction pixel values of each row are the pixel values of pixel I, pixel J, pixel K, and pixel L.

Also, in prediction mode 2, the pixels of the prediction block are equally replaced by the average of the pixel values of the upper pixels A, B, C and D and the left pixels I, J, K and L.

On the other hand, the pixels of the prediction block in the prediction mode 3 are interpolated in the lower left direction at a 45 ° angle between the lower-left and the upper-right, and the prediction in the prediction mode 4 The pixels of the block are extrapolated in the lower right direction at a 45 ° angle. In addition, the pixels of the prediction block in the prediction mode 5 are extrapolated in the lower right direction at an angle of about 26.6 degrees (width / height = 1/2) from the vertical. In addition, the pixels of the prediction block in the prediction mode 6 are extrapolated in the lower right direction at an angle of about 26.6 ° horizontally, and the pixels of the prediction block in the prediction mode 7 are in the lower left direction at an angle of about 26.6 ° vertically. Extrapolated, the pixels of the predictive block in the case of the prediction mode 8 are interpolated in an upward direction of about 26.6 degrees from the horizontal.

In prediction mode 3 to 8, the pixels of the prediction block may be generated from a weighted average of pixels A to M of the reference block to be decoded in advance. For example, in the prediction mode 4, the pixel d located at the top right of the prediction block may be estimated as in Equation 1. Here, round () is a function that rounds to integer places.

d = round (B / 4 + C / 2 + D / 4)

Meanwhile, as described above, the 16 × 16 prediction model for the luminance component includes four modes of prediction mode 0, prediction mode 1, prediction mode 2, and prediction mode 3.

In prediction mode 0, the pixels of the prediction block are extrapolated from the upper pixels, and in prediction mode 1, the pixels are extrapolated from the left pixels. In addition, in the prediction mode 2, the pixels of the prediction block are calculated as an average of upper pixels and left pixels. Finally, for prediction mode 3, a linear "plane" function is used that fits the upper and left pixels. This mode is more suitable for areas where the luminance changes smoothly.

As described above, in the H.264 standard, in each prediction mode except for the DC mode, the pixel value of the prediction block is generated according to the direction corresponding to each mode based on the adjacent pixels of the prediction block to be currently encoded.

However, in most cases, the current directional mode may be sufficient. However, since the direction of each prediction mode is limited depending on the image, it may be difficult to accurately predict the pixel value of the prediction block due to poor coding efficiency. have. In this case, the gain of entropy coding cannot be properly seen due to incorrect intra prediction, which causes a problem that the bit rate is unnecessarily increased.

One embodiment of the present invention is to solve the above-described problems, to provide a directional intra prediction device and a prediction method that can minimize the increase of overhead while increasing the accuracy and efficiency of intra prediction in video encoding The purpose.

In order to achieve the above object, the directional intra prediction apparatus according to an embodiment of the present invention uses the sum of pixel values of N columns using prediction mode 0 or prediction mode 1 from adjacent pixels of an N × N block to be encoded. A pixel column value generation unit for predicting a value and generating an N × 1 matrix; A pixel row value generation unit for predicting a sum of pixel values of N rows using prediction mode 0 or prediction mode 1 from adjacent pixels of an N × N block and generating an N × 1 matrix; An overall pixel value predictor for predicting a sum of all pixel values using prediction mode 0 or prediction mode 1 from neighboring pixels of the N × N block; A reference matrix generation for generating a reference matrix Q of [(2N + 1) x 1] using each matrix generated by the pixel column value generator and the pixel row value generator, and the values generated by the total pixel value predictor part; And a pixel value predictor for predicting each pixel value of the N × N prediction block by using the reference matrix Q generated by the reference matrix generator and the preset transformation matrix P.

Here, it is preferable that the pixel column value generator generates an N × 1 matrix representing the sum of pixel values of N columns from N adjacent pixels adjacent to the upper side of the N × N block using prediction mode 0. FIG.

In addition, the pixel row value generator may generate an N × 1 matrix representing the sum of pixel values of N rows from N adjacent pixels adjacent to the left side of the N × N block using prediction mode 1. FIG.

Directional intra prediction method according to an embodiment of the present invention for achieving the above object, predicts a matrix C _sum representing the sum of pixel values of four columns from the pixels adjacent to the upper side of the 4 x 4 block to be encoded Predicting a matrix R _sum representing a sum of pixel values of four rows from pixels adjacent to the left of the 4 × 4 block, and predicting a sum B _sum of all pixel values in the prediction block; Generating a reference matrix Q of 9 × 1 using each value predicted by the prediction step; And predicting each pixel value of the prediction block using the reference matrix Q and the preset transformation matrix P of 9 × 16.

Preferably, the directional intra prediction method further comprises determining whether the mode to be executed is prediction mode 0 or prediction mode 1. In this case, the prediction step is preferably performed when the prediction mode executed by the determination step is determined to be prediction mode 0 or prediction mode 1.

In addition, the directional intra prediction method may further include comparing a set threshold with a variance of pixels adjacent to the 4 × 4 block. In this case, the prediction step is preferably performed when the variance of pixels adjacent to the 4x4 block is larger than the set threshold by the comparison step.

Here, the matrix C _sum may be predicted as follows.

In this case, A, B, C, and D represent pixels adjacent to the upper side of the 4 × 4 block. In addition, the matrix R _sum can be predicted as follows.

In this case, I, J, K, and L represent pixels adjacent to the left side of the 4 × 4 block. In addition, the sum B _sum of the total pixel values can be predicted as follows.

Preferably, it is assumed that the reference matrix Q has the following relationship with the transformation matrix P.

Q = PI

here,

이며,

Is,

I represents the 4 x 4 prediction block in a 16 x 1 matrix.

As described above, according to the embodiment of the present invention, compared to the H.264 standard, it is possible to increase the accuracy of intra prediction while reducing the bit rate without significantly increasing the overhead of the bitstream generator.

Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings. In adding reference numerals to the components of each drawing, it should be noted that the same reference numerals are assigned to the same components as much as possible even though 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 only for distinguishing the components from other components, and the nature, order or order of the components are not limited by the terms. If a component is described as being "connected", "coupled" or "connected" to another component, that component may be directly connected or connected to that other component, but between components It will be understood that may be "connected", "coupled" or "connected".

6 is a diagram schematically illustrating an apparatus for directional intra prediction according to an embodiment of the present invention. Referring to the drawings, the directional intra prediction apparatus 600 includes a pixel column value generator 610, a pixel row value generator 620, a total pixel value predictor 630, a reference matrix generator 640, and a prediction block pixel. A value predictor 650 is provided.

The pixel column value generator 610 predicts the sum of pixel values of the N columns by using prediction mode 0 or prediction mode 1 from adjacent pixels of the N × N block to be encoded, and generates an N × 1 matrix.

For example, when the block to be encoded is a 4 × 4 block as shown in FIG. 3, the pixel column value generator 610 may include pixels A, B, C, D, E, adjacent to the 4 × 4 block. Predict the sum of pixel values of four columns using prediction mode 0 or prediction mode 1 from F, G, H, I, J, K, L, and M and generate a 4 × 1 matrix. Here, the pixel column value generator 610 predicts the sum of pixel values of four columns by using prediction mode 0 from pixels A, B, C, and D adjacent to the top of the 4 x 4 block, and generates a 4 x 1 matrix. It is desirable to. In this case, when the resulting N x 1 matrix column values by the pixel generator 610 as C _sum, C _sum can be expressed as Equation (2).

The pixel row value generator 620 predicts the sum of pixel values of the N rows by using prediction mode 0 or prediction mode 1 from adjacent pixels of the N × N block to be encoded, and generates an N × 1 matrix.

For example, when the block to be encoded is a 4 × 4 block as shown in FIG. 3, the pixel row value generator 620 may include pixels A, B, C, D, E, adjacent to the 4 × 4 block. Predict the sum of pixel values in four rows using prediction mode 0 or prediction mode 1 from F, G, H, I, J, K, L, and M and generate a 4 × 1 matrix. Here, the pixel row value generator 620 predicts the sum of pixel values of four rows from the pixels I, J, K, and L adjacent to the left of the 4 × 4 block by using prediction mode 1, and generates a 4 × 1 matrix. It is desirable to produce. In this case, assuming that the N x 1 matrix generated by the pixel row value generator 620 is R _sum , R _sum may be expressed by Equation 3 below.

The overall pixel value predictor 630 predicts the sum of all pixel values using the prediction mode 0 or the prediction mode 1 from adjacent pixels of the N × N block. In this case, the sum of all pixel values predicted by the all pixel value predictor 630 may vary depending on whether the mode to be executed is the prediction mode 0 or the prediction mode 1. When the sum of all pixel values predicted by the all pixel value predictor 630 is B _sum , B _sum may be expressed as in Equation 4.

The reference matrix generator 640 may include an N × 1 matrix C _sum generated by the pixel column value generator 610, an N × 1 matrix R _sum generated by the cell row value generator 620, and an overall pixel value predictor ( A reference matrix Q is generated by using the sum B _sum predicted by 630. In this case, the reference matrix Q may be expressed as Equation 5 based on a value obtained through Equations 2 to 4.

Since the matrix C _sum and the matrix R _sum are each N x 1 vectors, the reference matrix Q is [(2N + 1) x 1] together with the total pixel value sum B _sum predicted by the total pixel value predictor 630. It can be represented as a vector. For example, when the block to be encoded is a 4 x 4 block as shown in FIG. 3, a 4 x 1 matrix is generated by the pixel column value generator 610 and 4 x by the pixel row value generator 620. Since one matrix is generated, the reference matrix generator 640 may be displayed as a 9 × 1 vector together with the total pixel value sum predicted by the total pixel value predictor 630.

On the other hand, it is assumed that the reference matrix Q has a relationship of Q = PI with respect to the transformation matrix P. Here, I represents each pixel in the N × N block as an N ² × 1 matrix, and represents each pixel value of the prediction block to be encoded. For example, when the block to be encoded is 4 × 4, each pixel value in the prediction block may be represented by a matrix I of 16 × 1 vectors. In this case, in order to satisfy Q = PI, the transformation matrix P may be defined as Equation 6 as a 9 × 16 vector.

Here, the transformation matrix P is not a square matrix, but because the number of columns is larger than the number of rows, it can be regarded as an underdetermined system having an infinite solution. Therefore, we must find the solution with the smallest error. To this end, as in Equation 7,

Differentiate with respect to I and set the result to 0.

At this time, multiplying P by both sides gives Equation (9).

Substituting Q = PI yields the following equation (10).

Therefore, equation (11) can be derived from equation (10).

From Equation 11, the prediction block I can be obtained as shown in Equation 12.

As such, the 4 × 4 block may be predicted through Equation 12. Here, P ^T ( PP ^T ) ⁻¹ is called pseudo-inverse. In the actual experiment, a fixed pseudo-inverse matrix as shown in FIG. 7 was used to reduce the amount of computation.

FIG. 8 is a flowchart illustrating a directional intra prediction method by the directional intra prediction apparatus of FIG. 6.

Referring to the drawings, since the directional intra prediction method determines a prediction block using pixels adjacent to a block to be encoded, the prediction mode is mode 0 (Vertical prediction) for more accurate intra prediction. Mode) or mode 1 (Horizontal prediction mode), it is preferably applied. That is, it is determined whether the prediction mode to be executed is the vertical prediction mode or the horizontal prediction mode (S801), and when the prediction mode is the vertical prediction mode or the horizontal prediction mode, the directional intra prediction method according to the embodiment of the present invention is applied. Otherwise, it is preferable to perform intra prediction according to the general H.264 standard (S803).

In addition, since the directional intra prediction method according to the embodiment of the present invention uses pixels adjacent to the block to be encoded, it is applied only when the variance σ of the adjacent pixels is larger than the minimum threshold th. (S805) Otherwise, since the desired result is not obtained, it is preferable to perform intra prediction according to the general H.264 standard (S803). Here, a comparison value of the variance of adjacent pixels, that is, a threshold value, follows a preset value.

In the following description, a case where a block to be encoded is 4 x 4 will be described as an example to facilitate explanation. However, this does not mean that the present invention is limited to determining 4 × 4 prediction blocks.

If the prediction mode is the vertical prediction mode or the horizontal prediction mode, and the variance of the pixels adjacent to the block to be encoded is greater than a preset threshold, the pixel string value generator 610 may determine the upper neighbor pixels of the 4 × 4 block to be encoded. Predict the sum of pixel values of four columns using prediction mode 0 from A, B, C and D and generate a 4 x 1 matrix. In this case, the 4 x 1 matrix C _sum generated by the pixel column value generator 610 may be expressed as Equation 2 below.

In addition, the pixel row value generator 620 predicts the sum of pixel values of four rows from the pixels I, J, K, and L adjacent to the left side of the 4 × 4 block to be encoded using the prediction mode 1, and estimates 4 ×. Create a matrix of 1. In this case, the 4 × 1 matrix R _sum generated by the pixel row value generator 620 may be expressed by Equation 3 below.

As described above, in the embodiment of the present invention, when the prediction mode is the vertical prediction mode or the horizontal prediction mode, the upper neighboring pixels and the left neighboring pixels of the 4 × 4 block to be encoded are different from the general vertical prediction mode or the horizontal prediction mode. Consider them at the same time.

In addition, the overall pixel value predictor 630 predicts the sum of all pixel values using the prediction mode 0 or the prediction mode 1 from adjacent pixels of the 4 × 4 block to be encoded (S707). In this case, the sum of the total pixel values predicted by the total pixel value predictor 630 may vary as shown in Equation 4 according to whether the execution mode is the prediction mode 0 or the prediction mode 1.

The reference matrix generator 640 is a 4 x 1 matrix C _sum generated by the pixel column value generator 610, a 4 x 1 matrix R _sum generated by the cell row value generator 620, and an overall pixel value predictor ( A reference matrix Q is generated by using the sum B _sum predicted by the step 630 (S709). In this case, the reference matrix Q may be expressed as Equation 5. In this case, Q is a 9 x 1 matrix.

On the other hand, it is assumed that the reference matrix Q has a relationship of Q = PI with respect to the transformation matrix P. Here, I is a 16 x 1 matrix consisting of the pixel values of each of the 4 x 4 prediction blocks. In this case, in order to satisfy Q = PI, the transformation matrix P may be defined as Equation 6 as a 9 × 16 vector.

In order to find a solution for I representing a 4 x 4 prediction block, the calculation is performed as in Equations 7 to 11, whereby I becomes as in Equation 12.

In this manner, when the mode is the prediction mode 0 or the prediction mode 1, the prediction block can be predicted more accurately than the prediction based on general H.264.

While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented in one independent hardware, each or all of the components may be selectively combined to perform some or all functions combined in one or a plurality of hardware. It may be implemented as a computer program having a. 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.

In addition, the terms "comprise", "comprise" or "having" described above mean that the corresponding component may be inherent unless specifically stated otherwise, and thus excludes other components. It should be construed that it may further include other components instead. All terms, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art unless otherwise defined. Terms used generally, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be interpreted in an ideal or excessively formal sense unless explicitly defined in the present invention.

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 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, the embodiment of the present invention is applied to the field of video encoding and decoding, and improves the accuracy of intra prediction while reducing the bit rate without significantly increasing the overhead of the bitstream generator compared to the H.264 standard. It is a very useful invention that produces a possible effect.

1 is a diagram illustrating a general inter prediction.

2 is a diagram illustrating directionality of the intra prediction mode.

FIG. 3 is a diagram illustrating an example of labeling for explaining an intra prediction mode of FIG. 2.

FIG. 4 is a diagram illustrating each of the intra prediction modes of FIG. 2.

FIG. 5A is a diagram illustrating the prediction mode 0 of the intra prediction modes of FIG. 2, and FIG. 5B is a diagram illustrating the prediction mode 1 of the intra prediction modes of FIG. 2. to be.

6 is a diagram schematically illustrating an apparatus for directional intra prediction according to an embodiment of the present invention.

7 is a diagram illustrating an example of a pseudo-inverse matrix used in an experiment.

FIG. 8 is a flowchart illustrating a directional intra prediction method by the directional intra prediction apparatus of FIG. 6.

<Description of Symbols for Main Parts of Drawings>

600: directional intra prediction device 610: pixel column value generator

620: pixel row value generator 630: total pixel value predictor

640: reference matrix generator 650: prediction block pixel value predictor

Claims (12)

In the directional intra prediction apparatus, Predict the sum of pixel values of N columns using prediction mode 0 or prediction mode 1 from adjacent pixels of the N × N block to be encoded, and generate an N × 1 matrix using the predicted sum of pixel values of the N columns. A pixel column value generator; Predict the sum of pixel values of N rows using prediction mode 0 or prediction mode 1 from adjacent pixels of the N × N block, and generate an N × 1 matrix using the predicted sum of pixel values of the N rows. A pixel row value generator; A total pixel value predictor for predicting a sum of all pixel values using prediction mode 0 or prediction mode 1 from adjacent pixels of the N × N block; A reference matrix Q of [(2N + 1) x 1] using each matrix generated by the pixel column value generator and the pixel row value generator, and the sum of all pixel values predicted by the total pixel value predictor A reference matrix generating unit generating a; And A pixel value predictor for predicting each pixel value of the N × N prediction block by using the reference matrix Q generated by the reference matrix generator and a preset transformation matrix P Directional intra prediction device, wherein N is a natural number. The method of claim 1, The pixel column value generator generates an N × 1 matrix representing a sum of pixel values of N columns from N adjacent pixels adjacent to the N × N block using the prediction mode 0. Prediction device. The method of claim 1, The pixel row value generator generates an N × 1 matrix representing a sum of pixel values of N rows from N adjacent pixels adjacent to the left side of the N × N block using the prediction mode 1. Intra prediction device. The method of claim 1, When N = 4, the transformation matrix P is

Directional intra prediction device, characterized in that. In the directional intra prediction method, Predict a matrix C _sum representing the sum of pixel values of four columns from pixels adjacent to the upper side of the 4x4 block to be encoded, and pixel values of four rows from the pixels adjacent to the left of the 4x4 block. Predicting a matrix R _sum that represents the sum of and predicting a sum of all pixel values in the prediction block; Generating a reference matrix Q of 9 × 1 using each value predicted by the prediction step; And Predicting each pixel value of the prediction block using the reference matrix Q and a preset transformation matrix P of 9 × 16. Directional intra prediction method comprising a. The method of claim 5, Determining whether the mode being executed is prediction mode 0 or prediction mode 1 More, And when the mode executed by the determining step is determined to be prediction mode 0 or prediction mode 1, performing the prediction step. The method of claim 5, Comparing the variance of the pixels adjacent to the 4x4 block with a set threshold More, And the prediction step is performed when the variance of the pixels adjacent to the 4x4 block is greater than the set threshold value by the comparing step. The method of claim 5, The matrix C _sum is predicted as follows.

Here, A, B, C and D represent pixels adjacent to the top of the 4 x 4 block. The method of claim 5, The matrix R _sum is predicted as follows.

Where I, J, K and L represent pixels adjacent to the left of the 4 × 4 block. The method of claim 5, The sum of the total pixel values is predicted as follows.

Here, B _sum represents the _sum of all pixel values. The method of claim 5, Wherein the reference matrix Q is assumed to have a relationship with the transform matrix P as follows: Q = PI here,

Is, I represents the 4 x 4 prediction block in a 16 x 1 matrix. A computer-executable recording medium storing a program for executing the directional intra prediction method according to any one of claims 5 to 11.

Images