Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberCN101385347 A
Publication typeApplication
Application numberCN 200780005074
PCT numberPCT/KR2007/001042
Publication date11 Mar 2009
Filing date2 Mar 2007
Priority date3 Mar 2006
Also published asCN101385347B, EP1992171A1, EP1992171A4, EP1992171B1, US8165195, US20070206872, WO2007100221A1
Publication number200780005074.2, CN 101385347 A, CN 101385347A, CN 200780005074, CN-A-101385347, CN101385347 A, CN101385347A, CN200780005074, CN200780005074.2, PCT/2007/1042, PCT/KR/2007/001042, PCT/KR/2007/01042, PCT/KR/7/001042, PCT/KR/7/01042, PCT/KR2007/001042, PCT/KR2007/01042, PCT/KR2007001042, PCT/KR200701042, PCT/KR7/001042, PCT/KR7/01042, PCT/KR7001042, PCT/KR701042
Inventors宋秉哲
Applicant三星电子株式会社
Export CitationBiBTeX, EndNote, RefMan
External Links: SIPO, Espacenet
Method of and apparatus for video intraprediction encoding/decoding
CN 101385347 A
Abstract
Provided is a method of and apparatus for video intraprediction encoding/decoding. The method of video intraprediction encoding includes dividing pixels of an input block into a plurality of groups, performing intraprediction encoding on pixels of a first group among the plurality of groups using pixels of a previously processed neighbor block, reconstructing the intraprediction-encoded pixels of the first group, determining a directivity of each pixel of the remaining groups using the reconstructed pixels of the first group and reconstructed pixels of previously processed other groups, and performing predictive encoding on the pixels of the remaining groups using the reconstructed pixels of the first group and the reconstructed pixels of the previously processed other groups according to the determined directivities and sequentially processing the pixels of the remaining groups in units of a group according to a predetermined order.
Claims(22)  translated from Chinese
1. 一种视频帧内预测编码方法,该方法包括:将输入块的像素划分为多个组;利用先前处理的邻近块的像素,对该多个组中的第一组的像素进行帧内预测编码;重构第一组的帧内预测编码的像素;利用第一组的重构的像素、以及先前处理的其他组的重构的像素,确定剩余组的像素的方向性;以及根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,对剩余组的像素进行预测编码,并且根据预定的顺序以组为单位依次处理剩余组的像素。 An intra-prediction video coding, the method comprising: dividing an input block of pixels into a plurality of groups; the use of blocks of pixels neighboring previously processed, the plurality of groups of pixels in the first group of intra- predictive coding; reconstructed intra prediction encoding of the first group of pixels; pixel reconstruction reconstructed using the first group, and other groups previously processed pixels, to determine the directionality of the remaining groups of pixels; and in accordance with the determine the direction, the use of reconstructed pixels of the first group, as well as the reconstruction of previously processed pixels of other groups, the remaining groups of pixels predictive coding, and according to a predetermined sequence in group units sequentially processing the remaining groups pixels.
2. 如权利要求l所述的方法,其中所述确定方向性包括: 设置相对于剩余组的每个像素的多个方向,并且计算根据该多个方向的每一个的、剩余组的每个像素的邻近像素的绝对值差的平均值;以及比较才艮据该多个方向的、所计算的平均值,以确定该多个方向中具有所计算的平均值中的最小平均值的方向。 2. The method of claim l, wherein said determining directional comprising: setting a plurality of directions with respect to each of the remaining groups of pixels, and calculates based on the plurality of directions, each of the remaining groups each the absolute value of the difference between the average value of the pixel neighboring pixels; and comparing it to the plurality of directions, according to Gen, the calculated average value to determine the average value of the plurality of directions having the calculated average value in the direction of the minimum.
3. 如权利要求l所述的方法,其中所述进行预测编码包括: 在第一组的像素以及先前处理的其他组的像素中,确定参考像素;以及利用所确定的参考像素的平均值,预测剩余组的像素。 3. The method of claim l, wherein said prediction coding comprising: a first group of pixels in the pixel group, and other previously processed, it is determined reference pixels; and an average value with the determined reference pixels, Prediction of the remaining groups of pixels.
4. 如权利要求l所述的方法,其中所述输入块尺寸为MxN,其中M-2m, N=2n,并且m与n为正整数,并且将该输入块划分为:第一组,包括输入块像素中的偶数行中偶数号像素;第二组,包括位于第一组的像素的左侧与右侧的像素;第三组,包括位于第一組的像素之上与之下的像素;以及第四组, 包括没有包含在第一组至第三组中的剩余像素。 The first group, comprising: l The method according to claim 4, wherein the input block size is MxN, where M-2m, N = 2n, and m and n are positive integers, and the input block is divided into input block of pixels in the even-numbered even-numbered pixel rows; the second group, the left and right pixel comprises a first group of pixels; the third group including the pixel located above the first group of pixels and below ; and a fourth group including not included in the first to third sets of the remaining pixels.
5. 如权利要求4所述的方法,其中所述剩余组的像素按照第二组、然后第三组、然后第四组的顺序处理。 5. The method of claim 4, wherein the remaining groups of pixels according to the second group and the third group, and the fourth group of sequential processing.
6. 如权利要求4所述的方法,其中所述剩余组的像素按照第三组、然后第二组、然后第四组的顺序处理。 6. The method of claim 4, wherein the remaining groups of pixels according to the third group, then the second group, and the fourth group of sequential processing.
7. 如权利要求1所述的方法,还包括:将包括在帧内预测编码中使用的方向性信息的帧内预测模式信息添加到比特流的头部。 7. The method of claim 1, further comprising: adding the intra prediction mode information includes directional information used in the intra prediction coding to the bit stream header.
8. —种视频帧内预测编码装置,该装置包括: 块划分单元,其将输入块的像素划分为多个组; 方向确定单元,其确定该多个组的^^素的方向性;以及帧内预测单元,其利用先前处理的邻近块的像素,对该多个组中的第一组的像素进行帧内预测编码,根据预定的顺序以组为单位依次处理剩余组的像素,并且根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,对剩余组的像素进行预测编码。 8. - kind of a video frame predictive coding apparatus comprising: a block division unit, its input pixel block is divided into a plurality of groups; direction determination unit that determines the plurality of groups of ^^ prime directivity; and intra-prediction unit, a pixel adjacent to the block that utilizes previously processed, the plurality of groups of pixels in the first group of intra prediction encoding according to a predetermined sequence to set the remaining groups of pixels are sequentially processed as a unit, and in accordance with the determined directivity using reconstructed pixels of the first group, and the reconstruction of previously processed pixels of other groups, the remaining groups of pixel prediction coding.
9. 如权利要求8所述的装置,其中所述方向确定单元设置相对于第一像素的多个方向,并且计算根据该多个方向的每一个的、第一像素的邻近像素的绝对值差的平均值,并且比较根据该多个方向的、所计算的平均值,以确定该多个方向中具有所计算的平均值中的最小平均值的方向。 9. The apparatus of claim 8, wherein said direction determining means provided with respect to the direction of the first plurality of pixels, and calculates the absolute value of each of the first plurality of pixels adjacent to the pixel of the difference in the direction of the the average value, and based on the comparison of a plurality of directions, the calculated average value to determine the average value of the plurality of directions having calculated the direction of the minimum average.
10. 如权利要求8所述的装置,其中所述帧内预测单元在第一组的像素以及先前处理的其他组的像素中确定参考像素,并且利用所确定的参考像素的平均值,预测剩余组的像素。 10. The apparatus of claim 8, wherein the intra-prediction unit determines reference pixels of the first pixel group and the other groups of pixels previously processed, and use the average of the determined reference pixels of the prediction residual group of pixels.
11. 如权利要求8所述的装置,其中所述输入块尺寸为MxN,其中M-2m, N=2n,并且m与n为正整^:,并且将该输入块划分为:第一组,包括输入块像素中的偶数行中偶数号像素;第二组,包括位于第一组的像素的左侧与右侧的像素;第三组,包括位于第一组的像素之上与之下的像素;以及第四组, 包括没有包含在第一组至第三组中的剩余像素。 11. The apparatus of claim 8, wherein the input block size is MxN, where M-2m, N = 2n, and m is a positive integer and n ^ :, and the input block is divided into: a first group , including input block of pixels in even-numbered even-numbered rows of pixels; the second group, the left and right pixel comprises a first group of pixels; the third group including top located below the first set of pixels pixels; and a fourth group including not included in the first to third sets of the remaining pixels.
12. 如权利要求11所述的装置,其中所述剩余组的像素按照第二组、然后第三组、然后第四组的顺序处理。 12. The apparatus of claim 11, wherein the remaining groups of pixels according to the second group and the third group, and the fourth group of sequential processing.
13. 如权利要求11所述的装置,其中所述剩余组的像素按照第三组、然后第二组、然后第四组的顺序处理。 13. The apparatus of claim 11, wherein the remaining groups of pixels according to the third group, then the second group, and the fourth group of sequential processing.
14. 如权利要求8所述的装置,其中所述帧内预测单元将包括在帧内预测编码中使用的方向性信息的帧内预测模式信息添加到比特流的头部。 14. The apparatus of claim 8, wherein the intra prediction unit comprises adding directivity information intra prediction mode information used in the intra prediction coding to the bit stream header.
15. —种纟见频帧内预测解码方法,该方法包括:接收比特流,该比特流包含:通过划分输入块而获得的多个组中的第一组的利用邻近块的像素进行了帧内预测编码的像素的数据,以及利用第一组的重构的像素以及先前处理的其他组的重构的像素进行了帧内预测编码的像素的数据;确定剩余组的像素的方向性;以及根据所确定的方向性,利用第一组的解码的像素以及先前解码的其他组的像素,预测剩余组的像素。 15. - see Frequency seed Si intra prediction decoding, the method comprising: receiving a bitstream, the bitstream comprising: a plurality of groups obtained by dividing the input block using neighboring blocks of pixels of the first group were frame intra prediction encoded pixel data, and using the first set of reconstructed reconstructed pixel and other groups previously processed pixel intra prediction encoding pixel data; determining the remaining groups of pixels directivity; and according to the determined direction, utilizing the decoded pixels of the first group and the other groups of pixels of previously decoded prediction residual group of pixels.
16. 如权利要求15所述的方法,其中所述确定方向性包括: 计算第一像素的位于相对于第一像素预定角度方向上的邻近像素的绝对值差的平均值;以及比较根据所述预定角度方向的所计算的平均值,以确定所述预定角度方向中具有所计算的平均值中的最小平均值的方向。 16. The method of claim 15, wherein the orientation determining comprises: calculating a first pixel is located with respect to the absolute value of the difference between the average value of adjacent pixels of the first pixel on a predetermined angular direction; and in accordance with the comparison the average value of the calculated predetermined angular direction to determine the angular direction in the predetermined direction with the calculated average value of the minimum average.
17. 如权利要求15所述的方法,其中所述预测剩余组的像素包括: 在第一组的像素以及先前处理的其他组的像素中,确定参考像素;以及利用所确定的参考像素的平均值,预测剩余组的像素。 17. The method of claim 15, wherein the prediction of the remaining groups of pixels comprising: a first group of pixels and other groups of pixels previously treated, determines reference pixels; and using the determined reference average pixel values, the prediction of the remaining groups of pixels.
18. —种视频帧内预测解码装置,该装置包括:帧内预测模式确定单元,其接收比特流,该比特流包含:通过划分输入块而获得的多个组中的第一组的利用邻近块的像素进行了帧内预测编码的像素的数据,以及利用第一组的重构的像素以及先前处理的其他组的重构的像素进行了帧内预测编码的像素的数据;并且利用在收到的比特流中包含的帧内预测模式信息,确定帧内预测模式;以及帧内预测单元,其对第一组的像素进行帧内预测解码,确定剩余组的像素的方向性,并且根据所确定的方向性,利用第一组的解码的像素以及先前处理的其他组的解码的像素,预测剩余组的像素。 18. - Species intra prediction video decoding apparatus comprising: an intra prediction mode determining unit that receives a bitstream, the bitstream comprising: a plurality of groups obtained by dividing the input block using a first set of adjacent pixel block to be intra prediction encoding pixel data, and the use of reconstructed pixels of the first group and the reconstruction of the other groups previously processed pixel intra prediction encoding pixel data; and the use of the closing to the intra prediction mode information included in the bit stream, determining an intra prediction mode; and intra-prediction unit, a first group of pixels for which intra prediction decoding, determines directionality remaining groups of pixels, and in accordance with the determine the direction, use the decoded pixels of the first group and the decoding of the other groups previously processed pixels, the pixels of the remaining groups predicted.
19. 如权利要求18所述的装置,其中所述帧内预测单元计算第一像素的位于相对于第一像素预定角度方向上的邻近像素的绝对值差的平均值,并且比较根据所述预定角度方向的所计算的平均值,以确定所述预定角度方向中具有所计算的平均值中的最小平均值的方向。 19. The apparatus of claim 18, wherein the intra-prediction unit calculates a first pixel is located with respect to the absolute value of the difference between the average value of adjacent pixels of the first pixel on a predetermined angular direction, and based on the predetermined comparison The calculated average angular direction to determine the angular direction in the predetermined direction with the calculated average value of the minimum average.
20. 如权利要求18所述的装置,其中所述帧内预测单元在第一组的像素以及先前处理的其他组的像素中,确定参考像素,并且利用所确定的参考像素的平均值,预测剩余组的^f象素。 20. The apparatus of claim 18, wherein the intra-prediction unit in the pixels of the first group of pixels and other groups previously treated, determines reference pixels, and use the average of the determined reference pixels of the prediction The remaining group of ^ f pixels.
21. —种视频帧内预测编码方法,该方法包括:对于预定色彩空间中的每个颜色分量,将输入块的像素划分为多个组; 利用先前处理的邻近块的像素,预测该多个组中的第一组的像素; 生成第一组的预测的像素的余量,对所生成的余量进行色彩变换,并且编码该色彩变换的余量; 重构第一组的编码的像素;利用第一组的重构的像素、以及先前处理的其他组的重构的l象素,确定剩余组的像素的方向性;以及根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,预测剩余组的像素;以及生成剩余组的预测的像素的余量,对所生成的余量进行色彩变换,并且编码该色彩变换的余量,其中根据预定顺序以组为单位依次处理剩余组的像素。 21. - Species intra prediction video coding, the method comprising: a predetermined color space for each color component of the input pixel block is divided into a plurality of groups; the use of blocks of pixels neighboring previously processed, the plurality of prediction group in the first group of pixels; generating a first set of prediction margin pixels of the generated margin for color conversion, and the balance of the color transformation coding; reconstructed coded first group of pixels; reconstructed using the first set of pixels, and the reconstruction of previously processed l pixel other groups, the remaining groups of pixels determined directivity; and based on the determined direction, with the first set of reconstructed pixels , and the reconstructed pixel previously processed other groups, the remaining groups of predicted pixels; and generating prediction margin remaining groups of pixels, of the generated margin for color conversion, and the balance of the color transformation coding, wherein the remaining groups of pixels are sequentially processed in group units in accordance with a predetermined sequence.
22.如权利要求21所述的方法,其中所述预定色彩空间为RGB色彩空间。 22. The method according to claim 21, wherein the predetermined color space to RGB color space.
Description  translated from Chinese

视频帧内预测编码/解码的装置与方法 Intra-prediction video coding / decoding apparatus and method

技术领域 Technical Field

符合本发明的装置与方法涉及视频的帧内预测,更具体地,涉及视频帧内预测编码/解码,其中将当前块划分为多个组,利用当前块的先前处理的邻近块的像素,对第一组的像素进行帧内预测,并且利用相应先前组的重构的像素,对剩余组的像素进行帧内预测。 Frame apparatus and method according to the invention relates to a video prediction, and more particularly, relates to video intra prediction encoding / decoding, wherein the current block is divided into a plurality of groups, using the pixel of the previously processed neighboring blocks of the current block, of The first group of pixel intra prediction and reconstructed using the corresponding pixel of the previous group, the remaining groups of pixels intra-prediction.

背景技术 Background

才艮据诸如运动画面专家组(MPEG) -1、 MPEG-2、 MPEG-4、 H.264/MPEG-4/ 高级视频编码(AVC)等视频压缩标准,将画面划分为宏块用于视频编码。 Gen It was like Moving Picture Experts Group (MPEG) -1, MPEG-2, MPEG-4, H.264 / MPEG-4 / Advanced Video Coding (AVC) and other video compression standards, the screen is divided into macro-blocks for video coding. 在以所有帧间预测(interprediction )以及帧内预测(intraprediction )编码才莫式编码了每个宏块之后,根据编码宏块所需的比特率、以及原始宏块与解码的宏块之间的许可失真,选择适当的编码模式。 In order to all inter prediction (interprediction) and intra prediction (intraprediction) was Morse coding between encoding after each macro block, according to the coding macroblock required bit rate, and the original macroblock and the decoded macroblock License distortion, select the appropriate coding mode. 然后以所选择的编码模式编码宏块。 Then coded macroblock in the selected encoding mode.

在帧内预测时,与参照参考画面不同,使用在空间上邻近要编码的宏块的像素的值,估计要编码的宏块的预测值,并且当编码当前画面的宏块时, 编码预测值与原始像素值之间的差。 When the intra prediction with reference to a reference picture different from the values used to encode spatially adjacent pixel macroblock, the prediction value is estimated to be encoded macroblock, and when coding the current picture macroblock, encoding prediction value the difference between the original pixel values with. 帧内预测模式分为用于亮度分量的4 x4 帧内预测模式、(高简档(high profile)情况下)8x8帧内预测模式、16x16 帧内预测模式、以及用于色度分量的帧内预测模式。 Intra-prediction mode for the luminance component is divided into 4 x4 intra prediction mode, (high profile (high profile) case) 8x8 intra prediction mode, 16x16 intra prediction modes for chrominance components as well as intra prediction mode.

图1显示根据H.264标准的用于亮度分量的16x16帧内预测模式,图2 显示根据H.264标准的用于亮度分量的4x4帧内预测模式。 Figure 1 shows according to the H.264 standard for the luminance component intra 16x16 prediction mode, Figure 2 shows in accordance with the H.264 standard for the luminance component intra 4x4 prediction mode.

参照图1,有四种16x16帧内预测模式,即垂直模式、水平模式、直流(DC)模式、以及平面模式。 Referring to Figure 1, there are four 16x16 intra prediction modes, i.e. a vertical mode, horizontal mode, direct current (DC) mode, and a plane mode. 参照图1,有九种4x4帧内预测模式,即垂直模式、水平模式、DC模式、向左下对角线模式、向右下对角线模式、垂直向右模式、垂直向左模式、水平向上模式、以及水平向下模式。 Referring to Figure 1, there are nine kinds of 4x4 intra prediction modes, i.e. a vertical mode, horizontal mode, DC mode, a diagonal down left mode, a diagonal down right mode, a vertical right mode, a vertical left mode, a horizontal up mode, and a horizontal down mode.

例如,当以模式0 (即图2的垂直模式)预测编码4x4当前块时,4x4当前块之上邻近的像素A至D的像素值被预测作为4x4当前块的像素值。 For example, when the mode 0 (ie, vertical mode Fig. 2) when the current block predictive coding 4x4, 4x4 pixel adjacent above the current block A to D of the pixel values are predicted as the pixel value 4x4 current block. 换言之,像素A的像素值被预测作为4x4当前块的第一列的四个像素的像素值, In other words, the pixel value of the pixel A is predicted as four pixel values of the first column of the 4x4 current block,

像素B的像素值被预测作为4x4当前块的第二列的四个像素的像素值,像素C的像素值被预测作为4x4当前块的第三列的四个像素的像素值,像素D的像素值被预测作为4x4当前块的第四列的四个像素的像素值。 Pixel values of the pixel B is predicted as four 4x4 current block pixel values of the second column, the pixel C is predicted as four 4x4 current block pixel values of the third column, the pixels of the pixel D value is predicted as four pixel values of the fourth column of the 4x4 current block. 接着,获得并且编码使用像素A至D预测的4x4当前块的像素的像素值与在原始4x4当前块中包含的像素的实际像素值之间的差。 Then, the difference between the actual pixel value and obtain encoded using pixels A to D pixels of the 4x4 current block predicted pixel value contained in the original 4x4 current block of pixels between.

在根据H.264标准的视频编码中,以总共13种模式编码当前宏块,包括4x4帧内预测模式与16x16帧内预测模式,然后以具有最小成本的编码模式对当前宏块进行帕内预测编码。 In accordance with the H.264 video coding standard, a total of 13 kinds of modes encoding the current macroblock, including intra 4x4 prediction mode and an intra 16x16 prediction mode, and then the encoding mode having the smallest cost of the current macroblock prediction Pa coding. 这意味着以四种16x16帧内预测才莫式帧内预测当前宏块,并且选"^奪具有最小成本的一种。以九种4x4帧内预测^t式帧内预测当前宏块的每个4x4子块,并且为每个子块选择具有最小成本的一种。 比较所选的16x16帧内预测才莫式的成本以及所选的4x4帧内预测才莫式的成本的和,并且选择具有最小成本的模式。 This means that in four 16x16 intra prediction only Morse intra prediction of the current macroblock, and choose "^ wins to have a minimal cost. In nine 4x4 intra prediction ^ t every type of intra prediction of the current macroblock a 4x4 sub-blocks, and select one with the smallest cost for each sub-block. Compare selected 16x16 intra prediction only Mohs costs and selected 4x4 intra prediction only Mohs and cost, and choose a The minimum cost model.

通过这种方式,根据相关技术的帧内预测使用从要帧内预测的当前块的邻近块采样的像素,而不使用在当前块中包含的像素。 In this manner, according to the related art using intra prediction from a neighboring block to the intra prediction of the current pixel block sampling instead of pixels contained in the current block. 结果,当当前块的视频非常不同于邻近块的视频时,帧内预测的块与实际块之间的差可能较大。 As a result, when the video of the current block is very different from neighboring blocks of the video, the difference between the block and the actual block between intra prediction may be larger. 因为根据相关技术的帧内预测仅使用邻近块的像素信息而不使用要帧内预测的当前块的像素信息,所以预测和编码效率可能受到限制。 Because according to the related art intra prediction using only pixel information of neighboring blocks without using the pixel information of the current block to be intra prediction, the prediction and coding efficiency may be limited.

发明内容技术方案 Technical Solution invention

本发明提供一种视频帧内预测编码/解码的装置与方法,其中使用当前块的重构像素的内插像素以及当前块的邻近块的像素,作为视频帧内预测时的预测算子,由此提高预测效率。 The present invention provides a video frame predictive coding / decoding apparatus and method, in which the pixels of the current block of reconstructed pixels interpolated pixel and the neighboring block of the current block, prediction operator as a video frame prediction by This increase prediction efficiency.

根据本发明的一方面,提供了一种视频帧内预测编码方法。 According to an aspect of the present invention, there is provided a video frame prediction coding method. 该方法包括: 将输入块的像素划分为多个组;利用先前处理的邻近块的像素,对该多个组中的第一组的像素进行帧内预测编码;重构第一组的帧内预测编码的像素; 利用第一组的重构的像素、以及先前处理的其他组的重构的像素,确定剩余组的每个像素的方向性;以及根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,对剩余组的像素进行预测编码, 并且根据预定的顺序以组为单位依次处理剩余组的像素。 The method comprising: dividing an input block of pixels into a plurality of groups; the use of blocks of pixels neighboring previously processed, the plurality of groups of pixels in a first set of intra-prediction coding; reconstructed frame of the first group predictive coding pixel; reconstructed pixels reconstructed using the first pixel group, and other groups previously processed, each pixel is determined directionality remaining groups; and based on the determined direction, with the first group reconstructed pixels, and the reconstruction of the other groups previously processed pixel, the pixel prediction coding residual group, and according to a predetermined order of the group of pixels are sequentially processed in units of the remaining groups.

根据本发明的另一方面,提供了一种视频帧内预测编码装置。 According to another aspect of the present invention, there is provided an intra prediction video coding apparatus. 该装置包 The device package

括:块划分单元、方向确定单元、以及帧内预测单元。 Comprising: a block division unit, the direction determination unit, and the intra-prediction unit. 块划分单元将输入块的像素划分为多个组。 Block division unit of the input block of pixels into a plurality of groups. 方向确定单元确定该多个组的每个像素的方向性。 Direction determination unit determines for each pixel of the plurality of groups of directionality. 帧内预测单元利用先前处理的邻近块的像素,对该多个组中的第一组的像素进行帧内预测编码,根据预定的顺序以组为单位依次处理剩余组的像素,并且根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,对剩余组的像素进行预测编码。 Adjacent pixel block intra prediction unit using the previously processed, the plurality of groups of pixels in the first group of intra prediction encoding according to a predetermined order sequentially processed in units of a pixel group of the remaining groups, and determined in accordance with direction, using the first set of reconstructed reconstructed pixel, as well as other groups previously processed pixel, the remaining groups of pixel prediction coding.

根据本发明的另一方面,提供了一种视频帧内预测解码方法。 According to another aspect of the present invention, there is provided a video decoding method of intra prediction. 该方法包括:接收比特流,该比特流包含:通过划分输入块而获得的多个组中的第一组的利用邻近块的像素进行了帧内预测编码的像素的数据,以及利用第一组的重构的像素以及先前处理的其他组的重构的像素进行了帧内预测编码的像素的数据;确定剩余组的每个像素的方向性;以及根据所确定的方向性,利用第一组的解码的像素以及先前解码的其他组的像素,预测剩余组的像素。 The method comprises: receiving a bitstream, the bitstream comprising: a plurality of groups obtained by dividing an input block using neighboring blocks in the first group of pixels were intra prediction encoding pixel data, and the use of a first set of It reconstructed reconstructed pixel and the other groups previously processed pixel intra prediction encoding pixel data; determining for each pixel of the remaining groups directivity; and based on the determined direction, with the first group The decoding of the pixels and a pixel previously decoded other groups, predicted pixels of the remaining groups.

根据本发明的另一方面,提供了一种视频帧内预测解码装置。 According to another aspect of the present invention, there is provided an intra-prediction video decoding apparatus. 该装置包括:帧内预测模式确定单元与帧内预测单元。 The apparatus comprises: an intra prediction unit and the intra prediction mode determining unit. 帧内预测模式确定单元接收比特流,该比特流包含:通过划分输入块而获得的多个组中的第一组的利用邻近块的像素进行了帧内预测编码的像素的数据,以及利用第一组的重构的像素以及先前处理的其他组的重构的像素进行了帧内预测编码的像素的数据; 并且利用在收到的比特流中包含的帧内预测模式信息,确定帧内预测模式。 Intra-prediction mode determining unit receives a bitstream, the bitstream comprising: a plurality of groups obtained by dividing the input block using neighboring blocks of pixels of the first group were intra prediction encoding pixel data, and the use of the first Reconstruction of a set of pixels and the reconstructed previous processing of other pixels of the group of intra prediction encoding pixel data; and the use of the intra prediction mode information included in the received bit stream, determining an intra prediction mode. 帧内预测单元对第一组的像素进行帧内预测解码,确定剩余组的每个像素的方向性,并且根据所确定的方向性,利用第一组的解码的像素以及先前处理的其他组的解码的像素,预测剩余组的像素。 Intra-prediction unit pixels of the first group of intra prediction decoding, the directivity of each pixel to determine the remaining groups, and according to the determined direction, using the first set of decoded pixels and other groups of the previously processed decoded pixel prediction of the remaining groups of pixels.

技术效果 Technical effect

根据本发明的示范性实施例,当前块的像素之间的相关性以及当前块的邻近块用于^L频帧内预测,由此提高了预测效率和编码效率。 According to an exemplary embodiment of the present invention, the correlation between the current block and neighboring blocks of pixels between the current block for intra prediction ^ L frequency, thereby improving the prediction efficiency and coding efficiency.

附图说明 Brief Description

图1显示根据H.264标准的用于亮度分量的16x16帧内预测模式; 图2显示根据H.264标准的用于亮度分量的4x4帧内预测模式; 图3为使用根据本发明示范性实施例的视频帧内预测编码装置的视频编码器的方框图; Figure 1 shows according to the H.264 standard for the luminance component intra 16x16 prediction mode; Figure 2 shows in accordance with the H.264 standard for the luminance component intra 4x4 prediction mode; 3 is used in accordance with an exemplary embodiment of the present invention, FIG. intra prediction video block diagram showing an example of a video encoder of the coding apparatus;

图4为才艮据本发明示范性实施例的图3的帧内预测单元的方框图; 图5显示被图4的块划分单元划分的输入块; 图6显示根据本发明示范性实施例的帧中块的处理顺序; 图7显示根据本发明示范性实施例的被图4的块划分单元划分的第一组的像素; Figure 4 is a block diagram of data before Gen exemplary embodiment of the present invention, the intra-prediction unit 3; FIG. 5 shows a block division unit divided input block of FIG. 4; FIG. 6 shows a frame according to an exemplary embodiment of the present invention, Processing sequence block; Figure 7 shows the division unit divided according to the first set of pixel block is diagram of an exemplary embodiment of the present invention, 4;

图8为用来解释才艮据本发明示范性实施例的相对于第一组的像素的帧内预测的视图; 8 is used to explain the data before Gen embodiment of the present invention with an exemplary embodiment of the first group of pixels of the intra prediction view;

图9显示根据本发明示范性实施例的相对于第二组的像素的帧内预测所需的第一组的像素; Figure 9 shows a first example of a group of pixels with respect to the second group of pixels required for intra-prediction according to an exemplary embodiment of the present invention;

图IO与图IIA至图111为用来解释#4居本发明示范性实施例的围绕第二组的像素确定方向性的处理的视图; Figure IO FIG IIA to 111 for explaining # 4 pixels surrounding the second group of Habitat exemplary embodiment of the present invention to determine the directionality of view of a process;

图12显示根据本发明示范性实施例的相对于第三组的像素的帧内预测所需的第二组的像素; Figure 12 shows a second example of the group of pixels with respect to the third group of pixels required for intra prediction according to an exemplary embodiment of the present invention;

图13与图14A至图141为用来解释根据本发明示范性实施例的围绕第三组的像素确定方向性的处理的视图; 13 and FIG. 14A to FIG. 141 is a view for explaining the process to determine the directionality of the third group of surrounding pixels according to an exemplary embodiment of the present invention;

图15显示根据本发明示范性实施例的相对于第四组的像素的帧内预测所需的第一组至第三组的重构的像素; Figure 15 shows an example with respect to the fourth group of pixels in a first desired intra prediction to third sets of reconstructed pixel according to an exemplary embodiment of the present invention;

图16显示根据本发明另一示范性实施例的相对于第三组的像素的帧内预测所需的第一组的像素; Figure 16 shows a first example of a group of pixels with respect to the third group of pixels required for intra prediction in accordance with another exemplary embodiment of the present invention;

图17显示根据本发明另一示范性实施例的相对于第二组的像素的帧内预测所需的第一组与第三组的像素; 17 shows a first group of pixels and the third group with respect to an example of the second group of pixels required intra prediction according to another exemplary embodiment of the present invention;

图18显示根据本发明另一示范性实施例的相对于第四组的像素的帧内预测所需的第一组至第三组的像素; Figure 18 shows the first group to the third group of pixels with respect to the fourth embodiment of the group of pixels required for intra-prediction in accordance with another exemplary embodiment of the present invention;

图; Figure;

图20为显示根据本发明另一示范性实施例的视频帧内预测编码方法的流程图; Figure 20 is a flowchart showing a method of encoding a video frame in accordance with another exemplary embodiment of the present invention, the prediction;

图21为使用根据本发明示范性实施例的视频帧内预测解码装置的视频解码器的方框图; Figure 21 is a block diagram of a video decoder prediction decoding apparatus according to the present invention, a video frame to an exemplary embodiment of the use;

图22为才艮据本发明示范性实施例的图21的帧内预测单元的方框图; 22 is a block diagram according to FIG Burgundy only an exemplary embodiment of the invention the intra-prediction unit 21;

图。 FIG.

具体实施方式 DETAILED DESCRIPTION

图3为使用根据本发明示范性实施例的视频帧内预测编码装置的视频编码器300的方框图。 Figure 3 is the use of a video frame according to an exemplary embodiment of the present invention, a block diagram of a video encoder prediction encoding apparatus 300. 在以下描述中,为了解释方便,根据本发明示范性实施例的视频帧内预测编码装置用于H.264视频编码器。 Video intraframe prediction encoding apparatus in the following description, for convenience of explanation, in accordance with an exemplary embodiment of the present invention for H.264 video encoder. 但是,根据本发明示范性实施例的^f见频帧内预测编码装置也可以用于使用帧内预测的其他压缩方法。 However, according to an exemplary embodiment of the present invention, the frequency ^ f see intraframe prediction encoding apparatus may also be used for other compression methods using intra prediction.

参照图3,视频编码器300包括:运动估计单元302、运动补偿单元304、 帧内预测单元330、变换单元308、量化单元310、重排单元312、熵编码单元314、逆量化单元316、逆变换单元318、滤波器320、帧存储器3"、以及控制单元325。帧内预测单元330为根据本发明示范性实施例的视频帧内预测编码装置。 Referring to Figure 3, the video encoder 300 comprises: a motion estimation unit 302, the motion compensation unit 304, intra prediction unit 330, conversion unit 308, a quantization unit 310, a rearrangement unit 312, the entropy encoding unit 314, an inverse quantization unit 316, an inverse conversion unit 318, a filter 320, a frame memory 3 ", and a control unit 325. The intra prediction unit 330 as a video frame prediction encoding apparatus according to an exemplary embodiment of the present invention.

运动估计单元302与运动补偿单元304在参考画面中搜索当前画面的宏块的预测值进行帧间预测。 Motion estimation unit 302 and the motion compensation unit 304 searches the prediction value of the current picture prediction interframe macroblock in the reference picture.

帧内预测单元330在当前画面中搜索当前块的预测值进行帧内预测。 Intra-prediction unit 330 searches the current block in the current prediction value intra picture prediction. 具体地,根据本发明示范性实施例的帧内预测单元330接收要预测的当前块, 并且以图1与图2所示的16 x16帧内预测模式、4x4帧内预测模式、或者8x8 帧内预测模式、以及色度帧内预测模式进行帧内预测。 Intra-prediction unit In particular, according to an exemplary embodiment of the present invention 330 receives a current block to be predicted, and as shown in Figures 1 and 2 of 16 x16 intra prediction mode, 4x4 intra prediction mode, intra 8x8 or prediction mode, intra prediction mode and chroma intra prediction. 帧内预测单元330还将当前块划分为多个组,与在相关技术中一样利用先前处理的邻近块对从该多个组中选择的一个组进行帧内预测,并且利用相应先前组的重构的像素, 对剩余的组进行预测。 Intra-prediction unit 330 will be divided into a plurality of groups of the current block, and the same in the related art utilizing previously processed neighboring blocks of a group selected from the group of the plurality of intra prediction, and the use of the previously set corresponding weight pixel structure, the remaining group forecast.

更具体地,帧内预测单元330将当前块划分为多个组,并且利用当前块的邻近块的像素,对该多个组中的第一组的像素进行帧内预测。 More specifically, the intra prediction unit 330 the current block is divided into a plurality of groups, and adjacent to the current block using the pixel blocks, the plurality of groups of pixels in a first set of intra-prediction. 帧内预测单元330利用按照围绕要预测的像素的预定方向上的根据方向性从第一组的像素中选择的参考像素以及其他先前处理的组的像素的平均值,对剩余组的像素进行帧内预测。 The use of intra-prediction unit 330 according to the direction in accordance with a predetermined direction around the pixel to be predicted on the selected group of pixels from the first reference pixels and the average pixel previously processed other groups, the remaining groups of pixels from frame in the forecast. 通过这种方式,因为要帧内预测的当前块的一部分被首先帧内预测、并且使用先前帧内预测的部分的重构的信息帧内预测其他部分, 所以当前块内的像素信息以及邻近块的像素可以用于帧内预测,由此提高预测效率。 In this way, since the first part of the intra prediction of the current block to be intra prediction, intra prediction and the use of part of the previously reconstructed intra prediction information other portions, so that the pixel information of the current block and the neighboring block within The intra prediction pixel may be used, thereby improving prediction efficiency.

控制单元325控制视频编码器300的组件,并且确定用于当前块的预测 325 controls the video encoder assembly 300, and determines the prediction for the current block control unit

模式。 Mode. 例如,控制单元325可以确定将帧间预测或者帧内预测的块与原始块 For example, the control unit 325 may determine inter prediction or intra prediction block and the original block

之间的差最小化的预测模式,作为用于当前块的预测模式。 Minimize the difference between the prediction mode as a prediction mode for the current block. 为此,控制单元 To this end, the control unit

325计算帧间预测的画面与帧内预测的画面的成本,并且确定具有最小成本的预测模式作为最终的预测模式。 325 with the calculated inter-frame prediction intra picture prediction picture cost, and determine a prediction mode having the smallest cost as the final prediction mode. 此处,可以使用各种方法进行成本计算, 例如绝对差和(SAD)成本函数、绝对变换差和(SATD)成本函数、差平方和(SSD)成本函数、平均绝对差(MAD)成本函数、以及拉格郎日成本函数。 Here, you can use various methods to calculate costs, such as absolute difference (SAD) cost function, differential and absolute transformation (SATD) cost function, the difference between the sum of squares (SSD) cost function, the mean absolute difference (MAD) cost function, and Lagrange cost function. SAD为4x4块的余量的绝对值的和。 SAD is the balance and the absolute value of 4x4 blocks. SATD为通过对4x4块的余量应用Hadamard 变换而获得的系数的绝对值的和。 And the absolute value of the coefficient of SATD through application of margin 4x4 Hadamard transform blocks obtained. SSD为4x4块预测样本的余量的平方的和。 SSD is predicted samples margin 4x4 blocks squared. MAD为4x4块预测样本的余量的绝对值的平均。 MAD is an average of 4x4 block prediction samples of the absolute value of the balance. 拉格郎日成本函数为包含比特流信息的改进成本函数。 Lagrange cost function is to improve cost function contains a bit stream of information.

一旦通过帧间预测或者帧内预测找到了当前宏块要参照的预测数据,就将其从当前宏块中减去,以生成余量。 Once inter prediction or intra prediction of the current macroblock found to be referred forecast data, it will be subtracted from the current macroblock to generate margin. 通过变换单元308,将所生成的余量变换到频率域,然后由量化单元310量化。 By converting unit 308, the generated margin transformed to the frequency domain, then quantized by the quantization unit 310. 熵编码单元314编码量化的余量, Entropy encoding unit 314 encodes the quantized margin,

以專lr出比4争流。 Lr out of contention with special flow ratio 4.

为了获得用于帧间预测的参考画面,由逆量化单元316与逆变换单元318 处理量化后的画面,由此重构当前画面。 In order to obtain a reference picture for inter prediction, inverse quantization unit 316 by the picture conversion unit 318 inverse quantization after treatment, thereby reconstruct the current frame. 所重构的当前画面由进行解块化滤波的滤波器320处理,然后存储在帧存储器322中,以用于下一画面的帧间预测。 The reconstructed current frame by a de-blocking filter processing filter 320, and then stored in the frame memory 322, for the next inter-picture prediction. 在解块化滤波之前的第一组的重构的视频数据被输入到帧内预测单元330,以用作为用于预测剩余组的像素的参考数据。 Prior to deblocking filter of a first set of reconstructed video data is input to the intra-prediction unit 330 to be used as reference data for prediction of pixels of the remaining groups.

图4为根据本发明示范性实施例的图3的帧内预测单元330的方框图, 并且图5显示被图4的块划分单元331划分的输入块。 Figure 4 is an exemplary embodiment of the present invention, FIG. 3 is a block diagram of an intra prediction unit 330 of the embodiment, and FIG. 5 shows a block division unit 331 is divided into four of the input block.

参照图4,根据本发明示范性实施例的帧内预测单元330包括:块划分单元331、帧内预测器332、以及方向确定单元333。 Referring to Figure 4, according to an exemplary embodiment of the present invention, the intra prediction unit 330 comprises: a block dividing unit 331, an intra predictor 332, and the direction determination unit 333. 帧内预测器332包括用于处理通过划分输入块而获得的多个组中的每一个组的帧内预测器。 Intra prediction processing unit 332 includes a plurality of groups obtained by dividing an input block in the intra-frame predictor for each group. 在图4 中,帧内预测器332包括第一帧内预测器332a、第二帧内预测器332b、第三帧内预测器332c、以及第四帧内预测器332d,用来处理通过划分输入块而获得的四个组。 In Figure 4, the intra predictor 332 includes a first intra prediction unit 332a, a second intra prediction unit 332b, a third intra-frame predictor 332c, and a fourth intra prediction unit 332d, for processing by dividing input blocks obtained four groups.

块划分单元331将输入块的像素划分为包含至少两个组的多个组。 Block division unit 331 of the input block of pixels into at least two groups of multiple groups. 例如, 如图5所示,块划分单元331可以将输入的16x16块的像素划分为四个组。 For example, as shown in Figure 5, 16x16 pixel block divided block division unit 331 may be input into four groups. 参照图5,第一组包括输入块像素中的偶数行偶数列中的像素,第二组包括位于第一組的像素的左侧与右侧的像素,第三组包括位于第一组的像素之上 Referring to FIG. 5, a first set of input block comprising even-numbered rows of pixels in the even-numbered columns of pixels, including the pixel in the first group of pixels of the left and right of the second group, the third group comprises a first group of pixels above

与之下的像素,而第四组包括没有包含在第一组至第三组中的剩余像素。 Pixel below, and the fourth group includes the remaining pixels not included in the first group to the third group. Category

似地,即使当输入块尺寸为MxN (M=2m, N=2n,并且m与n为正整数)而非16x16时,也可以将输入块的像素划分为组。 Similarly, even when the input block size is MxN (M = 2m, N = 2n, and m and n are positive integers) rather than 16x16, the input pixel may be divided into groups of blocks. 第一组包括才艮据如图6所示的光栅扫描处理顺序(即其中从左到右从上至下依次处理块的块处理顺序)选 (Block processing sequence that is left to right, top to bottom where the processing blocks) The first group includes only Burgundy raster scan data processing sequence shown in FIG. 6 Options

择用于相对于其他组的高效空间内插预测的像素。 Optional for efficient space relative to other groups within the predicted pixel interpolation.

图7显示根据本发明示范性实施例的被图4的块划分单元331划分的第 Figure 7 is an exemplary display according to an embodiment of the present invention, the block dividing unit 331 dividing the first 4

一组的像素,图8为用来解释根据本发明示范性实施例的相对于第一组的像素的帧内预测的视图。 A group of pixels, with FIG. 8 for explaining an example of the first group of pixels of the intra-prediction view of an exemplary embodiment of the present invention. 在图8中,、y指示输入当前块800的第x行第y列中的像素。 Figure 8 ,, y indicates a first input of the current block 800 x row y column of pixels.

第一帧内预测器332a利用当前块的邻近块的像素,首先对第一组的像素进行帧内预测。 The first intra-frame predictor 332a using the pixel blocks adjacent to the current block, the first group of pixels of the first intra-frame prediction. 此时,可以使用根据H.264标准的帧内预测或者使用邻近块的像素的其他帧内预测方法。 In this case, you can use the intra prediction based on the H.264 standard or using neighboring blocks of pixels other intra prediction methods. 例如,在根据垂直模式的帧内预测中,第一帧内预测器332a首先预测在当前块之上邻近并且位于与第一组的像素相同列中的像素V2, V4, V6, ...V14, V16的像素值为当前块的第一組的像素的像素值。 For example, in the vertical intra prediction mode, the first intra-frame predictor 332a and is positioned adjacent the first prediction of the pixel in the same column of the first pixel group V2, V4, V6 on the current block, ... V14 , V16 of the pixel values of the first group of pixels of the current block of pixel values. 换言之,预测像素V2的像素值为第一组的第二列的8个像素的像素值,预测 In other words, the prediction pixel value V2 of the first group of pixels of eight pixels of the second column of the pixel values, prediction

像素V4的像素值为第一组的第四列的8个像素的像素值,预测像素V6的像 Pixel pixel value V4 fourth column of the first group of eight pixel values, as predicted pixel V6

素值为第一组的第六列的8个像素的像素值。 Pixel value of the pixel values of the first group of eight pixels in the sixth column. 通过这种方式,使用当前块的邻近块的像素,帧内预测第一组的像素。 In this way, the pixels of the current block using neighboring blocks of the intra prediction of the first group of pixels. 在第一帧内预测器332a根据各种帧内预测模式(例如水平模式)进行了帧内预测之后,其根据在每种帧内预测模式下帧内预测的第一组与原始图像对应于第一组的部分之间的差,比较帧内预测模式的成本,以确定用于第一组的帧内预测模式。 In the first intra-frame predictor 332a according to various intra prediction modes (e.g., horizontal mode) after the intra prediction, based in each intra-prediction mode of the first group and the intra prediction image corresponding to the original the cost difference between the part of a group, comparator intra prediction mode, to determine a first set of intra-prediction mode.

为帧内预测的第一组与当前块对应于第一组的原始像素之间的差的余量由变换单元308变换,然后由量化单元310量化。 For the intra prediction of the current block corresponding to the first group and the remainder of the difference between the first set of original pixels in transformation by the transformation unit 308, and then quantized by the quantization unit 310. 量化后的第一组的余量经过逆量化单元316与逆变换单元318,并且被加到帧内预测的第一组进行重构。 The balance of the first group of quantized through the inverse quantization unit 316 and inverse transform unit 318, and is supplied to a first set of intra-prediction reconstructed. 重构的第一组被存储在帧存储器322中。 Reconstruction of a first group is stored in the frame memory 322.

图9显示根据本发明示范性实施例的相对于第二组的像素的帧内预测所需的第一组的像素。 Figure 9 shows a first example of a group of pixels with respect to intra prediction of the second group of pixels required according to an exemplary embodiment of the present invention.

参照图9,利用先前处理的第一组的像素,预测标记有三角的第二组的像素。 Referring to FIG. 9, the first group using the previously processed pixel, the prediction pixels marked with triangular second group. 例如,可以使用在虚线910中包含的第一组的像素,预测第二组的像素911。 For example, a first group of pixels contained in the broken line 910, a second set of predicted pixels 911. 此时,为了确定用于预测第二组的像素的第一组的像素的参考像素, 方向确定单元333确定围绕第二组的像素的方向性。 In this case, in order to determine for the first group of pixels of the second group of pixels of the prediction reference pixel, the direction determination unit 333 determines the second group of pixels surrounding the directivity.

图IO与图IIA至图111为用来解释根据本发明示范性实施例的围绕第二组的像素确定方向性的处理的视图。 FIG IIA to FIG IO 111 is a view for explaining the process to determine the directivity of the second group of surrounding pixels according to an exemplary embodiment of the present invention. 在图10中,要预测的第二组的像素1000 由C指示,并且邻近像素1000的左侧与右侧的像素由Li和Rj指示(i,j=-4,-3,-2,-1,0, 1,2,3,4)。 In Figure 10, the second group of pixels to be predicted by the C 1000 indicates, and the left and right pixels adjacent to the pixel 1000 by Li and Rj directions (i, j = -4, -3, -2, - 1,0, 3, 4). 此处,U, L—2, L。 Here, U, L-2, L. , L2, L4, R—4, R—2, R。 , L2, L4, R-4, R-2, R. , R2, l为第一组的重构的像素。 , R2, l for the first set of reconstructed pixels. 在Ln中,如果n为奇数,则U具有Lw与L^之间的平均值。 In Ln, if n is odd, then U has an average value between Lw and L ^. 在L中,如果n为奇数,则l具有Rw与IU之间的平均值。 In L, if n is odd, l has an average value between Rw and IU.

参照图11A至图111,通过计算第二组的像素C的邻近像素的绝对值差的平均值,如下计算方向性系数D1至D9: Referring to FIG. 11A to 111, the average value calculated by the absolute value of the difference of the pixel C of the second set of adjacent pixels, is calculated as follows directivity D1 to D9:

<formula>formula see original document page 13</formula>方向确定单元333选择所计算的方向性系数D1至D9中的最小方向性系数,并且确定在对应于所选方向性系数的方向上的第一组的重构的像素为参考像素。 <Formula> formula see original document page 13 </ formula> direction determination unit 333 selects the directivity coefficient calculated in the D1 to D9 minimum directivity, and determines the direction corresponding to the selected directivity of the first reconstructed pixel group as reference pixels. 此处,对于方向的确定不限于上述描述,而是可以根据从邻近像素到第二组的像素的距离,向第二组的像素的邻近像素的绝对值差施加权重, 或者可以使用其他各种方向性测量算法。 Here, for the determination of the direction it is not limited to the above description, but according to a second group of pixels from adjacent pixels of the distance, to the absolute value of the difference between the second group of pixels neighboring the pixel of heavy weight is applied, or may use a variety of other directional measurement algorithm.

返回参照图10,如果选择了方向性系数D1,则第二帧内预测器332b确定对应于方向性系数D1的第一组的重构的像素L。 Referring back to FIG. 10, if you select the directivity D1, the second intra predictor 332b determines the direction corresponding to the first set of coefficients D1 reconstructed pixel L. 与R。 And R. 为参考像素,并且使用参考像素的平均值(L。+R。) /2作为第二組的像素C的预测算子。 Reference pixels, and use the average of reference pixels (L. + R.) / 2 as the pixel C is predicted second group operator. 换言之, 预测第二组的像素C为(L。+R。)/2。 In other words, the prediction for the pixel C of the second group (L. + R.) / 2. 如果选择了方向性系数D2,则第二帧内预测器332b使用(L-2+R。+L。+R2) /4作为第二组的像素C的预测算子。 If you choose the directivity D2, the second intra predictor 332b use (L-2 + R. + L. + R2) / 4 as a second set of pixel C prediction operator. 如果选择了方向性系数D3,则第二帧内预测器332b使用(L-2+R2) /2作为第二组的像素C的预测算子。 If you choose the directivity D3, the second intra predictor 332b use (L-2 + R2) / 2 as the second group of pixels C prediction operator. 如果选择了方向性系数D4,则第二帧内预测器332b使用(L-4+R2+U+R4) /4作为第二组的像素C的预测算子。 If you choose the directivity D4, the second intra predictor 332b use (L-4 + R2 + U + R4) / 4 as a second set of pixel C prediction operator. 如果选择了方向性系数D5,则第二帧内预测器332b使用(U+R4) /2作为第二组的像素C的预测算子。 If you choose the directivity D5, the second intra predictor 332b use (U + R4) / 2 as the pixel C of the second set of prediction operator. 如果选择了方向性系数D6,则第二帧内预测器332b使用(L。+R-2+L2+R。) /4作为第二组的像素C的预测算子。 If you choose the directivity D6, the second intra predictor 332b use (L. + R-2 + L2 + R.) / 4 as a second set of pixel C prediction operator. 如果选择了方向性系数D7,则第二帧内 If you choose the directivity D7, the second frame

预测器332b使用(L2+R-2) /2作为第二组的像素C的预测算子。 Predictor 332b use (L2 + R-2) / 2 as the pixel C of the second set of prediction operator. 如果选择了 If you select

方向性系数D8,则第二帧内预测器332b使用(L2+R-4+L4+R-2) /4作为第二组 Directivity D8, the second intra-frame predictor 332b using (L2 + R-4 + L4 + R-2) / 4 as the second group

的像素C的预测算子。 Pixel C of prediction operator. 如果选择了方向性系数D9,则第二帧内预测器332b If you choose the directivity D9, the second intra predictor 332b

使用(L4+R-4) /2作为第二组的像素C的预测算子。 Use (L4 + R-4) / 2 as the second group of pixels C prediction operator.

可以仅利用第一组的重构的像素来如下生成方向性系数D1至D9: <formula>formula see original document page 14</formula> You can only use the reconstructed pixels of the first set to the following generation directivity D1 to D9: <formula> formula see original document page 14 </ formula>

对方向的确定以及对用于相对于剩余组的预测的参考像素的确定与上述的类似,由此不再描述。 Determination and the determination of the orientation with respect to the remaining groups prediction reference pixel and similar to the above, thereby will not be described.

如上所述,根据方向确定单元333确定的方向,利用先前处理的第一组的重构的像素,预测第二组的每个像素。 As described above, according to the direction determination unit 333 determines the direction using the reconstructed pixels of the first group previously processed, the prediction for each pixel of the second group. 一旦生成了第二组所有像素的预测值,则计算预测值与输入块中相应于第二组的像素之间的余量。 Once the second set of predicted value is generated for all the pixels, the predicted value calculating block corresponding to the input margin between the second group of pixels. 该余量由变换单元308变换,然后由量化单元310量化。 Balance converted by the conversion unit 308, and then quantized by the quantization unit 310. 第二组的量化的余量经过逆量化单元316与逆变换单元318,被加到第二组的预测值上进行重构,然后存储在帧存储器322中。 The balance of the second set of quantized through the inverse quantization unit 316 and inverse transform unit 318, it is added to the predicted value of the second set of reconstructed, and then stored in the frame memory 322.

图12显示根据本发明示范性实施例的相对于第三组的像素的帧内预测所需的第一组与第二组的像素。 Figure 12 shows a first group and a second group of pixels with respect to an example of a third set of pixel intra prediction required according to an exemplary embodiment of the present invention.

参照图12,利用先前处理的第一组与第二组的重构的像素,预测第三组的像素。 Referring to FIG. 12, the use of the previously reconstructed pixels of the first group and the second processing group, the third group of pixels of the prediction. 例如,可以利用在虚线1201中的当前块以及当前块的先前处理的邻近块中包含的第一组与第二组的重构的像素,预测第三组的像素1211。 For example, the dotted line 1201 can be used in the current block and the neighboring block previously processed current block included in the first group and the second group of reconstructed pixels, the prediction of a third set of pixels 1211. 与相对于第二组的像素的预测一样,为了确定用于相对于第三组的像素的预测的第一組与第二组的参考像素,方向确定单元333确定围绕第三组的像素的方向性。 With respect to the prediction of the second group of pixels as to determine a third set of prediction with respect to a first group of pixels and a second group of reference pixels, the direction determination unit 333 determines a pixel group around a third direction sex.

图13与图14A至图141为用来解释才艮据本发明示范性实施例的围绕第三组的像素确定方向性的处理的视图。 13 and 14A to 141 for explaining Burgundy It was the third group of pixels around an exemplary embodiment of the present invention to determine the directionality of view of the process. 在图13中,第三组的像素1300由C指示,并且邻近像素1300之上与之下的像素由仏和dj指示(i,j=_4,-3,-2,-l,0,l,2,3,4)。 In Figure 13, the third group of pixels 1300 by the C instruction, and the pixel on the adjacent pixels by 1300 and under the instructions 仏 and dj (i, j = _4, -3, -2, -l, 0, l , 2,3,4). 此处,U-4, U—2, U。 Here, U-4, U-2, U. , U2, U4, d—4, d—2, d。 , U2, U4, d-4, d-2, d. , d2,山为第一组的重构的像素,而U力U-3,仏,U3, db d—3,山,d3为第二组的重构的像素。 , D2, mountain reconstructed pixels of the first group, and U force U-3, Fo, U3, db d-3, Hill, d3 for the second group of reconstructed pixels.

参照图14A至图141通过计算第三组的像素C的邻近像素的绝对值差的平均值,如下计算方向性系数D1至D9: Referring to FIG. 14A through FIG. 141 to calculate the average of the absolute value of the difference of the pixel C of the third set of neighboring pixels, is calculated as follows directivity D1 to D9:

Dl=( Dl = (

D2=( D2 = (

D3=( D3 = (

D4=( D4 = (

D5= D5 =

D6=( D6 = (

D7=( D7 = (

D8-( D8- (

D9= D9 =

U -d I + I Ud U -d I + I Ud

• 1. *1 ,、0 0 • 1. * 1 ,, 0 0

Ud I + IU -d Ud I + IU -d

-2 0 -11 -20-11

Ud I + IU -d Ud I + IU -d

U,d2 I + IU,(^ U, d2 I + IU, (^

U,d+ |(图14E); U, d + | (FIG. 14E);

Ud I + I Ud Ud I + I Ud

U>d I + I Ud U> d I + I Ud

Ud + Ud Ud + Ud

2 "4 3-3 2 "43-3

U -d U -d

+ + + +

+■ + ■

+ +

ii Ud ii Ud

0 2 02

U k1 U k1

-1 3 -13

U -d U -d

-2 4 -twenty Four

I )/3 (图14A); I )/3 (图14B); I )/3 (图14C); I )/3 (图14D); I) / 3 (FIG. 14A); I) / 3 (FIG. 14B); I) / 3 (FIG. 14C); I) / 3 (FIG. 14D);

+ I uV I )/3 + I uV I) / 3

,3-1 3-1

+ I Ud I )/3 + I Ud I) / 3

-2 -2

(图14F) (图14(;) (图14H ) (FIG. 14F) (FIG. 14 (;) (FIG. 14H)

•4 • 4

(图141 );公式[3] (FIG. 141); Equation [3]

方向确定单元333选择所计算的方向性系数D1至D9中的最小方向性系数,并且确定在对应于所选方向性系数的方向上的第一组与第二组的重构的像素为参考像素。 Directivity direction determination unit 333 select calculated the minimum directivity D1 to D9, and determines the direction of the selected directivity factor corresponding to the first and second sets of reconstructed pixel reference pixels .

一旦方向确定单元333确定了围绕第三组的像素的方向性系数,第三帧内预测器332c就根据所确定的方向性,利用第一组与第二组的重构的像素, 预测第三组的像素。 Once the direction determination unit 333 determines the orientation coefficients around a third set of pixels, the third intra-frame predictor 332c is determined in accordance with the directivity using the first and second sets of reconstructed pixel, the third prediction group of pixels.

返回参照图13,如果选择了方向性系数D1,则第三帧内预测器"&确定对应于方向性系数Dl的第一组的重构的像素U。与d。为参考像素,并且使用参考像素的平均值(U。+d。) /2作为第三组的像素C的预测算子。换言之, 预测第三组的像素C为(U。+d。)/2。如果选择了方向性系数D2,则第三帧内预测器332c使用(U—,+山)/2作为第三组的像素C的预测算子。如果选择了方向性系数D3,则第三帧内预测器332c使用(U-2+d2) /2作为第三组的像素C的预测算子。如果选择了方向性系数D4,则第三帧内预测器33&使用(U-3+d3) /2作为第三组的像素C的预测算子。如果选择了方向性系数D5,则第三帧内预测器332c使用(U-4+d4) /2作为第三组的像素C的预测算子。如果选择了方向性系数D6,则第三帧内预测器332c使用(UrHi-】)/2作为第三组的像素C的预测算子。如果选择了方向性系数D7,则第三帧内预测器33"使用(U2+d-2) /2作为第三组的像素C的预测算子。 Referring back to FIG. 13, if you select the directivity D1, the third intra predictor "& OK Dl corresponding to the directivity of the first set of reconstructed pixel U. and d. As reference pixels, and use of reference the average value of the pixel (U. + d.) / 2 as the third set of pixel C prediction operator. In other words, the prediction of a third set of pixels C (U. + d.) / 2. If you select the direction coefficient D2, the third intra predictor 332c use (U -, + Mountain) / 2 as the third set of pixel C prediction operator if you selected directivity D3, then use a third intra predictor 332c. (U-2 + d2) / 2 as the third set of pixel C prediction operator. If you choose the directivity D4, the third intra predictor 33 & Use (U-3 + d3) / 2 as the third Group C of the pixel prediction operator. If you choose the directivity D5, the third intra predictor 332c use (U-4 + d4) / 2 as the third set of pixel C prediction operator. If you select directivity D6, the third intra predictor 332c use (UrHi -]) / 2 as the third set of pixel C prediction operator if you selected directivity D7, the third intra predictor 33. " Use (U2 + d-2) / 2 as the third group of pixels C prediction operator. 如果选择了方向性系数D8,则第三帧内预测器332c使用(U3+d-3) /2作为第三组的像素C的预测算子。 If you choose the directivity D8, then use a third intra predictor 332c (U3 + d-3) / 2 as the third group of pixels C prediction operator. 如果选择了方向性系数D9,则第三帧内预测器332c使用(U4+d-4) /2作为第三组的像素 If you choose the directivity D9, the third intra predictor 332c use (U4 + d-4) / 2 as the third group of pixels

C的预测算子。 C's prediction operator.

如上所述,根据方向确定单元333确定的方向,利用先前处理的第一组与第二组的重构的像素,预测第三组的每个像素。 As described above, according to the direction determination unit 333 determines the direction using the first and second sets of reconstructed previously processed pixels, each pixel of the prediction of the third group. 一旦生成了第三组所有像素的预测值,则计算第三组的预测值与输入块中相应于第三组的像素之间的余量。 Once generated the third set of prediction values of all pixels of the third group is calculated prediction value corresponding to the input block and the balance a third set of pixels. 该余量由变换单元308变换,然后由量化单元310量化。 Balance converted by the conversion unit 308, and then quantized by the quantization unit 310. 第三组的量化的余量经过逆量化单元316与逆变换单元318,被加到第三组的预测值上进行重构,然后存储在帧存储器322中。 A third set of quantized balance through the inverse quantization unit 316 and inverse transform unit 318, it is added to the predicted value of the third group reconstructed, and then stored in the frame memory 322.

图15显示根据本发明示范性实施例的相对于第四组的像素的帧内预测所需的第一组至第三组的重构的像素。 Figure 15 shows an example of intra prediction with the fourth group of pixels required for the first to third sets of reconstructed pixel according to an exemplary embodiment of the present invention.

参照图15,利用先前处理的第一组至第三组的重构的像素,预测第四组的像素。 Referring to FIG. 15, using the previously processed first to third sets of reconstructed pixels, the prediction of the fourth group of pixels. 例如,可以利用在虛线1510中的当前块以及当前块的先前处理的邻近块中包含的第一组至第三组的重构的像素,预测第四组的像素1511。 For example, the dotted line 1510 can be used in the current block and the neighboring block previously processed current block included in the first group to the third group of reconstructed pixels, the prediction of the fourth group of pixels 1511. 与相对于第二组与第三组的像素的预测一样,为了确定用于相对于第四组的像素的预测的第一组至第三组的参考像素,方向确定单元333确定围绕第四组的像素的方向性。 With respect to the prediction of the second group and the third group of pixels as to determine a prediction with respect to the fourth group of pixels of the first group to the third group of reference pixels, the direction determination unit 333 determines a fourth group of around directional pixel. 与参照图13与图14A至图14I描述的相对于第三组的像素的预测一样,使用第一组至第三组的重构的像素,预测第四组的像素。 Phase with reference to FIG. 13 and FIG. 14A to 14I described for predicting pixels of the third group of the same, using the first to third sets of reconstructed pixels, the prediction of the fourth group of pixels. 但是, 在确定围绕第四组的像素的方向性时,可以利用位于第四组的像素的左侧与右侧的第三组的像素的绝对值差的平均值,另外计算方向性系数DIO。 However, in determining a fourth group of pixels surrounding the directivity, you can use the average of the absolute value of the difference between the third group of the fourth group of pixels located at the left and right pixels, and additionally calculate directivity DIO.

一旦利用第一组至第三组的重构的像素生成了第四组所有像素的预测值,则计算第四组的预测值与输入块中相应于第四组的像素之间的余量。 Once the use of the first to third sets of reconstructed pixel values to generate a fourth set of prediction for all the pixels, the calculation of the fourth group of the input block and the predicted value corresponding to the balance of the fourth group of pixels. 该余量由变换单元308变换,然后由量化单元310量化。 Balance converted by the conversion unit 308, and then quantized by the quantization unit 310. 第四组的量化的余量经过逆量化单元316与逆变换单元318,被加到第四组的预测值上进行重构, 然后存储在帧存储器322中。 The balance of the fourth group of quantization to inverse quantization unit 316 and inverse transform unit 318, is added to the predicted value of the fourth group will be reconstructed, and then stored in the frame memory 322.

在以上本发明的示范性实施例中,依次预测通过划分输入块而获得的第一组至第四组的像素。 In the above exemplary embodiment of the present invention, in order to predict the first group to the fourth group of pixel blocks obtained by dividing the input. 在本发明另一示范性实施例中,第二組与第三组的处理顺序相互交换,并且利用第一组的重构的像素,预测第三组的像素,而利用第一组与第三组的重构的像素,预测第二组的像素。 In another exemplary embodiment of the present invention, the second set of interchangeable with the processing procedure of the third group, and the use of a first set of reconstructed pixels, the prediction of a third set of pixels, and the use of the first group and the third reconstructed pixel group, the prediction of the second group of pixels.

图16显示根据本发明另一示范性实施例的相对于第三组的像素的帧内预测所需的第一组的像素。 Figure 16 shows a first example of a group of pixels with respect to a third set of pixel intra prediction needed according to another exemplary embodiment of the present invention.

参照图16,利用先前处理的第一组的重构的像素,预测第三组的像素。 Referring to FIG. 16, the use of a first set of previously processed reconstructed pixels, the prediction of a third set of pixels. 例如,可以利用在虚线1610中的当前块以及当前块的先前处理的邻近块中包 For example, the dotted line 1610 can be used in the current block and the neighboring block previously processed packet of the current block

含的第一组的像素,预测第三组的像素1611。 The first group of pixels contained, forecast a third set of pixels 1611. 与本发明以上示范性实施例一样,方向确定单元333首先确定围绕第三组的^^素的方向性,并且利用对应于所确定的方向性的第一组的重构的像素,预测第三组的像素。 The present invention with the above exemplary embodiment, the direction determination unit 333 first determines a third group of surrounding pixel ^^ directivity, and reconstructed using the first set corresponding to the determined directivity pixel, the third prediction group of pixels.

图17显示根据本发明另一示范性实施例的相对于第二组的像素的帧内预测所需的第一组与第三组的像素。 17 shows a first group of pixels and a third set of intra prediction with respect to a second embodiment of the group of pixels required in accordance with another exemplary embodiment of the present invention.

参照图17,利用先前处理的第一组与第三组的重构的像素,预测第二组的像素。 Referring to FIG. 17, with the first group and the third group of reconstruction of previously processed pixels, the prediction of the second group of pixels. 例如,可以利用在虚线1710中的当前块以及当前块的先前处理的邻 For example, the dotted line 1710 can be used in the current block and the adjacent block previously processed current

近块中包含的第一组与第三組的像素,预测第二组的像素nii。 The first group and the third group of pixels near the block contained in the second set of predicted pixels nii. 与本发明以 In the present invention,

上示范性实施例一样,方向确定单元333首先确定围绕第二组的像素的方向性,并且利用对应于所确定的方向性的第一组与第三组的重构的像素,预测第二组的像素。 The exemplary embodiment, the direction determination unit 333 first determines directionality about the second group of pixels, and utilizes the directivity corresponding to the determined first group and the third group of reconstructed pixels, the second set of prediction pixels.

图18显示根据本发明另一示范性实施例的相对于第四组的像素的帧内预测所需的第一组至第三组的像素。 Figure 18 shows the first group to the third group of pixels with respect to the embodiment of the fourth group of the intra prediction pixels required in accordance with another exemplary embodiment of the present invention. 参照图18,利用先前处理的第一组至第三组的重构的像素,预测第四组的像素。 Referring to FIG. 18, using the previously processed first to third sets of reconstructed pixels, the prediction of the fourth group of pixels. 例如,可以利用在虚线1810中的当前块以及当前块的先前处理的邻近块中包含的第一组至第三组的像素,预测第四组的像素1811。 For example, the dotted line 1810 can be used in the current block and the first to third sets of previously processed pixels adjacent blocks included in the current block, the prediction of the fourth group of pixels 1811.

图,图20为显示根据本发明另一示范性实施例的视频帧内预测编码方法的流程图。 FIG, 20 is a flowchart showing a predictive encoding method of video frames according to another exemplary embodiment of the present invention.

参照图19,在操作1910,将输入块的像素划分为多个组。 Referring to FIG. 19, in operation 1910, the input pixel is divided into a plurality of block groups. 在操作1920, 利用输入块的先前处理的邻近块的像素,帧内预测该多个组中的第一组的像素。 In operation 1920, using the input block of pixels of the previously processed neighboring blocks, intra prediction in the plurality of groups of the first group of pixels. 计算第一组的帧内预测的像素的预测值与输入块的原始像素之间的余量, 然后将其变换和量化。 The balance of intra prediction calculating a first set of predicted pixel values between the input block of original pixels, and then transform and quantization. 与预测模式信息一起,量化的余量被熵编码。 Together with the prediction mode information, quantization margin is entropy coding. 对第一 First

组的量化的余量进行逆量化与逆变换,并且将第一组的逆变换的余量加到第一组的预测值进行重构,然后存储在存储器中。 Quantify the margin set for inverse quantization and inverse transform, and the inverse transform margin is added to the predicted value of the first group to reconstruct the first group, and then stored in the memory. 在操作1930,利用第一组的重构的像素,帧内预测第二组的像素。 In operation 1930, the use of reconstructed pixels of the first group, the second group of pixels intra prediction. 一旦生成了第二组的帧内预测的像素的预测值以及第二组的原始像素的预测值,就计算第二组的预测值与输入块中对应于第二组的原始像素之间的余量,然后使其经过变换和量化。 Once generated the prediction value predicted value of the second set of intra-prediction and a second set of pixels of the original pixels, to calculate the predicted value of the second set of the input block and the second group than between pixels corresponding to the original volume, then passed through a transform and quantization. 与预测模式信息一起,量化的余量被熵编码。 Together with the prediction mode information, quantization margin is entropy coding. 对第二组的量化的余量进行逆量化与逆变换,并且将第二组的逆变换的余量加到第二组的预测值进行重构,然后存储在存储器中。 The balance of the second set of quantized inverse quantization and inverse transform, and the remainder of the second group is added to the predicted value of the inverse transform of the reconstructed second set, and then stored in the memory. 在操作1940,利用第一组与第二组的重构的像素,帧内预 In operation 1940, with the first group of pixels and the second group of reconstruction, intra prediction

测第三组的像素。 The third group of pixels measured. 如上所述,计算第三组的预测值与输入块中对应于第三组的原始像素之间的余量,然后使其经过变换和量化。 As described above, the predicted values calculated with a third set of input block corresponding to the balance between the original pixels of the third group, and then passed through a transformation and quantization. 与预测模式信息一起, 量化的余量被熵编码。 Together with the prediction mode information, quantization margin is entropy coding. 对第三组的量化的余量进行逆量化与逆变换,并且将第三组的逆变换的余量加到第三组的预测值进行重构,然后存储在存储器中。 The remaining amount of the third group of quantized inverse quantization and inverse transform, and the balance a third set of inverse transform is applied to the prediction to reconstruct the value of the third group, and then stored in the memory.

在操作1950,利用第一组至第三组的重构的像素,帧内预测第四组的像素。 In operation 1950, using the first to third sets of reconstructed pixels, a fourth set of pixel intra prediction.

参照图20,根据本发明另一示范性实施例的视频帧内预测编码方法类似于根据本发明以上示范性实施例的视频帧内预测编码方法,但是第二组与第三组的处理顺序如上所述相互交换。 Referring to FIG. 20, similar to the intra-prediction video coding method according to the present invention, the above exemplary embodiment of the intra-prediction video coding method according to another exemplary embodiment of the present invention, but the second group and the third group of the above processing sequence The mutual exchange.

根据本发明示范性实施例的帧内预测中的余量值可能会大于根据相关技术的帧内预测中的余量值。 According to the intra prediction exemplary embodiment of the present invention, the remainder may be larger than the magnitude of intra prediction according to the related art in more than magnitude. 在这种情况下,可以将指示应用本发明示范性实施例的视频帧内预测编码方法的语法信息以画面、切片、或者画面组(GOP) 为单位加到编码的比特流的头部,以自适应地应用才艮据本发明示范性实施例的视频帧内预测编码方法。 In this case, you can use syntax information indicating an exemplary embodiment of the present invention, a video frame predictive coding method to screen, slice, or group of pictures (GOP) head units added to the coded bit stream to According to adaptively apply only Burgundy video frame exemplary embodiment of the present invention, a method of predictive coding. 例如可以为每个8x8块设置一比特语法,并且如 For example, one bit may be provided for each 8x8 block syntax, and as

流的头部;否则将0加到比特流的头部。 Otherwise, 0 is added to the bit stream head; the head of the stream. 在没有分离的方向模式信息的情况 In the absence of separation direction mode information

流。 Stream. 这是因为,根据本发明示范性实施例,在帧内预测解码对应像素期间, 可以利用围绕对应像素的邻近块的像素的绝对值差来确定方向。 This is because, according to an exemplary embodiment of the present invention, during the intraframe prediction decoding corresponding to the pixels, can be used around the pixel corresponding to the absolute value of the difference between adjacent blocks to determine the direction.

另外,在根据本发明示范性实施例的帧内预测时,编码端与解码端可以预先确定根据方向来分类组,例如,在模式1下检查方向性系数D1至D3、 在模式2下检查方向性系数D1至D6、以及在模式3下检查方向性系数D1至D9,而不是检查围绕每个组的每个像素的所有方向性,并且在视频帧内预测编码中,以画面、切片、或者GOP为单位将模式信息加到比特流的头部。 In addition, according to the intra predicting an exemplary embodiment of the invention, the encoding side and decoding side can be pre-determined by the direction of taxonomic groups, for example, in Mode 1 Check directivity D1 to D3, in mode 2 Check direction coefficient D1 to D6, and check in Mode 3 directivity D1 to D9, instead of checking around each pixel of each group all directional, and intra-prediction video coding, to the picture, slice, or GOP unit of the mode information is added to the head of the bit stream.

除上述的方向性确定方法之外,还可能有其他性能更好的方向性确定方法。 In addition to the above-described method for determining directionality addition, there may be a better method of determining the directionality of other properties. 例如,假设有N种方法,包括具有低复杂度与低性能的方向性确定方法直至具有高复杂度与高性能的方向性确定方法。 For example, suppose there are N ways, including those with low complexity and low performance of the method of determining the direction until a high complexity and high performance directional determining method. 在这种情况下,编码端对每个单元(每个画面、每个切片、或者每个GOP)应用该N种方法,以选择最优方法,并且将指示所选方法的索引信息加到每个单元的比特流的头部以发送到解码端。 In this case, the encoding side applications for each unit (each picture, each slice, or each GOP) of the N ways, in order to choose the best method, and index information indicating the selected method is added to each the head of a unit of the bit stream for transmission to the decoder.

图21为使用根据本发明示范性实施例的视频帧内预测解码装置的视频解码器2100的方框图。 Figure 21 is a video frame to use the present invention, according to an exemplary embodiment of a block diagram of a video decoder of the predictive decoding apparatus 2100.

参照图22,视频解码器2100包括:熵解码单元2110、重排单元2120、 逆量化单元2130、逆变换单元2140、运动补偿单元2150、帧内预测单元2160、 以及滤波器2170。 Referring to FIG. 22, the video decoder 2100 includes: an entropy decoding unit 2110, the rearrangement unit 2120, an inverse quantization unit 2130, an inverse transform unit 2140, a motion compensation unit 2150, intra-prediction unit 2160, and a filter 2170. 帧内预测单元2160对应于根据本发明示范性实施例的视频帧内预测解码装置。 Intra-prediction unit 2160 in accordance with the video frame corresponding to an exemplary embodiment of the present invention, in the predictive decoding means.

熵解码单元2110与重排单元2120接收压缩的比特流,并且进行熵解码, 由此生成量化的系数。 Entropy decoding unit 2110 and the rearrangement unit 2120 receive a compressed bitstream and perform entropy decoding, thereby generating quantized coefficients. 逆量化单元2130与逆变换单元2140对量化的系数进行逆量化与逆变换,由此提取变换编码系数、运动向量信息、头部信息、以及帧内预测模式信息。 Inverse quantization unit 2130 and 2140 pairs of quantized coefficients inverse transform unit performing inverse quantization and inverse transform, thereby extracting transformation encoding coefficients, motion vector information, header information, and intra prediction mode information. 此处,帧内预测模式信息可以包含预定语法,其指示输入比特流是否为被划分为多个组以通过根据本发明示范性实施例的视频帧内预测编码方法进行编码的比特流。 Here, the intra prediction mode information may comprise a predetermined syntax, which indicates whether the input bit stream is divided into a plurality of groups by the present invention, according to an exemplary embodiment of the intra-prediction video coding method for encoding the bitstream. 如上所述,可以从邻近块的像素的绝对值差,计算每个组的像素的预测的方向信息,并且由此可以不将方向信息包含在比特流中。 As described above, the absolute value of the difference between adjacent blocks of pixels from the predicted direction information calculated for each group of pixels, and thereby the direction information may not be included in the bitstream. 比特流还可以包含模式信息,其指示多个方向性中在实际编码中使用的方向性。 The bit stream can also contain pattern information indicating a plurality of directional directivity for use in the actual coding.

运动补偿单元2150与帧内预测单元2160利用解码的头部信息,根据编码的画面类型,生成预测的块,并且将预测的块加到误差D,n,以生成uF,n。 Motion compensation unit 2150 and the use of intra-prediction unit 2160 decodes header information, based on the picture type coded, the prediction block is generated, and is added to the prediction error block D, n, to generate uF, n. uF、由滤波器2170处理,由此生成重构的画面F,n。 uF, processed by filter 2170, thereby generating a reconstructed picture F, n.

具体地,根据本发明示范性实施例的帧内预测单元2160利用在收到的比特流中包含的帧内预测模式信息,确定在编码要解码的当前块时使用的帧内 Intra-prediction unit In particular, according to an exemplary embodiment of the present invention, the use of 2160 in the intra prediction mode information included in the received bit stream to determine when intra-coding a current block to be decoded using

测了收到的比特流时,帧内预测单元2160对第一组的像素进行帧内预测解码,并且利用先前处理的组的重构的像素,解码剩余组的像素。 Measuring the received bit stream, the intra-prediction unit 2160 pairs of the first group of pixels of intra prediction decoding, and the use of the reconstructed previously processed pixel group, the remaining groups of pixels decoded.

图22为才艮据本发明示范性实施例的图21的帧内预测单元2160的方框图。 According to FIG. 22 is only Burgundy exemplary embodiment of the present invention, the intra-prediction unit 21 is a block diagram of 2160.

参照图22,帧内预测单元2160包括帧内预测模式确定单元2161与帧内预测器2162。 Determining unit 2161 and the intra prediction unit 2162 with reference to FIG. 22, intra-prediction unit 2160 includes an intra prediction mode. 帧内预测器2162包含用于处理通过划分输入块而获得的组的帧内预测器。 Comprising intra-frame predictor 2162 intra-frame predictor for processing group by dividing the input block is obtained. 在图22中,输入块被划分为四个组,并且帧内预测器2162包括第一帧内预测器2162a、第二帧内预测器2162b、第三帧内预测器2162c、 以及第四帧内预测器2162d,来处理这四个组。 In Figure 22, the input block is divided into four groups, and the intra prediction unit 2162 includes a first intra prediction unit 2162a, a second intra prediction unit 2162b, the third intra-prediction unit 2162c, and the fourth frame predictor 2162d, to deal with these four groups.

帧内预测模式确定单元2161根据从收到的比特流中提取的帧内预测模式信息,在包括根据本发明示范性实施例的视频帧内预测模式的各种帧内预测模式中,确定要帧内预测解码的当前块被帧内预测编码的帧内预测模式, Intra prediction mode determining unit 2161 according to the extracted from the received bitstream intra prediction mode information, including video frame prediction mode according to an exemplary embodiment of the present invention, various intra prediction modes, is determined to be a frame intra-prediction decoded current block is intra prediction coded intra prediction mode,

以确定当前块的帧内预测模式。 To determine the intra prediction mode of the current block.

第一帧内预测器2162a解码在收到的比特流中包含的第一组的编码的视频数据,并且输出第一组的解码的视频数据。 Encoding a first intra prediction decoding unit 2162a included in the received bit stream of the first set of video data, and outputs a first set of decoded video data.

第二帧内预测器2162b、第三帧内预测器2162c、以及第四帧内预测器2162d通过相对于第二至第四组的像素计算位于预定角度上的邻近像素的绝对值差,确定剩余组的像素的方向性,根据所确定的方向性确定第一组的解码的像素和/或先前处理的其他组的解码的像素中的参考像素,并且利用所确定的参考像素的平均值,以组为单位对第二至第四组的像素进行帧内预测解码。 Second intra predictor 2162b, a third intra predictor 2162c, 2162d and the fourth intra predictor by computing relative to the absolute value of the pixel to the fourth group of the second predetermined angle is located adjacent pixels on the poor, to determine the remaining directional pixel group, determining the decoded pixels of the first group and / or decoding of the other groups previously processed pixels according to the determined reference pixels directivity, and use the average of the determined reference pixels, to group as a unit to the fourth group of the second pixel intra prediction decoding.

图23为显示根据本发明示范性实施例的视频帧内预测解码方法的流程图。 Figure 23 is a flowchart of display according to an exemplary embodiment of the present invention is a video decoding method of intra prediction.

参照图23,在操作2310,收到比特流。 Referring to FIG. 23, in operation 2310, received bitstream. 比特流包括通过划分输入块而获 Bitstream including by dividing an input block and eligible

得的多个组中的第一组的像素的数据,其利用邻近块的像素进行了帧内预测编码,以及利用第一组的重构的像素以及先前处理的其他组的重构的像素进行了帧内预测编码的像素的数据。 A plurality of groups of pixels obtained in a first set of data, utilizing the pixels of the neighboring blocks intra prediction encoding, and the use of a first set of reconstructed reconstructed pixel and the other groups previously processed pixels intra prediction coding pixel data.

在操作2320,对收到的第一组的像素进行帧内预测解码。 In operation 2320, the first group of pixels received intra-prediction decoding. 在操作2330, 确定剩余组的每个像素的方向性,并且根据所确定的方向性,使用第一组的解码的像素以及先前处理的其他组的解码的像素,预测剩余组的像素。 In operation 2330, it is determined for each pixel of the remaining groups directivity, and directivity according to the determined, using the decoded first group of pixels and the decoding of the other groups previously processed pixel, the prediction of the remaining groups of pixels.

本发明不仅可以用于YUV区域中的编码,而且也可以用于RGB区域。 The present invention is applicable not only to YUV coding region, but can also be used for RGB area. 在MPEG正在进行标准化的H.264 High 444简档或者高级444简档中,考虑余量颜色变换(RCT )。 In the ongoing standardization of MPEG H.264 High 444 444 profiles or Advanced profile, considering margin color conversion (RCT). 在2004年7月发布的JVT文件"Text of ISO/正C 14496 10 Advanced Video Coding 3rd Edition" (ISO/IEC JTC 1/2C 29/WG11 N6540)中, 详细描述了关于RCT的更多的信息。 In the July 2004 release of JVT file "Text of ISO / n-C 14496 10 Advanced Video Coding 3rd Edition" (ISO / IEC JTC 1 / 2C 29 / WG11 N6540) are described in detail more information about the RCT. 根据相关技术,RGB分量的视频数据被变换到另一色彩空间,例如YCbCr,以进行编码,并且重构YCbCr分量的视频数据,并且将YCbCr分量的重构的视频数据变换到RGB色彩空间以进行解码,由此在色彩格式变换期间造成了色彩失真。 According to the related art, the RGB components of video data is converted into another color space, such as YCbCr, for encoding, and the reconstructed video data component YCbCr, and converting the reconstructed video data component YCbCr to RGB color space for decoding, whereby during the color format conversion caused a color distortion. 为了减少色彩格式变换期间的色彩失真,人们正在研究用于在RGB区域中的视频编码与解码的RCT。 In order to reduce color distortion in color format conversion period, it is being investigated for the RGB video region encoding and decoding of RCT.

通过将RCT用于本发明的示范性实施例,在视频编码期间,对于根据本发明示范性实施例划分的组,可以在RGB区域中在紧接余量块变换之前进行RCT。 By RCT used in the present invention an exemplary embodiment, during video encoding, for the group according to an exemplary embodiment of the present invention is divided, in the region immediately RGB margin block transform prior to RCT.

更具体地,当将RCT用于本发明的示范性实施例时,对于RGB色彩空间 More specifically, when the RCT for the exemplary embodiment of the present invention, when, for the RGB color space

中的每个颜色分量,将输入块的像素划分为多个组。 Each color component of the input pixel block is divided into a plurality of groups. 接着,利用先前处理的邻近块的像素,预测该多个组中的第一组的像素。 Next, a pixel adjacent to the previously processed block, the prediction of the plurality of groups in the first group of pixels. 生成第一组的预测的像素 Generating the predicted pixels of the first group

的余量,并且对第一组的预测的像素的余量进行RCT。 Margin, and the margin of the first group of pixels prediction was RCT. 对第一组的像素的经RCT变换的余量进行DCT、量化、以及熵编码,以进行编码。 RCT transformed by the margin of the first group of pixels in the DCT, quantization and entropy coding to be encoded. 第一组的编码的像素被重构,然后被存储在存储器中,以用于预测剩余组的像素。 Encoding a first group of pixels is reconstructed, and then stored in the memory to be used for prediction of pixels of the remaining groups. 如上所述, 在确定了剩余组的像素的方向性之后,根据所确定的方向性,利用第一组的重构的像素、以及先前处理的其他组的重构的像素,预测剩余组的像素。 As described above, in determining the directionality of the remaining groups of pixels after, according to the determined directivity using reconstructed pixels of the first group, and the reconstruction of the other groups previously processed pixel, the prediction of the remaining groups of pixels . 通过从原始视频的对应区域中减去预测值,生成剩余组的预测的像素的余量, 并且对生成的余量进行色彩变换,并且编码该色彩变换的余量。 By subtracting the prediction value from the corresponding region of the original video, the pixels of the generated predicted margin remaining groups, and the color balance of the generated transform, and encoding the remainder of the color transformation.

除RCT之外,本发明的示范性实施例还可以用于涉及对余量的色彩变换的其他色彩变换。 In addition to the RCT, an exemplary embodiment of the present invention may also be used to refer to other color conversion of the balance of color transformation.

本发明的示范性实施例还可以实现为计算机可读记录介质上的计算机可读代码。 Exemplary embodiments of the present invention can also be implemented as a computer-readable computer readable recording medium. 计算机可读记录介质为可以存储以后可以由计算机系统读取的数据的任意数据存储设备。 The computer readable recording medium is any data storage device that can be stored can be later read by a computer system. 计算机可读记录介质的例子包括:只读存储器(R0M)、 随机访问存储器(RAM)、 CD-R0M、磁带、软盘、光学数据存储设备、以及载波。 Examples of the computer readable recording medium comprising: a read-only memory (R0M), a random access memory (RAM), CD-R0M, magnetic tape, floppy disks, optical data storage devices, and carrier waves. 计算机可读记录介质也可以在网络耦合的计算机系统上分布,乂人而可以分布方式存储和执行计算纟几可读代码。 The computer-readable recording medium can also be on a computer system coupled to the distribution network, qe person can perform calculations in a distributed fashion store and a few Si-readable code.

虽然参照本发明的示范性实施例具体显示与描述了本发明,但是本领域技术人员应该理解,在不脱离权利要求限定的本发明的精神与范围的前提下, 可以进行形式和细节的各种变化。 While the reference to an exemplary embodiment of the present invention is particularly shown and described the present invention, those skilled in the art would understand that in the spirit and scope of the claims without departing from the present invention may be made in form and detail change.

Referenced by
Citing PatentFiling datePublication dateApplicantTitle
CN102469308A *15 Nov 201023 May 2012深圳市融创天下科技股份有限公司Video coding intra-frame prediction method
CN103119945A *15 Jul 201122 May 2013三星电子株式会社Method and apparatus for encoding and decoding image through intra prediction
CN103125118A *6 Sep 201129 May 2013索尼公司Image processing device and image processing method
CN105049857A *15 Jul 201111 Nov 2015三星电子株式会社Method and apparatus for encoding and decoding image through intra prediction
US96547956 Feb 201516 May 2017Samsung Electronics Co., Ltd.Method and apparatus for encoding and decoding image through intra prediction
US965479620 Feb 201516 May 2017Samsung Electronics Co., Ltd.Method and apparatus for encoding and decoding image through intra prediction
US965479720 Feb 201516 May 2017Samsung Electronics Co., Ltd.Method and apparatus for encoding and decoding image through intra prediction
US966134420 Feb 201523 May 2017Samsung Electronics Co., Ltd.Method and apparatus for encoding and decoding image through intra prediction
WO2012065448A1 *25 Jun 201124 May 2012Shenzhen Temobi Science & Tech Co., LtdIntra frame prediction method for video encoding
Classifications
International ClassificationH04N7/32
Cooperative ClassificationH04N19/105, H04N19/147, H04N19/19, H04N19/59, H04N19/61, H04N19/136, H04N19/11, H04N19/593, H04N19/176
European ClassificationH04N7/34, H04N7/26A8B, H04N7/26A6C, H04N7/26A4C1, H04N7/50
Legal Events
DateCodeEventDescription
11 Mar 2009C06Publication
6 May 2009C10Entry into substantive examination
31 Oct 2012C14Grant of patent or utility model