CN101365136A - Method and apparatus for intra-frame prediction - Google Patents
Method and apparatus for intra-frame prediction Download PDFInfo
- Publication number
- CN101365136A CN101365136A CN 200810211844 CN200810211844A CN101365136A CN 101365136 A CN101365136 A CN 101365136A CN 200810211844 CN200810211844 CN 200810211844 CN 200810211844 A CN200810211844 A CN 200810211844A CN 101365136 A CN101365136 A CN 101365136A
- Authority
- CN
- China
- Prior art keywords
- buffer memory
- border
- piece
- row
- prediction piece
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The invention discloses a method and a device for intra-frame prediction. The method for the intra-frame prediction comprises the steps of (1) determining the row caching index of a lower boundary pixel value which is near to an upper prediction block of a current prediction block, and determining line caching index of a right boundary pixel value which is near to a left prediction block of the current prediction block; (2) taking a corresponding pixel value of the row caching index of the lower boundary pixel value which is near to the upper prediction block obtained in the lower boundary cache as the lower boundary pixel value which is near to the upper prediction block of the current prediction block; regarding the corresponding pixel value of the line caching index of the left boundary pixel value obtained in the right boundary cache as the right boundary pixel value which is near to the left prediction block of the current prediction block; (3) calculating the pixel value of the current prediction block according to the obtained lower boundary pixel value which is near to the upper prediction block and the right boundary pixel value which is near to the left prediction block. The scheme greatly saves prediction time and improves decoding speed of images.
Description
Technical field
The present invention relates to the picture decoding technology, relate in particular to method of intra-prediction and device in the picture decoding.
Background technology
Referring to Fig. 1, the picture decoding process in the standard H.264 may further comprise the steps:
Step 101 is carried out the entropy decoding to the coded data that receives, and obtains the decoded data of entropy.
Step 102 is carried out inverse quantization to the decoded data of entropy and is handled, and obtains conversion coefficient.
Step 103 is carried out inverse discrete cosine transform (DTC, Discrete CosineTransform) to conversion coefficient, obtains residual image.
Each pixel value of residual image is the poor of the pixel value of the original image respective point pixel value that deducts the predicted picture respective point, so-called predicted picture, just approaching with original image image.
Every width of cloth image consists of a plurality of macro blocks, and each macro block comprises the pixel value of 16 * 16 matrixes, macro block in image by from left to right sequence arrangement, has arranged delegation's macro block after, arrange from left to right since the second capable macro block again, until the lower right corner of image.For comprising the more part of details in the image, often the macroblock partitions with appropriate section is a plurality of sub-pieces, for example, 16 * 16 macro block matrix can be divided the sub-piece of 16 4 * 4 matrixes or be divided into the sub-piece of 48 * 8 matrixes.
Step 104 is carried out infra-frame prediction or inter prediction to the coded data that receives, and obtains predicted picture.
Step 105 is handled the predicted picture addition that obtains in the residual image obtain and the step 104 with step 103, obtains the image closely similar with original image.
So far, decode procedure finishes.
In the step 104 of above-mentioned flow process, when carrying out infra-frame prediction, with the macro block is that unit predicts, according to order from left to right, the macro block of first row in the macro block predict one by one, finish the prediction of the image first row macro block after, carry out the second row macro block prediction from left to right, the rest may be inferred, till entire image has all been predicted.If macro block is divided into a plurality of sub-pieces, when then this macro block being predicted, be that unit predicts with each height piece.As seen, current prediction piece may be a macro block, also may be sub-piece.
Each macro block has a sequence number, is called the macro block sequence number, according to macro block numeric order ground each macro block is predicted during infra-frame prediction.Each height piece in each macro block has sub-piece sequence number, and sub-piece sequence number has embodied the order that each height piece in this macro block is predicted.
When current prediction piece is predicted, need use the following border pixel values of the prediction piece adjacent with the border, top of current prediction piece, and the right border pixel values of the prediction piece that faces mutually with the left boundary of current prediction piece.When predicting, the absolute coordinate of each pixel of following border in image of prediction piece on the vicinity of at first definite current prediction piece, and the absolute coordinate of each pixel of border, the right in image of contiguous left side prediction piece; The absolute coordinate on the border, the right of piece is predicted on an absolute coordinate and a contiguous left side according to the following border of prediction piece on the vicinity of the current prediction piece of determining, inquires the pixel value of correspondence position in buffer memory; A pixel value and a contiguous left side according to the following border of prediction piece on the vicinity that inquires predict that the calculated for pixel values on the border, the right of piece goes out the pixel value of current prediction piece; And, the absolute coordinate of each pixel of border, the right of the current prediction piece of storage and the corresponding relation between its pixel value in buffer memory, and the absolute coordinate of each pixel of following border of current prediction piece and the corresponding relation between its pixel value, stored relation is used for follow-up other prediction pieces being used when carrying out pixel value prediction in buffer memory.
Flow process below by Fig. 2 is specifically described carrying out method of intra-prediction in the prior art, and this method may further comprise the steps:
The information that includes the infra-frame prediction type of each macro block in the coded data that decoding end receives, the corresponding type of prediction of each macro block sequence number, when current predicted macroblock is carried out infra-frame prediction, just can know the type of prediction of current predicted macroblock according to the macro block sequence number.
The method that this step is calculated absolute coordinate specifically comprises: the relative coordinate of each pixel of following border of prediction piece on the vicinity that calculates in the step 202 is added contiguous top macro block that step 203 the obtains coordinate with respect to the image upper left corner respectively, obtain the contiguous absolute coordinate of each pixel of following border in image of predicting piece of going up; The relative coordinate of each pixel of border, the right of the contiguous left side prediction piece that calculates in the step 202 is added contiguous left side macro block that step 203 the obtains coordinate with respect to the image upper left corner respectively, obtain the absolute coordinate of each pixel of border, the right in image that piece is predicted on a contiguous left side.
The width of this buffer memory is a picture traverse, and the length of this buffer memory is image length.
Obtain after the right border pixel values of the following border pixel values of prediction piece on the vicinity of current predicted macroblock and contiguous left side prediction piece, just can calculate the pixel value of current predictor piece, its computational methods are technology as well known to those skilled in the art, specifically can not give unnecessary details here referring to standard H.264.
Storing corresponding relation in the buffer memory into is used for follow-up other macro blocks being predicted.
The method that this step is calculated absolute coordinate specifically comprises: the relative coordinate of each pixel of following border of prediction piece on the vicinity that calculates in the step 208 is added that respectively the prediction piece is with respect to the coordinate of macro block under it on the vicinity that step 209 obtains, add that the macro block under the prediction piece obtains the contiguous absolute coordinate of following boundary pixel in image of going up the prediction piece with respect to the coordinate in the image upper left corner on the vicinity that step 210 obtains; A vicinity left side that calculates in the step 208 is predicted that the relative coordinate of each pixel of border, the right of piece adds the contiguous left coordinate of predicting piece with respect to macro block under it that step 209 obtains respectively, add the contiguous left coordinate of predicting the affiliated macro block of piece with respect to the image upper left corner that step 210 obtains, obtain the absolute coordinate of the right boundary pixel in image that piece is predicted on a contiguous left side.
All predictor pieces according to step 208～214 pair current predicted macroblock are predicted, just can calculate all pixel values in the current macro.
The method that this step is calculated absolute coordinate specifically comprises: the relative coordinate of each pixel of following border of prediction piece on the vicinity that calculates in the step 215 is added that respectively the prediction piece is with respect to the coordinate of macro block under it on the vicinity that step 216 obtains, add that the macro block under the prediction piece obtains the contiguous absolute coordinate of following boundary pixel in image of going up the prediction piece with respect to the coordinate in the image upper left corner on the vicinity that step 217 obtains; A vicinity left side that calculates in the step 215 is predicted that the relative coordinate of each pixel of border, the right of piece adds the contiguous left coordinate of predicting piece with respect to macro block under it that step 216 obtains respectively, add the contiguous left coordinate of predicting the affiliated macro block of piece with respect to the image upper left corner that step 217 obtains, obtain the absolute coordinate of the right boundary pixel in image that piece is predicted on a contiguous left side.
All predictor pieces according to step 215～221 pair current predicted macroblock are predicted, just can calculate all pixel values in the current macro.
Predict according to step 201～221 pair current predicted macroblock, calculate after all pixel values in the current macro, other macro blocks are carried out 201～221 similarly, just can calculate the pixel value of entire image.
In the existing intra-frame prediction method, need at first to obtain the following boundary pixel of prediction piece on the vicinity of current prediction piece and the absolute coordinate of the right boundary pixel in image of contiguous left side prediction piece through complicated calculating, inquire the following border pixel values of prediction piece on the vicinity and the right border pixel values that piece is predicted on a contiguous left side according to absolute coordinate in buffer memory, these computational methods are very consuming time.Especially when predicting with 4 * 4 matrixes in the frame, the process of calculating the right border pixel values of contiguous left side prediction piece of each 4 * 4 sub-piece and the contiguous following border pixel values that goes up the prediction piece is complicated more, and this has influenced decoding speed greatly.And H.264 video compression standard is very complicated, and computer need carry out a large amount of calculating just can decode image, if the computational speed of computer is not enough or computational methods are too complicated, then decoding speed can't reach the requirement that people watch video.
Summary of the invention
The invention provides a kind of method of intra-prediction, this method can improve the speed of picture decoding.
The invention provides a kind of device of infra-frame prediction, this device can improve the speed of picture decoding.
A kind of method of intra-prediction, this method comprises:
Each row pixel value of each macro block in the image is carried out serial number according to capable number, row buffer memory index as each row pixel value in each macro block, border, the right buffer memory is set, be used for the pixel value of rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width; Every capable pixel value in the image number is carried out serial number according to row, row cache index as each row pixel value in the image, following border buffer memory is set, be used for the pixel value of the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse;
Determine the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction piece, and determine the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction piece;
The row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece;
A following border pixel values and a contiguous left side according to prediction piece on the vicinity of obtaining predict that the right border pixel values of piece calculates the pixel value of current prediction piece.
A kind of device of infra-frame prediction, this device comprise neighborhood pixels value acquisition module, computing module, following border buffer memory and border, the right buffer memory;
Border, described the right buffer memory, the pixel value that is used for rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with each row of each macro block in the image according to row in each macro block;
Described following border buffer memory, the pixel value that is used for the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with every capable pixel in the image according to row in the image;
Described neighborhood pixels value acquisition module is used for determining the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction piece, and determines the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction piece; The row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece;
Described computing module, the following border pixel values and the contiguous left side that are used for prediction piece on the basis vicinity of obtaining predict that the right border pixel values of piece calculates the pixel value of current prediction piece.
From such scheme as can be seen, among the present invention, border, the right buffer memory that is provided with is used for the pixel value of rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with each row of each macro block in the image according to row in each macro block, the following border buffer memory that is provided with is used for the pixel value of the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, and the row cache index of each row pixel is for number carrying out sequence number after the serial number with every capable pixel in the image according to row in the image; When carrying out infra-frame prediction, determine the row buffer memory index of the right border pixel values of the following border pixel values of prediction piece on the vicinity of current prediction piece and contiguous left side prediction piece, the row buffer memory index corresponding pixel value of obtaining and determining in following border buffer memory and border, the right buffer memory is to calculate the pixel value of current prediction piece.Like this, do not need to carry out the right border pixel values and the contiguous absolute coordinate of going up the following border pixel values of prediction piece that the contiguous left side prediction piece of current prediction piece is determined in a large amount of calculating, with the right border pixel values and the contiguous following border pixel values that goes up the prediction piece that obtains contiguous left side prediction piece, thereby, save predicted time, improved the decoding speed of image.
Description of drawings
Fig. 1 is the picture decoding flow chart in the standard H.264;
Fig. 2 carries out the method for intra-prediction flow chart in the prior art;
Fig. 3 carries out the method for intra-prediction flow chart for the present invention;
Fig. 4 carries out the apparatus structure schematic diagram of infra-frame prediction for the present invention.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with embodiment and accompanying drawing, the present invention is described in more detail.
Carry out in the infra-frame prediction process, need obtain the following border pixel values of prediction piece on the vicinity of current prediction piece and the right border pixel values of contiguous left side prediction piece, according to the following border pixel values of prediction piece on the vicinity of obtaining and and the right border pixel values of contiguous left side prediction piece calculate the pixel value of current prediction piece.Among the present invention, the following border pixel values of prediction piece is stored in the buffer memory of following border on the vicinity of current prediction piece, and the right border pixel values of the contiguous left side prediction piece of current prediction piece is stored in the buffer memory of border, the right.Each row pixel value of each macro block in the image is carried out serial number according to capable number, row buffer memory index as each row pixel value in each macro block, border, the right buffer memory is set, be used for the pixel value of rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width; Every capable pixel value in the image number is carried out serial number according to row, row cache index as each row pixel value in the image, following border buffer memory is set, be used for the pixel value of the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse.Current prediction piece may be current predicted macroblock, also may be the current predictor piece after current predicted macroblock is divided.
The right border buffer memory stored 16 pixel values, can be expressed as 16 * 1 matrix, comprise 1 row, 16 row, from left to right, suppose that the row buffer memory concordance list of each pixel value correspondence is shown: 1,2,3......16; Image is unit with the macro block, and the pixel value of each macro block is 16 * 16 matrix, comprises 16 row, 16 row, and the pixel value of every row is 16, and so correspondingly, the row buffer memory index of each row pixel value is in the macro block: 1,2, and 3......16.
With A presentation video width, then following border buffer memory stored A pixel value, can be expressed as the matrix of A * 1, comprise 1 row A row, from left to right, suppose that the row cache concordance list of each pixel value correspondence is shown 1,2,3......A; So correspondingly, the row cache index of each row pixel value of image is shown by sequence list from left to right: 1,2, and 3......A.
The order that the present invention predicts each macro block in the image is same as the prior art, just at first according to from left to right order, macro block in the first row macro block is predicted one by one, after finishing the prediction of the image first row macro block, carry out the second row macro block prediction from left to right, the rest may be inferred, till entire image has all been predicted.If macro block is divided into a plurality of sub-pieces, when then this macro block being predicted, be that unit predicts with each height piece.
When each macro block is predicted, the row cache index of the following border pixel values of prediction piece on the vicinity of at first definite current prediction piece, and the row buffer memory index of the right border pixel values of piece is predicted on the contiguous left side of definite current prediction piece; Predict the row cache index corresponding pixel value of the following border pixel values of piece then on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece; Calculate the pixel value of current prediction piece again according to the right border pixel values of the following border pixel values of predicting piece on the vicinity of obtaining and contiguous left side prediction piece.
If prediction piece and contiguous left side prediction piece all exist on the vicinity of current prediction piece, then carry out in the forecasting process, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for prediction piece on the vicinity of current prediction piece stores its following border pixel values into the pixel value at the corresponding row buffer memory index place of following border buffer memory after calculating pixel value; The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is predicted piece for the contiguous left side of current prediction piece and its right border pixel values is cached to the pixel value at the respective column buffer memory index place of border, the right buffer memory after calculating pixel value.
If the contiguous left side prediction piece of current prediction piece exists, and the contiguous prediction piece of going up does not exist, for example, current prediction piece is the macro block of image first row, last row, then carry out in the forecasting process, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for predicting the corresponding initial value of row cache index of the following border pixel values of piece in the buffer memory of following border with on the vicinity of determining; The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is predicted piece for the contiguous left side of current prediction piece and its right border pixel values is cached to the pixel value at the respective column buffer memory index place of border, the right buffer memory after calculating pixel value.
If the contiguous left side prediction piece of current prediction piece does not exist, and the contiguous prediction piece of going up exists, for example, current prediction piece is the macro block of image second row, first row, then carry out in the forecasting process, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for prediction piece on the vicinity of current prediction piece stores its following border pixel values into the pixel value at the corresponding row buffer memory index place of following border buffer memory after calculating pixel value; The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is the contiguous left corresponding initial value of storing in the buffer memory of border, the right with determine of row buffer memory index of predicting the right border pixel values of piece.
If the contiguous left side prediction piece of current prediction piece does not exist, and the prediction piece does not exist yet on the vicinity, for example, current prediction piece is the macro block of image first row, first row, then carry out in the forecasting process, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for predicting the corresponding initial value of row cache index of the following border pixel values of piece in the buffer memory of following border with on the vicinity of determining; The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is the contiguous left corresponding initial value of storing in the buffer memory of border, the right with determine of row buffer memory index of predicting the right border pixel values of piece.
Calculate after the pixel value of current prediction piece, the right border pixel values of current prediction piece is stored into the position of respective column buffer memory index in the buffer memory of border, the right, if the corresponding row buffer memory of the right border pixel values index place with current prediction piece in the buffer memory of border, the right has stored pixel value, then upgrade the pixel value at respective column buffer memory index place in the buffer memory of border, the right with the right border pixel values of current prediction piece; And, the following border pixel values of current prediction piece is stored into the position at corresponding line buffer memory index place in the buffer memory of following border, if the corresponding row cache index of the following border pixel values place with current prediction piece in the buffer memory of following border has stored pixel value, then upgrade the pixel value at corresponding line buffer memory index place in the buffer memory of following border with the following border pixel values of current prediction piece.
Store the following border pixel values of the current prediction piece that calculates in the buffer memory of following border relevant position, store the right border pixel values of the current prediction piece that calculates in the buffer memory of border, the right relevant position.Like this, in the prediction of back, just can directly from border, the right buffer memory and following border buffer memory, obtain the contiguous left side of current prediction piece and predict the right border pixel values of piece and the following border pixel values of contiguous prediction piece, and do not need to carry out the right border pixel values and the contiguous absolute position of going up the following border pixel values of prediction piece that the contiguous left side prediction piece of current prediction piece is determined in a large amount of calculating, with the right border pixel values and the contiguous following border pixel values that goes up the prediction piece that obtains contiguous left side prediction piece.
Below in conjunction with Fig. 3 the present invention program is specifically described, uses LeftPixelBuf[16] border, expression the right buffer memory, use PicWidthPixelBuf[PicWidth] the following border of expression buffer memory.When initial, because the right border pixel values of the contiguous left side prediction piece of first macro block of the image upper left corner is 0, and, the pixel value on the following border of prediction piece all is 0 on the vicinity of image, therefore, can be with LeftPixelBuf[16] and PicWidthPixelBuf[PicWidth]] in the original pixel values of each buffer memory index correspondence all be made as 0, also can be with LeftPixelBuf[16] and PicWidthPixelBuf[PicWidth]] in the original pixel values of each buffer memory index correspondence all be made as sky.
Referring to Fig. 3,, in the present embodiment, suppose LeftPixelBuf[16 for method of intra-prediction flow chart of the present invention] in the row buffer memory index of pixel value of storage be: 1,2,3......16; With A presentation video width, suppose PicWidthPixelBuf[PicWidth]] in the row cache index of pixel value of storage be: 1,2,3......A, A are 16 multiple.The flow process of Fig. 3 may further comprise the steps:
Step 301 is judged the infra-frame prediction type of current predicted macroblock, if 16 * 16 matrix-type, then execution in step 302, if 8 * 8 matrix-type, then execution in step 306, if 4 * 4 matrix-type, then execution in step 310.
Step 302 is determined the row cache index of the following border pixel values of prediction piece on the vicinity of current predicted macroblock, determines the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current predicted macroblock.
When current predicted macroblock is carried out infra-frame prediction, just can determine the row cache index of top border pixel values of current predicted macroblock and the row buffer memory index of left boundary pixel value according to the macro block sequence number of current predicted macroblock, the contiguous row cache index of going up the following border pixel values of prediction piece of the just current prediction piece of the row cache index of the top border pixel values of current predicted macroblock, the row buffer memory index of the right border pixel values of the contiguous left side prediction of the just current prediction piece of the row buffer memory index of the left boundary pixel value of current predicted macroblock piece.
The method of row cache index of determining the top border pixel values of current predicted macroblock according to the macro block sequence number of current predicted macroblock can be the macro block sequence number that deducts initial macro block with the macro block sequence number of current macro, obtain subtracting each other the result, judge and to subtract each other the result whether less than the macroblock number of the every row of image, the macro block sequence number of current macro is represented with M, the macro block sequence number of initial macro block is represented with N, the result of M-N is represented with O, the macroblock number of the every row of image is represented with P, if O is less than P, then the row cache index of the top border pixel values of current predicted macroblock is: (M-N) * and 16+1, (M-N) * 16+2, (M-N) * 16+3...... (M-N) * 16+16; If O is more than or equal to P, then with O to the P complementation, obtain the complementation result, the result is expressed as Q with this complementation, then the row cache index of the top border pixel values of current predicted macroblock is: Q * 16+1, Q * 16+2, Q * 16+3......Q * 16+16.
Suppose that current predicted macroblock is first macro block in the upper left corner in the image, the macro block sequence number is 1, then the row cache index of the following border pixel values of prediction piece is on the vicinity of current predicted macroblock: 1,2,3......16, the row cache index of the right border pixel values of the contiguous left side prediction piece of current predicted macroblock is: 1,2, and 3......16.
For another example, suppose that current predicted macroblock is the macro block on first macro block the right, the image upper left corner, here should current predicted macroblock be called second macro block, the macro block sequence number is 2, and then the row cache index of the following border pixel values of prediction piece is on the vicinity of current predicted macroblock: 17,18,19......32, the row cache index of the right border pixel values of the contiguous left side prediction piece of current predicted macroblock is: 1,2, and 3......16.
Again such as, suppose that P is 3, M is 4, N is 1, O=M-N=3=P, and just current predicted macroblock is the macro block of image second row first row, with O to the P complementation, the complementation result who obtains is 0, and then the row cache index of the following border pixel values of prediction piece is on the vicinity of current predicted macroblock: 1,2,3......16, the row cache index of the right border pixel values of the contiguous left side prediction piece of current predicted macroblock is: 1,2, and 3......16.
Step 303, at PicWidthPixelBuf[PicWidth]] in the row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity obtaining and determine, at LeftPixelBuf[16] in the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine.
Suppose that current predicted macroblock is second macro block, then this step specifically comprises: at PicWidthPixelBuf[PicWidth]] in to obtain the row cache index be 17,18,19......32 pairing pixel value, following border pixel values as prediction piece on the vicinity of current predicted macroblock, at this moment, PicWidthPixelBuf[PicWidth]] in the row cache index be 17,18,19......32 pairing pixel value is an initial value, at LeftPixelBuf[16] in to obtain row buffer memory index be 1,2,3......16 pairing pixel value is as contiguous left the right border pixel values of predicting piece of current predicted macroblock.
Step 304 predicts that according to the following border pixel values and the contiguous left side of prediction piece on the vicinity of obtaining the right border pixel values of piece calculates the pixel value of current predicted macroblock.
Obtain after the right border pixel values of the following border pixel values of prediction piece on the vicinity of current predicted macroblock and contiguous left side prediction piece, just can calculate the pixel value of current predicted macroblock, its computational methods are prior art, do not give unnecessary details here.
Step 305, store the right border pixel values of the current predicted macroblock that calculates the position of respective column buffer memory index in the buffer memory of border, the right into, the following border pixel values of the current predicted macroblock that calculates is stored into the position of corresponding line buffer memory index in the buffer memory of following border.
Suppose that current predicted macroblock is second macro block, then the right border pixel values of current predicted macroblock is all pixel values of these predicted macroblock the 16th row, the row buffer memory index of each pixel value correspondence of the 16th row is: 1,2,3......16, because the right border pixel values of having stored first macro block in the buffer memory of border, the right, when therefore storing, the pixel value of respective column buffer memory index in the buffer memory of border on the right of upgrading with each pixel values of the 16th row of current predicted macroblock; The following border pixel values of current predicted macroblock is each pixel value of this predicted macroblock the 16th row, the row cache index of each pixel value correspondence of the 16th row is: 17,18,19......32, when storing, the 16th all pixel values of going of current macro are stored into the position of corresponding line buffer memory index in the buffer memory of following border.
Step 306 is determined the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction 8 * 8 sub-pieces, determines the right buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction 8 * 8 sub-pieces.
The contiguous row cache index of going up the following border pixel values of prediction piece of the just current predictor piece of the row cache index of the top border pixel values of current predictor piece, the row buffer memory index of the right border pixel values of the contiguous left side prediction of the just current predictor piece of the row buffer memory index of the left boundary pixel value of current predictor piece piece.
When current predictor piece is carried out infra-frame prediction, the row cache index of the following border pixel values of prediction piece on the vicinity of determining current predictor piece according to the macro block sequence number and the relative position of current predictor piece in affiliated macro block of macro block under the current predictor piece, determine the contiguous left row buffer memory index of predicting the right border pixel values of piece of current predictor piece according to the relative position of current predictor piece in affiliated macro block, for example, current predictor piece for 8 * 8, relative position in macro block be under it in macro block first the row secondary series sub-piece, then can know by inference simply, back 8 row cache index in the row cache index of the top border pixel values of corresponding its affiliated macro block of the row cache index of the top border pixel values of current predictor piece, similarly, can know by inference simply, preceding 8 in the row buffer memory index of the left boundary pixel value of corresponding its affiliated macro block of the row buffer memory index of the left boundary of current predictor piece ranks the buffer memory index.
Suppose that current predicted macroblock is first macro block in the upper left corner in the image, current predictor piece is first 8 * 8 sub-piece in the current predicted macroblock upper left corner, the row cache index of the following border pixel values of prediction piece is on the vicinity of then current predictor piece: 1,2,3......8, the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current predictor piece is: 1,2, and 3......8.
Step 307, at PicWidthPixelBuf[PicWidth]] in the row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity obtaining and determine, at LeftPixelBuf[16] in the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine.
Suppose that current predictor piece is first 8 * 8 sub-piece in the first macro block upper left corner, then this step specifically comprises: at PicWidthPixelBuf[PicWidth]] in to obtain the row cache index be 1,2,3......8 pairing pixel value, as the following border pixel values of prediction piece on the vicinity of current predictor piece, at LeftPixelBuf[16] in to obtain row buffer memory index be 1,2,3......8 pairing pixel value is as contiguous left the right border pixel values of predicting piece of current predictor piece.
Step 308 predicts that according to the following border pixel values and the contiguous left side of prediction piece on the vicinity of obtaining the right border pixel values of piece calculates the pixel value of current predictor piece.
Obtain after the right border pixel values of the following border pixel values of prediction piece on the vicinity of current predictor piece and contiguous left side prediction piece, just can calculate the pixel value of current predictor piece, its computational methods are prior art, do not give unnecessary details here.
Step 309, store the right border pixel values of the current predictor piece that calculates the position of respective column buffer memory index in the buffer memory of border, the right into, the following border pixel values of the current predictor piece that calculates is stored into the position of corresponding line buffer memory index in the buffer memory of following border.
Suppose that current predictor piece is first 8 * 8 sub-piece in the first macro block upper left corner, the right border pixel values of then current predictor piece is each pixel value of these predictor piece the 8th row, the row buffer memory index of each pixel value correspondence of the 8th row is: 1,2,3......8, when storing, the 8th each pixel value that is listed as of current predictor piece is stored into the position of respective column buffer memory index in the buffer memory of border, the right; The following border pixel values of current predictor piece is each pixel value of this predictor piece eighth row, the row cache index of each pixel value correspondence of eighth row is: 1,2,3......8, when storing, each pixel value of the eighth row of current predictor piece is stored into the position of corresponding line buffer memory index in the buffer memory of following border.
All predictor pieces according to step 306～309 pair current predicted macroblock are predicted, all pixel values in calculating current macro.
Step 310 is determined the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction 4 * 4 sub-pieces, determines the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction 4 * 4 sub-pieces.
The contiguous row cache index of going up the following border pixel values of prediction piece of the just current predictor piece of the row cache index of the top border pixel values of current predictor piece, the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of the just current predictor piece of row buffer memory index of the left boundary pixel value of current predictor piece.
When current predictor piece is carried out infra-frame prediction, the row cache index of the following border pixel values of prediction piece on the vicinity of determining current prediction piece according to the macro block sequence number and the relative position of current prediction piece in affiliated macro block of macro block under the current prediction piece, determine the contiguous left row buffer memory index of predicting the right border pixel values of piece of current prediction piece according to the relative position of current prediction piece in affiliated macro block, for example, current predictor piece for 4 * 4, relative position in macro block is the sub-piece that first row the 4th is listed as in macro block under it, then can know by inference simply, last four row cache index in the row cache index of the top border pixel values of corresponding its affiliated macro block of the index of the top border pixel values of current predictor piece, similarly, can know by inference simply, preceding four in the row buffer memory index of the left boundary pixel value of corresponding its affiliated macro block of the row buffer memory index of the left boundary of current predictor piece ranks the buffer memory index.
Suppose that current predicted macroblock is first macro block in the upper left corner in the image, current predictor piece is the sub-piece on first sub-piece the right, be herein referred to as the second sub-piece, the row cache index of the following border pixel values of prediction piece is on the vicinity of then current predictor piece: 5,6,7,8, the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current predictor piece is: 1,2,3,4.
Step 311, at PicWidthPixelBuf[PicWidth]] in the row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity obtaining and determine, at LeftPixelBuf[16] in the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine.
Suppose that current predictor piece is the second sub-piece in first macro block, then this step specifically comprises: at PicWidthPixelBuf[PicWidth]] in to obtain the row cache index be 5,6,7,8 pairing pixel values, following border pixel values as prediction piece on the vicinity of current predictor piece, at LeftPixelBuf[16] in to obtain row buffer memory index be 1,2,3,4 pairing pixel values are as contiguous left the right border pixel values of predicting piece of current predictor piece.
Step 312 predicts that according to the following border pixel values and the contiguous left side of prediction piece on the vicinity of obtaining the right border pixel values of piece calculates the pixel value of current predictor piece.
Step 313, store the right border pixel values of the current predictor piece that calculates the position of respective column buffer memory index in the buffer memory of border, the right into, the following border pixel values of the current predictor piece that calculates is stored into the position of corresponding line buffer memory index in the buffer memory of following border.
Suppose that current predictor piece is the second sub-piece in first macro block, the right border pixel values of then current predictor piece is each pixel value of these predictor piece the 4th row, the row buffer memory index of each pixel value correspondence of the 4th row is: 1,2,3,4, when storing, the 4th each pixel value that is listed as of current predictor piece is stored into the position of respective column buffer memory index in the buffer memory of border, the right; The following border pixel values of current predictor piece is each pixel value of this predictor piece the 4th row, the row cache index of each pixel value correspondence of the 4th row is: 5,6,7,8, when storing, the 4th each pixel value of going of current predictor piece is stored into the position of corresponding line buffer memory index in the buffer memory of following border.
All predictor pieces according to step 310～313 pair current predicted macroblock are predicted, all pixel values in calculating current macro.
Predict according to step 301～313 pair current predicted macroblock, calculate after all pixel values in the current macro, other macro blocks are carried out 301～313 similarly, just can calculate the pixel value of entire image.
If learn that in advance the type of prediction of each macro block in the image that need predict is all identical, need not execution in step 301 in the then above-mentioned flow process, do not need just to judge which kind of infra-frame prediction type is current predicted macroblock be.For example: if the infra-frame prediction type of image all is 16 * 16 matrix, then direct execution in step 302; If the infra-frame prediction type of image all is 8 * 8 matrix, then direct execution in step 306; If the infra-frame prediction type of image all is 4 * 4 matrix form, then direct execution in step 310.
Referring to Fig. 4, be the apparatus structure schematic diagram of infra-frame prediction of the present invention, this device comprises neighborhood pixels value acquisition module, computing module, following border buffer memory and border, the right buffer memory;
Border, described the right buffer memory, the pixel value that is used for rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with each row of each macro block in the image according to row in each macro block;
Described following border buffer memory, the pixel value that is used for the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with every capable pixel in the image according to row in the image;
Described neighborhood pixels value acquisition module is used for determining the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction piece, and determines the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction piece; The row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece;
Described computing module, the following border pixel values and the contiguous left side that are used for prediction piece on the basis vicinity of obtaining predict that the right border pixel values of piece calculates the pixel value of current prediction piece;
Alternatively, this device comprises memory module, the right border pixel values respective column buffer memory index stores of pixel value that is used for the current prediction piece that will calculate in the buffer memory of border, the right, with the following border pixel values corresponding row buffer memory index stores of the current prediction piece that calculates in the buffer memory of following border.
H.264 the method for the right border pixel values of the following border pixel values of prediction piece and contiguous left side prediction piece has solved and has calculated the slow-footed problem of adjacent edge circle pixel value in the standard on the vicinity of the current prediction piece of quick calculating that the present invention proposes.After the present invention predicts current prediction piece, store the following border pixel values of the current prediction piece that calculates in the buffer memory of following border relevant position, store the right border pixel values of the current prediction piece that calculates in the buffer memory of border, the right relevant position, like this, make the pixel value that follow-up prediction can directly be obtained to be needed from border, the right buffer memory and following border buffer memory, and do not need to carry out the right border pixel values and the contiguous absolute position of going up the following border pixel values of prediction piece that the contiguous left side prediction piece of current prediction piece is determined in a large amount of calculating, with the right border pixel values and the contiguous following border pixel values that goes up the prediction piece that obtains contiguous left side prediction piece, thereby, save predicted time greatly, improved the decoding speed of image.
Above-described specific embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the above only is specific embodiments of the invention; and be not intended to limit the scope of the invention; within the spirit and principles in the present invention all, any modification of being made, be equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (8)
1, a kind of method of intra-prediction is characterized in that, this method comprises:
Each row pixel value of each macro block in the image is carried out serial number according to capable number, row buffer memory index as each row pixel value in each macro block, border, the right buffer memory is set, be used for the pixel value of rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width; Every capable pixel value in the image number is carried out serial number according to row, row cache index as each row pixel value in the image, following border buffer memory is set, be used for the pixel value of the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse;
Determine the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction piece, and determine the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction piece;
The row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece;
A following border pixel values and a contiguous left side according to prediction piece on the vicinity of obtaining predict that the right border pixel values of piece calculates the pixel value of current prediction piece.
2, the method for claim 1, it is characterized in that, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for prediction piece on the vicinity of current prediction piece stores its following border pixel values into the pixel value at the corresponding row buffer memory index place of following border buffer memory after calculating pixel value;
The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is predicted piece for the contiguous left side of current prediction piece and its right border pixel values is cached to the pixel value at the respective column buffer memory index place of border, the right buffer memory after calculating pixel value.
3, the method for claim 1, it is characterized in that, when current prediction piece is positioned at described image the top, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for predicting the corresponding initial value of row cache index of the following border pixel values of piece in the buffer memory of following border with on the vicinity of determining;
The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is predicted piece for the contiguous left side of current prediction piece and its right border pixel values is cached to the pixel value at the respective column buffer memory index place of border, the right buffer memory after calculating pixel value.
4, the method for claim 1, it is characterized in that, when current prediction piece is positioned at described image Far Left, the row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is the contiguous left corresponding initial value of storing in the buffer memory of border, the right with determine of row buffer memory index of predicting the right border pixel values of piece.
5, the method for claim 1, it is characterized in that, when current prediction piece is positioned at described image the top and Far Left, the row cache index corresponding pixel value of the following border pixel values of prediction piece on that obtain and the vicinity that determine in the buffer memory of following border is for predicting the corresponding initial value of row cache index of the following border pixel values of piece in the buffer memory of following border with on the vicinity of determining;
The row buffer memory index corresponding pixel value of the right border pixel values of the vicinity left side prediction piece that obtains in the buffer memory of border on the right with determine is the contiguous left corresponding initial value of storing in the buffer memory of border, the right with determine of row buffer memory index of predicting the right border pixel values of piece.
6, the method for claim 1 is characterized in that, before the row cache index of the following border pixel values of prediction piece, this method comprises on the vicinity of described definite current prediction piece:
Judge the matrix form of current predicted macroblock, if the matrix form of current predicted macroblock is 16 * 16 matrix form, then described current prediction piece is current predicted macroblock, and the row cache index of the following border pixel values of prediction piece is on the described vicinity of determining current prediction piece: the row cache index of determining the following border pixel values of prediction piece on the vicinity of current predicted macroblock according to the macro block sequence number of current predicted macroblock;
If the matrix form of current predicted macroblock is 4 * 4 matrix form or 8 * 8 matrix form, then described current prediction piece is current predictor piece, the row cache index of the following border pixel values of prediction piece is on the described vicinity of determining current prediction piece: the row cache index of the following border pixel values of prediction piece on the vicinity of determining current predictor piece according to the macro block sequence number and the relative position of current predictor piece in affiliated macro block of macro block under the current predictor piece, the row buffer memory index of the right border pixel values of the described contiguous left side prediction piece of determining current predictor piece is: the row buffer memory index of the right border pixel values of determining the contiguous left side prediction piece of current predictor piece according to the relative position of current predictor piece in affiliated macro block.
7, a kind of infra-frame prediction device is characterized in that, this device comprises neighborhood pixels value acquisition module, computing module, following border buffer memory and border, the right buffer memory;
Border, described the right buffer memory, the pixel value that is used for rightmost one row after respective column buffer memory index stores is predicted the prediction piece, the width of this border, the right buffer memory is the macro block width, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with each row of each macro block in the image according to row in each macro block;
Described following border buffer memory, the pixel value that is used for the delegation bottom after corresponding row buffer memory index stores is predicted the prediction piece, the width of this bottom's border buffer memory is a picture traverse, and the row buffer memory index of each row pixel is for number carrying out sequence number after the serial number with every capable pixel in the image according to row in the image;
Described neighborhood pixels value acquisition module is used for determining the row cache index of the following border pixel values of prediction piece on the vicinity of current prediction piece, and determines the row buffer memory index of the right border pixel values of the contiguous left side prediction piece of current prediction piece; The row cache index corresponding pixel value of the following border pixel values of prediction piece on the vicinity of in the buffer memory of following border, obtaining and determining, following border pixel values as prediction piece on the vicinity of current prediction piece, the row buffer memory index corresponding pixel value of the left boundary pixel value that obtains and determine in the buffer memory of border on the right is as the right border pixel values of the contiguous left side prediction piece of current prediction piece;
Described computing module, the following border pixel values and the contiguous left side that are used for prediction piece on the basis vicinity of obtaining predict that the right border pixel values of piece calculates the pixel value of current prediction piece.
8, device as claimed in claim 7, it is characterized in that, this device comprises memory module, the right border pixel values respective column buffer memory index stores of pixel value that is used for the current prediction piece that will calculate in the buffer memory of border, the right, with the following border pixel values corresponding row buffer memory index stores of the current prediction piece that calculates in the buffer memory of following border.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810211844 CN101365136B (en) | 2008-09-09 | 2008-09-09 | Method and apparatus for intra-frame prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810211844 CN101365136B (en) | 2008-09-09 | 2008-09-09 | Method and apparatus for intra-frame prediction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101365136A true CN101365136A (en) | 2009-02-11 |
| CN101365136B CN101365136B (en) | 2011-01-26 |
Family
ID=40391226
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200810211844 Expired - Fee Related CN101365136B (en) | 2008-09-09 | 2008-09-09 | Method and apparatus for intra-frame prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101365136B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103621095A (en) * | 2011-06-13 | 2014-03-05 | 高通股份有限公司 | Border pixel padding for intra prediction in video coding |
| CN104025588A (en) * | 2011-10-28 | 2014-09-03 | 三星电子株式会社 | Method and device for intra prediction of video |
| CN104363455A (en) * | 2014-10-29 | 2015-02-18 | 复旦大学 | Hardware on-chip storage method applicable to infra-frame prediction reference pixels in HEVC (high efficiency video coding) standard |
| WO2015096812A1 (en) * | 2013-12-27 | 2015-07-02 | Mediatek Inc. | Method and apparatus for palette coding with cross block prediction |
| CN104933670A (en) * | 2015-07-02 | 2015-09-23 | 北京电子工程总体研究所 | Storage method and system of image pixel data |
| US10321141B2 (en) | 2013-12-18 | 2019-06-11 | Hfi Innovation Inc. | Method and apparatus for palette initialization and management |
| US10542271B2 (en) | 2013-12-27 | 2020-01-21 | Hfi Innovation Inc. | Method and apparatus for major color index map coding |
| US10743031B2 (en) | 2013-12-27 | 2020-08-11 | Hfi Innovation Inc. | Method and apparatus for syntax redundancy removal in palette coding |
| CN112022065A (en) * | 2020-09-24 | 2020-12-04 | 电子科技大学 | Method and system for quickly positioning time point of capsule entering duodenum |
-
2008
- 2008-09-09 CN CN 200810211844 patent/CN101365136B/en not_active Expired - Fee Related
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103621095B (en) * | 2011-06-13 | 2017-09-01 | 高通股份有限公司 | Boundary pixel for the infra-frame prediction in video coding is filled up |
| CN103621095A (en) * | 2011-06-13 | 2014-03-05 | 高通股份有限公司 | Border pixel padding for intra prediction in video coding |
| US9807399B2 (en) | 2011-06-13 | 2017-10-31 | Qualcomm Incorporated | Border pixel padding for intra prediction in video coding |
| US9621918B2 (en) | 2011-10-28 | 2017-04-11 | Samsung Electronics Co., Ltd. | Method and device for intra prediction video |
| CN107222744B (en) * | 2011-10-28 | 2020-10-27 | 三星电子株式会社 | Method and apparatus for intra prediction of video |
| US10506239B2 (en) | 2011-10-28 | 2019-12-10 | Samsung Electronics Co., Ltd. | Method and device for intra prediction video |
| CN104025588B (en) * | 2011-10-28 | 2017-08-15 | 三星电子株式会社 | Method and apparatus for the infra-frame prediction of video |
| US10893277B2 (en) | 2011-10-28 | 2021-01-12 | Samsung Electronics Co., Ltd. | Method and device for intra prediction video |
| CN107222744A (en) * | 2011-10-28 | 2017-09-29 | 三星电子株式会社 | Method and apparatus for the infra-frame prediction of video |
| US9883191B2 (en) | 2011-10-28 | 2018-01-30 | Samsung Electronics Co., Ltd. | Method and device for intra prediction video |
| CN104025588A (en) * | 2011-10-28 | 2014-09-03 | 三星电子株式会社 | Method and device for intra prediction of video |
| US10291919B2 (en) | 2011-10-28 | 2019-05-14 | Samsung Electronics Co., Ltd. | Method and device for intra prediction video |
| US10321141B2 (en) | 2013-12-18 | 2019-06-11 | Hfi Innovation Inc. | Method and apparatus for palette initialization and management |
| US10979726B2 (en) | 2013-12-18 | 2021-04-13 | Hfi Innovation Inc. | Method and apparatus for palette initialization and management |
| US10182242B2 (en) | 2013-12-27 | 2019-01-15 | Mediatek Inc. | Method and apparatus for palette coding with cross block prediction |
| CN106031171B (en) * | 2013-12-27 | 2019-08-30 | 联发科技股份有限公司 | Method and apparatus with the palette coding predicted across block |
| CN106031171A (en) * | 2013-12-27 | 2016-10-12 | 联发科技股份有限公司 | Method and apparatus for palette coding with cross block prediction |
| US10531119B2 (en) | 2013-12-27 | 2020-01-07 | Mediatek Inc. | Method and apparatus for palette coding with cross block prediction |
| US10542271B2 (en) | 2013-12-27 | 2020-01-21 | Hfi Innovation Inc. | Method and apparatus for major color index map coding |
| US10743031B2 (en) | 2013-12-27 | 2020-08-11 | Hfi Innovation Inc. | Method and apparatus for syntax redundancy removal in palette coding |
| WO2015096812A1 (en) * | 2013-12-27 | 2015-07-02 | Mediatek Inc. | Method and apparatus for palette coding with cross block prediction |
| US11166046B2 (en) | 2013-12-27 | 2021-11-02 | Hfi Innovation Inc. | Method and apparatus for syntax redundancy removal in palette coding |
| CN104363455A (en) * | 2014-10-29 | 2015-02-18 | 复旦大学 | Hardware on-chip storage method applicable to infra-frame prediction reference pixels in HEVC (high efficiency video coding) standard |
| CN104933670A (en) * | 2015-07-02 | 2015-09-23 | 北京电子工程总体研究所 | Storage method and system of image pixel data |
| CN112022065A (en) * | 2020-09-24 | 2020-12-04 | 电子科技大学 | Method and system for quickly positioning time point of capsule entering duodenum |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101365136B (en) | 2011-01-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101365136B (en) | Method and apparatus for intra-frame prediction | |
| US10542276B2 (en) | Data caching method and apparatus for video decoder | |
| KR100621137B1 (en) | Moving image encoding apparatus and moving image processing apparatus | |
| CN101848391B (en) | Dynamic image encoding device and dynamic image decoding device | |
| CN102547296B (en) | Motion estimation accelerating circuit and motion estimation method as well as loop filtering accelerating circuit | |
| WO2022104498A1 (en) | Intra-frame prediction method, encoder, decoder and computer storage medium | |
| CN104244007A (en) | Image compression method and device based on arbitrary shape matching | |
| US8902994B1 (en) | Deblocking filtering | |
| CN101505427A (en) | Movement estimation apparatus in video compression encoding algorithm | |
| CN104202602B (en) | Perform the device and method of Video coding | |
| CN107018418A (en) | Reference data method for reusing, band width estimation method and associated video decoder | |
| CN103414895A (en) | Encoder intra-frame prediction device and method applicable to HEVC standards | |
| CN101540911A (en) | Method and device for generating an image data stream, method and device for reconstructing a current image from an image data stream | |
| CN103188496A (en) | Fast motion estimation video encoding method based on motion vector distribution forecast | |
| CN106464880A (en) | Method and device for reducing a computational load in high efficiency video coding | |
| CN111741297B (en) | Inter-frame prediction method, video coding method and related devices | |
| CN102291581A (en) | Realizing method of self-adaptive motion estimation supporting frame field | |
| CN105850132A (en) | Method and apparatus for color index prediction | |
| US10999586B2 (en) | Image encoding method and equipment for implementing the method | |
| CN1968420B (en) | Methods of image processing for video encoder and decoder | |
| CN102088610A (en) | Video codec and motion estimation method thereof | |
| CN109495742B (en) | Video frame coding method, device and equipment | |
| CN117280691A (en) | Enhanced motion vector prediction | |
| WO2023155751A1 (en) | Video coding control method, apparatus, device, and storage medium | |
| CN101600111B (en) | Searching method for realizing secondary coding of self-adaptive interpolation filter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110126 |