WO2007063808A1 - 画像符号化/画像復号化方法及び画像符号化/画像復号化装置 - Google Patents
画像符号化/画像復号化方法及び画像符号化/画像復号化装置 Download PDFInfo
- Publication number
- WO2007063808A1 WO2007063808A1 PCT/JP2006/323606 JP2006323606W WO2007063808A1 WO 2007063808 A1 WO2007063808 A1 WO 2007063808A1 JP 2006323606 W JP2006323606 W JP 2006323606W WO 2007063808 A1 WO2007063808 A1 WO 2007063808A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- prediction
- pixel
- image
- mode
- signal
- Prior art date
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/103—Selection of coding mode or of prediction mode
- H04N19/11—Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/117—Filters, e.g. for pre-processing or post-processing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/124—Quantisation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/136—Incoming video signal characteristics or properties
- H04N19/14—Coding unit complexity, e.g. amount of activity or edge presence estimation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/147—Data rate or code amount at the encoder output according to rate distortion criteria
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
- H04N19/146—Data rate or code amount at the encoder output
- H04N19/152—Data rate or code amount at the encoder output by measuring the fullness of the transmission buffer
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
- H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
- H04N19/51—Motion estimation or motion compensation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
Definitions
- the present invention relates to an image encoding method, an image decoding method, an image encoding device, and an image decoding device for moving images or still images.
- H.264 a moving picture coding method that has greatly improved coding efficiency has been developed in collaboration with ITU-T and ISO / IEC in accordance with ITU-TRec. H.264 and ISOZIEC14496-10 (hereinafter, "H.264").
- Conventional encoding methods such as ISO / IECMPEG-1, 2, 4, ITU—TH. 2 61, H. 263 perform intra-frame prediction on the frequency domain (DCT coefficient) after orthogonal transform, and code the transform coefficient.
- H. 2 64 is designed to reduce the amount
- the conventional (ISOZlECMPEG-l, 2, 4) video is adopted by incorporating the direction prediction (Non-patent Document 1) in the spatial domain (pixel domain). Compared with the intra-frame prediction of the image coding method, it achieves high prediction efficiency.
- intra-frame prediction methods are defined for luminance signals, and one of them can be selected in units of macroblocks (16x16 pixel blocks). .
- the prediction methods are called 4x4 pixel prediction, 8x8 pixel prediction, and 16x1 6 pixel prediction, respectively.
- 16x16 pixel prediction In 16x16 pixel prediction, four encoding modes are defined, which are called vertical prediction, horizontal prediction, DC prediction, and plane prediction.
- the pixel value of the surrounding macroblock before applying the deblocking filter after the decoding process is used as the reference pixel value and used for the prediction process.
- 4x4 pixel prediction a luminance signal in a macroblock is divided into 16 4x4 pixel blocks, and one of nine modes is selected in block units for each 4x4 pixel block. Each of the nine modes has a prediction direction of 22.5 degrees, except for DC prediction (mode 2), which predicts with the average value of available reference pixels. A prediction value is generated by performing interpolation.
- mode 2 DC prediction
- One mode information for 4x4 pixel prediction 16 per macroblock are required.
- 4x4 pixel prediction is the unit power of the prediction process. Therefore, relatively efficient prediction can be performed even for images with complex textures, but prediction by simply copying the interpolated value in the prediction direction. There is a problem that the prediction error increases as the distance of the reference pixel force increases.
- 8x8 pixel prediction a luminance signal in a macroblock is divided into four 8x8 pixel blocks, and one of nine modes is selected in units of blocks for each 8x8 pixel block.
- the prediction mode is designed with the same framework as 4x4 pixel prediction, but encoding is performed by performing 3-tap filtering on the already encoded reference pixels and flattening the reference pixels used for prediction. A process of averaging the distortion is included.
- 4x4 pixel prediction there is a problem that the predicted value does not hit as the reference pixel force distance increases. In particular, since the distance power from the reference pixel is longer than the prediction of ⁇ 4 pixels, there is a problem that prediction accuracy cannot be expected for an image having a complex texture.
- 8x8 pixel prediction is a prediction unit defined only in V.264 high profile, and has been introduced especially for improving code efficiency in high-resolution images.
- the 4x4 pixel transform quantization block size is applied, and for 8x8 pixel prediction, the 8x8 pixel transform quantization block size is applied. That is, the transform quantization block size is defined by the predicted block shape. Since the prediction block shape considers the compatibility between the main profile and the high profile, it is possible to mix 4x4 pixel prediction and 8x8 pixel prediction in a macroblock!
- mode information is predicted by using the high correlation of mode information of adjacent blocks. Reduce the amount of code.
- the mode information is predicted, the 1-bit flag is encoded.
- the prediction is unsuccessful, the 3-bit data is encoded to reduce the code amount of the mode information.
- 4x4 pixel prediction is selected when there is almost no error signal in the macroblock, at least 16 bits (up to 64 bits) must be signed, which greatly increases the coding efficiency. May decrease.
- JP-A-2005-159947 as a technique for improving the prediction efficiency of an intra-frame prediction method, performs block matching in a screen and calculates a prediction block from an encoded reference image. Proposes a method to make predictions by supplementing the predicted values.
- This method is a prediction method based on the assumption that the image of an arbitrary coded block in the screen is similar to the image of the block to be predicted, and the prediction accuracy is low when the correlation in the screen is low. There is a problem that is bad.
- the amount of misregistration indicating the position of the reference image used for prediction must be encoded, and the amount of code of mode information may increase.
- an interpolation pixel corresponding to the encoded reference image power prediction mode is generated by the method specified in the H.264 high profile, and the direction specified in the prediction mode.
- the prediction error increases as the prediction pixel is further away from the prediction direction.
- Prediction block shapes cannot be mixed in the macroblock. There was a problem that the amount of code of information could not be reduced o
- an input image signal corresponding to one screen is divided into a plurality of pixel block signals, and the number of references changed according to the distance between the predicted pixel and the encoded reference pixel
- a prediction signal generating step for performing prediction within a screen to extrapolate the prediction pixel according to a direction defined by a plurality of prediction modes using pixels, and generating a prediction image signal and prediction mode information
- a prediction error signal is calculated from the block signal and the prediction image signal, a prediction mode selection step of selecting one prediction mode using the prediction error signal, and the selected! Based on the prediction mode
- a coding step of performing coding using a prediction error signal is
- a step of decoding an input encoded image signal to generate a decoded image signal including prediction mode information, and selective prediction based on prediction mode information of the decoded image signal Prediction is performed by changing the number of reference pixels used for prediction and the prediction pixel generation order according to the distance between the encoded reference pixel and the prediction pixel according to the mode, and by performing intra-screen prediction with extrapolation in units of pixels.
- a prediction signal generation step for generating an image signal, a prediction error signal generation step for generating a prediction error signal based on the decoded image signal, and a prediction image signal and the prediction error signal are added to generate a decoded image And a decoded image generation step.
- FIG. 1 is a block diagram showing a configuration of a moving picture coding apparatus according to a first embodiment of the present invention.
- FIG. 2 is a block diagram showing a configuration of an intra-frame prediction unit according to the embodiment.
- FIG. 3 is a flowchart of the image encoding device according to the embodiment.
- FIG. 4A is a diagram showing an outline of a prediction order according to the embodiment.
- FIG. 4B is a diagram showing an outline of a block shape according to the embodiment.
- FIG. 4C is a diagram showing an outline of a block shape according to the embodiment.
- FIG. 4D is a diagram showing an outline of a block shape according to the embodiment.
- FIG. 4E is a diagram showing an outline of a block shape according to the embodiment.
- FIG. 5A is a diagram showing direction prediction according to the embodiment.
- FIG. 5B is a diagram showing a relationship between a prediction block and a reference image in direction prediction according to the embodiment.
- FIG. 5C is a diagram showing a prediction method according to the embodiment.
- FIG. 5D is a diagram showing a prediction method according to the embodiment.
- FIG. 6A is a diagram showing a relationship between a prediction block and a reference image in pixel adaptive prediction according to the embodiment.
- FIG. 6B is a diagram showing pixel adaptive vertical prediction according to the embodiment.
- FIG. 6C is a diagram showing pixel adaptive prediction according to the embodiment.
- FIG. 7 is a view showing a weighting table of the number of reference images and filter coefficients used in pixel adaptive vertical prediction according to the embodiment.
- FIG. 8 is a diagram showing adjacent blocks used for prediction in the prediction mode according to the embodiment.
- FIG. 9 is a view showing mode names of direction prediction and pixel adaptive prediction according to the embodiment.
- FIG. 10 is a diagram showing a configuration of a moving image encoding apparatus according to the second embodiment of the present invention.
- FIG. 11 is a block diagram showing a configuration of an intra-frame prediction unit according to the embodiment.
- FIG. 12 is a flowchart of the image encoding device according to the embodiment.
- FIG. 13 is a schematic diagram of a syntax structure according to the embodiment.
- FIG. 14 is a diagram showing a data structure of sequence parameter set syntax according to the embodiment.
- FIG. 15 is a diagram showing a data structure of picture parameter set syntax according to the embodiment.
- FIG. 16 is a diagram showing a data structure of slice header syntax according to the embodiment.
- FIG. 17 is a diagram showing a data structure of macroblock layer syntax according to the embodiment.
- FIG. 18 is a diagram showing a data structure of sequence parameter set syntax according to the embodiment.
- FIG. 19 is a diagram showing a data structure of picture parameter set syntax according to the embodiment.
- FIG. 20 is a diagram showing a data structure of slice header syntax according to the embodiment.
- FIG. 21A is a diagram showing a data structure of macroblock layer syntax according to the embodiment.
- FIG. 21B is a diagram showing a data structure of the macroblock prediction syntax according to the embodiment.
- FIG. 22 is a block diagram showing a configuration of a moving picture encoding apparatus according to the third embodiment of the present invention.
- FIG. 23 is a diagram showing a data structure of sequence parameter set syntax according to the embodiment.
- FIG. 24 is a diagram showing a data structure of picture parameter set syntax according to the embodiment.
- FIG. 25 is a diagram showing a data structure of slice header syntax according to the embodiment.
- FIG. 26 is a diagram showing a data structure of macroblock layer syntax according to the embodiment.
- FIG. 27 is a diagram showing an mb-type data structure according to the embodiment.
- FIG. 28A is a diagram showing a data structure of block size switching according to the embodiment.
- FIG. 28B is a diagram showing a data structure of block size switching according to the embodiment.
- FIG. 29 is a block diagram showing a configuration of a video decoding apparatus according to the fourth embodiment of the present invention.
- FIG. 30 is a block diagram showing an intra-frame prediction unit in the video decoding apparatus according to the embodiment.
- FIG. 31 is a block diagram showing a configuration of a moving picture decoding apparatus according to the fifth embodiment of the present invention.
- a moving image encoding apparatus 100 shown in FIG. 1 is configured to encode a moving image signal by dividing it into a plurality of small pixel blocks.
- the moving image encoding device 100 is provided with a plurality of frame prediction modes having different block sizes and prediction image signal generation methods.
- Intraframe prediction is a prediction method in which prediction is closed within a frame, and a prediction target block is predicted using already encoded reference pixels.
- FIG. 4A it is assumed that sign left / right processing is performed toward the upper left force and lower right as shown in FIG. 4A.
- the moving image signal input to the moving image encoding apparatus 100 is divided into a plurality of small pixel blocks as the input image signal 115 by the screen dividing unit 101.
- a part of the divided input image signal 115 is input to the intra-frame prediction unit 102, and finally encoded by the encoding processing unit 107 via the mode selection unit 103, the transform and quantization unit 107.
- Sign It is output as digitized data 113.
- the screen dividing unit 101 divides a moving image signal into a plurality of pixel blocks for the purpose of each pixel block becoming a 16 ⁇ 16 pixel block shown in FIG. 4B.
- This 16x16 pixel block is called a macroblock and has the basic processing block size of the following encoding process.
- the moving picture encoding apparatus 100 performs an encoding process for each macroblock.
- the intra-frame prediction unit 102 extrapolates all the prediction modes that can be selected in units of macroblocks using the reference pixels temporarily stored in the reference image memory 106, and generates a prediction image signal 114. That is, the intra-frame prediction unit 102 generates all the predicted image signals 114 in modes that can be taken by the prediction pixel block by intra-screen prediction, for example, intra-frame prediction. However, the next prediction cannot be performed unless a locally decoded image is created in the macroblock as in H.264 intra-frame prediction (4x4 pixel prediction (see Fig. 4C) or 8x8 pixel prediction (see Fig. 4D)). In such a case, transformation and quantization, inverse quantization and inverse transformation may be performed within the intra-frame prediction unit 102, but a predicted image signal is generated only by the intra-frame prediction unit 102.
- the predicted image signal 114 generated by the intra-frame prediction unit 102 is sent to the mode selection unit 103.
- the mode selection unit 103 generates a prediction error signal 116 obtained by subtracting the prediction image signal 114 from the input image signal 115. Further, the mode selection unit 103 selects a prediction mode based on the mode information predicted by the intra-frame prediction unit 102 and the generated prediction error signal 116. More specifically, this embodiment uses a cost such as the following equation.
- the mode information is OH and the absolute sum of the prediction error signals is SAD, the following equation is used.
- ⁇ is given as a constant and is determined based on the quantization width and the value of the quantization parameter.
- the prediction mode is determined based on the cost obtained in this way. The prediction mode that gives the smallest cost repulsion is selected as the optimal prediction mode.
- the absolute sum of the mode information and the prediction error is used, but the prediction mode may be selected using only the mode information or only the absolute sum of the prediction errors.
- the absolute sum of mode information and prediction error may be Hadamard transformed, or values approximated to these may be used.
- the cost may be created using the activity of the input image signal, or the cost function may be created using the quantization width and the quantization parameter.
- the mode selection unit 103 is connected to the transform / quantization unit 104, and the mode information and the prediction error signal selected by the mode selection unit 103 are input to the transform / quantization unit 104.
- the transform / quantization unit 104 orthogonally transforms the input prediction error signal to generate transform coefficient data.
- the transform and quantization unit 104 converts the prediction error signal into a coefficient using discrete cosine transform or the like, but the prediction error signal is converted into a coefficient using a method such as wavelet transform or independent component analysis. May be converted to
- the transform and quantization unit 104 quantizes the transform coefficient.
- the quantization parameter required for the quantization is set by the sign key control unit 108.
- the quantized transform coefficient 117 is output to the code processing unit 107 together with prediction information 109 and information on a prediction method such as a quantization parameter.
- the code key processing unit 107 receives the prediction information and the like and inputs the quantized transform coefficient 117 as an entropy code (for example, Huffman code or arithmetic code).
- the data that has been entropy-encoded by the code processing unit 107 is multiplexed by the multiplexing unit 111, and is output as the code key data 113 through the output acknowledger 112.
- the inverse quantization and inverse transform unit 105 inversely quantizes the transform coefficient 117 quantized by the transform and quantization unit 104 according to the quantization parameter set by the sign key control unit 108,
- the obtained transform coefficient is subjected to inverse transform (for example, inverse discrete cosine transform), and the transform coefficient is decoded into the prediction error signal 116.
- inverse transform for example, inverse discrete cosine transform
- the prediction error signal 116 decoded by the inverse quantization and inverse transform unit 105 is added to the prediction image signal 114 of the selected prediction mode supplied from the mode selection unit 103 by the adder 118.
- the addition signal becomes a decoded image signal 119 and is input to the reference image memory 106.
- the reference image memory 106 stores the decoded image signal 119 as a reference image. In this way, the reference image stored in the reference image memory 106 is referred to when the prediction error signal or the like is generated by the intra-frame prediction unit 102.
- the encoding loop (in-frame prediction unit 102 ⁇ mode selection unit 103 ⁇ transformation and quantization unit 104 ⁇ inverse transform and inverse quantization unit 105 ⁇ reference image memory in FIG. Processing) is performed once for all prediction modes that can be selected in the macroblock.
- the input image signal 115 of the next block is input, and encoding is performed.
- the encoding control unit 108 performs feedback control of generated code amount, quantization characteristic control, mode control, and the like. Further, the code key control unit 108 controls rate control for controlling the amount of generated code.
- the control of the prediction unit and the control of the entire code are performed.
- FIG. 2 An example in which the moving image encoding apparatus 100 implements the moving image encoding method according to the present invention will be described with reference to FIG. As shown in FIG. 2, the same components as those in FIG.
- the intra-frame prediction unit 102 also performs an internal mode selection unit 204, an internal transform / quantization unit 206, an internal inverse quantization / inverse transform unit 207, A reference image memory 209 is included.
- the pixel adaptive prediction unit 201 and the direction prediction unit 202 each have a plurality of prediction modes, but the prediction methods are different.
- the fixed mode prediction unit 203 executes one of prediction modes in which mode information is not transmitted at the power macroblock level for performing prediction using the prediction methods of the pixel adaptive prediction unit 201 and the direction prediction unit 202.
- the direction prediction unit 202 and the pixel adaptive prediction unit 201 will be described in detail. These prediction units use the already decoded reference pixels stored in the reference image memory 106 to predict a block to be predicted. There are 9 prediction modes. As shown in Fig. 5A, except for mode 2, each has a different prediction direction by 22.5 degrees. Modes 0 to 8 excluding mode 2 are defined, and mode 2 is DC prediction in the direction prediction unit 202. The names of the direction prediction modes performed in the direction prediction unit 202 and the pixel adaptive prediction modes performed in the pixel adaptive prediction unit 201 are shown in FIG. The relationship between the prediction block for 4x4 pixel prediction and the reference pixel is shown in Fig. 5B. Pixels from uppercase A to M are reference pixels, and pixels from lowercase a to p are prediction pixels.
- the direction prediction unit 202 operates in mode 2 When the DC prediction is selected, the prediction pixel is calculated according to the following equation (2).
- the direction prediction unit 202 uses a prediction method that copies the prediction value supplemented from the reference pixel with respect to the prediction direction shown in FIG. 5A. Specifically, the prediction value generation method when mode 0 (vertical prediction) is selected will be described based on the following equation (3).
- This prediction mode can be selected only when reference pixels A to D are available.
- the details of the prediction method are shown in Fig. 5 (c).
- the luminance values of the reference pixels A to D are copied as they are to the prediction pixel in the vertical direction and supplemented as the prediction value.
- This prediction mode can only be selected when reference pixels A to D and I to M are available.
- the details of this prediction mode are shown in Fig. 5 (d).
- the value generated by the 3-tap filter is copied to the predicted pixel in the direction of 45 degrees to the lower right and compensated as the predicted value.
- the pixel application prediction unit 201 will be described.
- the pixel adaptive prediction unit 201 performs prediction by changing the number of reference pixels to be used according to the distance between the prediction pixel and the reference pixel, and the prediction value changes in units of pixels.
- Intraframe prediction is prediction that uses the spatial correlation of images, and the prediction value is created on the assumption that the luminance values between adjacent pixels are similar. If the distance between the predicted pixel and the available reference pixel is increased, this assumption is not satisfied, and the prediction error tends to increase. Therefore, when the distance between the prediction pixel and the reference pixel is increased, the decrease in the prediction error is reduced by increasing the number of reference pixels to be used according to the distance. Further, by changing the weighting table of the reference pixels to be used according to the distance, it is possible to generate a predicted value with high accuracy.
- FIG. 6A shows the relationship between the reference pixel and the prediction block.
- the reference pixels and prediction pixels shown in Fig. 6A and Fig. 5B have different indexes to facilitate explanation of the force prediction formulas corresponding to one-to-one.
- Pre-pixels ⁇ and pixels xlO force xl3, xl9 force et al. X22, x28 force et al. X31, x37 force et al. 16 pixels from pixel xl4 force xl7, x23 force etc. x26, x32 force etc. x35, x41 force etc. x44 are prediction auxiliary pixels and are used to improve the prediction accuracy.
- Figure 6B shows the prediction method for pixel adaptive vertical prediction.
- the prediction value is determined using more reference pixels as the distance between the reference pixel and the prediction pixel increases.
- a prediction value is generated using three reference pixels of ⁇ , ⁇ 1, and x02.
- a prediction value is generated using five reference pixels xOO, x01, x02, x03, and x04 at a prediction pixel x20 that is two distances away from the reference pixel in the prediction direction.
- prediction is performed using seven reference pixels xOO, x01, x02, x03, x04, x05, and x06. A value is generated.
- a prediction value is generated using nine reference pixels of ⁇ 0, x01, x02, x03, x04, x05, x06, x07, and x08.
- the predicted value generation method will be specifically described by the following equation (5).
- the predicted pixel is calculated using the following equation (5).
- X (n) (X (n-d- 1) + (X (n—d) ⁇ 1) + X (n— d + 1) + 2)>> 2 (5)
- blk—num is 4 for a 4x4 pixel block and 8 for an 8x8 pixel block.
- L represents the distance between the reference pixel and the prediction pixel.
- Vi is an index determined according to the corresponding prediction mode.
- hi represents the filter coefficient, and the number of taps varies with L. Specifically, hi and Vi will be described regarding pixel adaptive vertical prediction.
- Figure 7 shows an example of the weighting filter coefficient hi (weighting table) used according to the distance between the reference pixel and the predicted image during 4x4 pixel prediction.
- a prediction pixel that is one distance away from the reference pixel in the prediction direction is predicted by the following equation (8) using three reference pixels.
- X (n) (X (nd- 1) + (X (n—d) ⁇ 1) + X (n— d + 1) + 2)>> 2 (8)
- Reference pixel force The prediction pixel that is two distances from the prediction direction uses five reference pixels and
- X (n) (X (n— (d ⁇ 1) 2) + (X (n— (d ⁇ 1) 1) ⁇ 2) + (6X (n— (d ⁇ 1 ))) + (X (n— (d ⁇ 1) + 1) ⁇ 2) + X (n— (d ⁇ 1) +2) +8)>> 4 (9)
- Reference pixel force Prediction pixels that are three distances from the prediction direction use seven reference pixels.
- X (n) (X (n— 3d— 3) + (6X (n— 3d— 2)) + (15X (n— 3d— 1)) + (20X (n— 3d)) + ( 15X (n-3d + l)) + (6X (n— 3d + 2)) + X (n— 3d + 3) +32) >> 6 (10)
- the prediction pixel that is four distances from the reference pixel in the prediction direction uses nine reference pixels.
- X (n) (X (n— (d ⁇ 2) 4) + (X (n— (d ⁇ 2) — 3) ⁇ 3) + (28X (n— (d ⁇ 2) -2)) + (56X (n— (d ⁇ 2) — 1)) + (70X (n— (d ⁇ 2))) + (56X (n— (d ⁇ 2) +1 )) + (28X (n- (d ⁇ 2) +2)) + (X (n— (d ⁇ 2) + (X (n— (d ⁇ 2) + 3) ⁇ 3) + X (n— (d ⁇ 2) +4 ) + 128)>> 8 (11)
- the reference pixels used for prediction, the number of reference pixels, and the weighting table can be obtained by expanding Equation (5).
- X (31) (X (01) + (6X (02)) + (15X (03)) + (20X (04)) + (15X (05)) + (6
- prediction is performed by supplementing the last reference pixel value with the prediction formula. For example, when predicting xl9, the left reference pixel of xOO cannot be used. However, since a reference pixel of x09 can be used, it is predicted as the following equation (13).
- Eq. (12) can be derived by determining the pixels used for prediction in Eq. (5) and developing the necessary predicted values.
- the prediction formula for pixel adaptive horizontal prediction (mode 1) is expressed by the following equation (14).
- the necessary predicted image is assigned and developed, so that the necessary reference pixels, the number of reference pixels, and the weighting table are obtained. Bull can be determined.
- the reference pixel is compensated as in the pixel adaptive vertical prediction. For example, when x37 is predicted by pixel adaptive horizontal prediction, the reference pixel of ⁇ 45 cannot be used! Therefore, as shown in the following equation (22), this reference pixel is filled with ⁇ 36.
- the code efficiency can be improved. Due to improved prediction accuracy, the predicted image becomes closer to the input image, and this is a highly effective prediction method visually.
- the fixed mode prediction unit 203 performs prediction of mode information, and performs pixel adaptive prediction or direction prediction in a predetermined prediction mode.
- mode information of adjacent prediction blocks is used for prediction of mode information.
- Figure 8 shows the relationship between adjacent blocks in the case of 4x4 pixel prediction. Let the left adjacent block of prediction block C be ⁇ and the upper adjacent block be ⁇ . Given these two prediction mode forces 3 ⁇ 4rev-left-mode and prev-upper-mode, the prediction mode of the block C is determined by the following equation (23).
- current-mode is the prediction mode of prediction block C.
- the function min (A, B) returns the smaller of A and B.
- this mode is capable of greatly reducing the amount of code for encoding the prediction mode information.
- whether to generate a prediction image using the pixel adaptation prediction unit 201 or to generate a prediction image using the direction prediction unit 202 is determined by an ex-direct-intra flag described later. It is determined based on the ex-direct-intra flag information given from the encoding control unit 108.
- the prediction mode information of the current prediction block is determined based on the prediction mode information of adjacent prediction blocks A and B.
- the determination may be made based on prediction mode information of surrounding prediction blocks.
- the mode may be determined.
- the prediction mode with the highest frequency among the prediction modes of the peripheral prediction block, the median value of the prediction mode of the peripheral prediction block, or the average value of the prediction modes of the peripheral prediction block is determined as the prediction mode of the current prediction block. You can do it.
- the intra-frame prediction unit 102 shown in FIG. 2 A detailed operation of the intra-frame prediction unit 102 shown in FIG. 2 will be described.
- the input image signal 115 is input to the intra-frame prediction unit 102
- the signal is input to the pixel adaptive prediction unit 201, the direction prediction unit 202, and the fixed mode prediction unit 203.
- the pixel adaptive prediction unit 201, the direction prediction unit 202, and the fixed mode prediction unit 203 generate corresponding prediction image signals 114.
- the corresponding prediction unit when prediction prohibition information is input from the sign key control unit 108, the corresponding prediction unit does not generate a predicted image signal. Specifically, when the prediction prohibition information is input to the pixel adaptive prediction unit 201, the pixel adaptive prediction unit 201 does not perform the generation process of the predicted image signal 114. When the prediction prohibition information is input to the direction prediction unit 202, the direction prediction unit 202 does not generate the predicted image signal 114. When the prediction prohibition information is input to the fixed mode prediction unit 203, the prediction image signal 114 is not generated by the fixed mode prediction unit 203.
- the corresponding prediction unit and prediction switching switch 205 are not connected. It is not allowed to input prediction prohibition information to the pixel adaptive prediction unit 201 and the direction prediction unit 202 at the same time.
- the predicted image signal 114 and the prediction mode information predicted by each prediction unit are input to the internal mode selection unit 204.
- the internal mode selection unit 204 generates a prediction error signal 116 obtained by subtracting the prediction image signal 114 from the input image signal 115.
- the internal mode selection unit 204 performs mode selection based on the prediction error signal 116 generated by the prediction mode information generated by each prediction unit.
- the prediction mode is selected using the cost used in Equation (1).
- the selected directional prediction mode is the prediction mode predicted by the pixel adaptive prediction unit 201.
- the internal mode selection unit 204 transmits mode information representing the force that is the prediction mode predicted by the measurement unit 202 to the sign key control unit 108.
- ex-adaptive-intra-flag (described later) is TRUE, and when direction prediction is used, the flag is FALSE.
- the sign key control unit 108 controls the prediction switching switch 205 according to the given mode information. Whether or not the fixed mode prediction is performed is judged by ex- direct- intra- flag. When this flag is TRUE, fixed mode prediction is performed, and when it is FALSE, this prediction is not performed.
- the prediction switching switch 205 When the ex-adaptive-intra-flag is TRUE, the prediction switching switch 205 is connected to the pixel adaptive prediction unit 201. When the ex-adaptive-intra-flag is FALSE, the prediction switching switch 205 is connected to the direction prediction unit 202. If the ex-direct-intra-flag is TRUE, the sign control unit 108 connects the switch 205 to the fixed mode determination unit 203, respectively.
- the internal mode selection unit 204 When prediction has been completed for all small pixel blocks, the internal mode selection unit 204 outputs the predicted image signal 114 and mode information for the macroblock from the intra-frame prediction unit 102.
- the prediction switching switch 205 is connected to the internal conversion and quantization unit 206, and the mode information selected by the internal mode selection unit 204 and the prediction error signal 114 pass through the prediction switching switch 205. Input to the internal conversion and quantization unit 206.
- the internal transform / quantization unit 206 converts the input prediction error signal 114 into a transform coefficient using a discrete cosine transform or the like, and generates transform coefficient data.
- the prediction error signal 114 may be converted using a method such as wavelet conversion or independent component analysis.
- the internal transform / quantization unit 206 quantizes the transform coefficient.
- the quantization parameter required for the quantization is set by the encoding control unit 108.
- the internal inverse quantization and inverse transform unit 207 converts the transform coefficient 117 quantized by the internal transform and quantization unit 206 according to the quantization parameter set by the sign key control unit 108. Inverse quantization is performed, and inverse transform (for example, inverse discrete cosine transform) is performed on the obtained transform coefficient, and a prediction error signal 116 is output.
- the prediction error signal 116 decoded by the internal inverse quantization and inverse transform unit 207 is selected by the internal adder 208 to select the internal mode.
- the prediction image signal 114 of the mode supplied from the selection unit 204 is added. The added signal is input to the internal reference image memory 209 as a decoded image signal 119.
- the internal reference image memory 209 stores the decoded image signal 119 as a reference image.
- the reference image power pixel adaptive prediction unit 201 or the direction prediction unit 202 stored in the internal reference image memory 209 is referred to when the prediction image signal 114 and the like are generated by the fixed mode prediction unit 203.
- Internal prediction loop pixel adaptive prediction unit 201 or direction prediction unit 202 in FIG. 2, fixed mode prediction 203 ⁇ internal mode selection unit 204 ⁇ prediction switching switch 205 ⁇ internal conversion and quantization unit 206 ⁇ internal inverse conversion and inverse quantization
- the processing that flows in the order of section 207 ⁇ internal reference image memory 209) is a single loop when processing is performed for all the prediction modes that can be selected in the small pixel block in the macroblock.
- the encoding control unit 108 first connects the prediction switching switch 205 to the pixel adaptive prediction unit 201, performs 16 internal prediction loops, and selects an optimal mode for each macroblock of pixel adaptive prediction. Determine the combination.
- the sign key control unit 108 connects the prediction switching switch 205 to the direction prediction unit 202, and similarly performs 16 internal prediction loops.
- the sign key control unit 108 connects the prediction switching switch 205 to the fixed mode prediction unit 203, determines which prediction method the fixed mode prediction is predicted from the state of the internal flag, and determines the determined prediction. Make predictions with the method.
- the three modes obtained here are input to the mode selection unit 103, and the optimum mode of the macroblock is selected.
- the image dividing unit 101 divides one frame into a plurality of macroblocks and further divides it into a plurality of small pixel blocks. (Step S2).
- One divided macroblock is input to the intraframe prediction unit 102 as an input image signal 115.
- the mode selection unit 103 initializes an index indicating the mode and a cost (step S3).
- the intra-frame prediction unit 102 uses the input image signal 115 to generate a prediction image signal 114 for one prediction mode that can be selected in the prediction block (step S4).
- the mode selection unit 103 takes the difference between the predicted image signal 114 and the input image signal 115 and generates a prediction error signal 116.
- the cost cost is calculated from the absolute value sum SAD of the prediction mode code amount OH and the prediction error signal 116 (step S5).
- the mode selection unit 103 determines whether or not the calculated cost cost is smaller than the minimum cost min—cost (step S6). If it is smaller (YES), the minimum cost is updated with that cost. At the same time, the code mode at that time is held as the best_mode index (step S7). If the calculated cost cost is greater than the minimum cost min—cost (NO), the index indicating the mode number is incremented, and it is determined whether the index after the increment is the last mode (step S8).
- index is greater than MAX, which is the last number of the mode! /, (YES)
- the best-mode code key mode information and the prediction error signal 116 are sent to the transform and quantization unit 104, and the change is made. Conversion and quantization are performed (step S9).
- the quantized transform coefficient 117 is input to the code processor 107 and entropy-encoded by the code processor 107 together with the prediction information 109 and the prediction switching information 110 (step S10).
- the index is smaller than MAX, which is the last number of the mode (NO)
- the prediction image signal 114 of the sign key mode indicated by the next index is generated (step S4).
- the quantized transform coefficient 117 is input to the inverse quantization and inverse transform unit 105, and is inversely quantized and inverse transformed to be decoded into the prediction error signal 116. Is done.
- the decoded prediction error signal 116 is added by the adder 118 with the best-mode prediction image signal 114 provided from the mode selection unit 103, and the decoded image signal 119 is generated. Made.
- the decoded image signal 119 is stored in the reference image memory 106 as a reference image.
- step Sl l It is determined whether or not the sign of one frame has been completed (step Sl l). If the encoding process has been completed (YES), the next frame is input and the encoding process is performed again. When the encoding process for one frame is not completed (NO), the input signal of the next small pixel block is input to the intra-frame prediction unit 102, and the encoding process is continued.
- the moving image encoding device 100 encodes a moving image.
- prediction pixel calculation is performed using the filter weighting table shown in Fig. 7 in pixel adaptive prediction.
- prediction is performed from the prediction pixel !, and the pixel value after prediction is stored in the memory, this pixel value increases the distance between the reference pixel and the prediction pixel.
- pixel adaptive horizontal prediction (mode 1) generates prediction pixels from left to right. The generated prediction pixel is stored in a memory and used as a reference pixel for the next prediction as needed. As a result, the distance between the reference pixel and the predicted pixel is large, the calculation cost for generating the predicted value of the pixel can be suppressed, and the hardware cost can be reduced.
- the same prediction may be performed for the power 8x8 pixel block and the 16x16 pixel block, and the color difference signal described in detail for the pixel adaptive intra-frame prediction regarding the 4x4 pixel prediction.
- the distance between the reference pixel and the prediction pixel increases, so that the effect of pixel adaptive prediction is high.
- the number of reference pixels used in prediction has been increased according to the distance, but may be reduced to reduce the calculation cost.
- the processing target frame is divided into rectangular blocks of 16 ⁇ 16 pixel size, and the block force at the upper left of the screen is also directed toward the lower right, and sequentially encoded.
- the sign order may be another order. Encoding may be performed from the lower right to the upper left, or may be performed in a screen center force spiral. Upper-right force The process may be performed in the lower-left direction, or processing may be performed by directing power from the periphery of the screen to the center.
- a processing target frame is divided into macro blocks each having a size of 16 ⁇ 16 pixels, and an 8 ⁇ 8 pixel block or a 4 ⁇ 4 pixel block is used as a processing unit for intra-frame prediction.
- the processing target block does not need to be a uniform block shape. It can also be applied to block sizes such as 16x8 pixels, 8x16 pixels, 8x4 pixels, and 4x8 pixels. For example, 8x4 pixel prediction and 2x2 pixel prediction can be realized in the same framework.
- the transform quantization block size need not be a uniform block size, and can be applied to block sizes of 16x8 pixels, 8x16 pixels, 8x4 pixels, 4x8 pixels, and the like. Furthermore, it is possible to select different block sizes within a single macroblock, as long as it is not necessary to have a uniform block size. For example, as shown in Fig. 4E, 8x8 pixel prediction and 4x4 pixel prediction may be mixed in a macroblock. In this case, when the number of divided blocks increases, intra-frame prediction with higher prediction efficiency can be realized by the power of increasing the amount of codes for coding the divided blocks, and the prediction error can be reduced. Therefore, the block size may be selected in consideration of the balance between the code amount of the transform coefficient and the locally decoded image.
- the prediction mode has been described only with intra-frame prediction.
- the present method may be used as an intra-frame prediction method selected in a slice code in an inter-frame code that may be predicted using inter-frame prediction using correlation between screens.
- switching between intra-frame prediction and inter-frame prediction does not need to be performed in units of macro blocks, and may be performed for each 8 ⁇ 8 pixel block, or may be divided for each 8 ⁇ 4 pixel block.
- the transform and quantization unit 104 and the inverse quantization and inverse transform unit 105 are provided.
- transform quantization and inverse quantization and inverse transform are not necessarily performed for all prediction error signals. It is possible to code the prediction error signal as it is without being performed by the code processing unit 107 as it is, or to omit the quantization and inverse quantization processing. Similarly, the conversion process and inverse change It is not necessary to perform the conversion process.
- FIG. 13 shows an outline of the syntax structure used in the present embodiment.
- the syntax mainly consists of three parts: high level syntax (1301), slice level syntax (1304), and macroblock level syntax (1307).
- the high level syntax (1301) is packed with syntax information of higher layers above the slice.
- the slice level syntax (1304) specifies the information required for each slice, and the macro block level syntax (1307) specifies the quantization parameter change value and mode information required for each macro block. Has been.
- the high-level syntax includes sequence parameter set syntax (1302) and picture parameter set syntax (1303), etc. It consists of syntax.
- the slice level syntax (1304) includes a slice header syntax (1305), a slice data syntax (1306), and the like.
- Macroblock level syntax (130 7) is composed of macroblock layer syntax (1308), macroblock prediction syntax (1309), and more! RU
- required syntax information is sequence parameter set syntax (1302), picture parameter set syntax (1303), slice header syntax (1305), and macroblock layer syntax (1308).
- sequence parameter set syntax (1302)
- picture parameter set syntax (1303)
- slice header syntax (1305
- macroblock layer syntax (1308). The syntax of is described below.
- ex_adaptive—intra—in—seq—flag shown in the sequence parameter set syntax of FIG. 14 is a flag indicating whether or not the pixel adaptive prediction is changed for each sequence.
- the flag is TRUE Pixel adaptive prediction and direction prediction can be switched in sequence units.
- the flag is FALSE, pixel adaptive prediction cannot be used in the sequence.
- the ex-adaptive-intra-in-pic-flag shown in the picture parameter set syntax of FIG. 15 is a flag indicating whether or not the pixel adaptive prediction is changed for each picture.
- the flag is TRUE, pixel adaptive prediction and direction prediction can be switched on a picture-by-picture basis.
- the flag force is FALSE, pixel adaptive prediction cannot be used in the picture.
- ex-adaptive-intra-in-slice-flag shown in the slice header syntax of Fig. 16 is a flag indicating whether or not the pixel adaptive prediction is changed for each slice, and the flag is TRUE. Sometimes, pixel adaptive prediction and direction prediction can be switched in slice units. When the flag is FALSE, pixel adaptive prediction cannot be used in the slice.
- the ex-adaptive-intra-flag shown in the macroblock layer syntax in FIG. 17 is a flag indicating whether or not pixel adaptive prediction is used in the macroblock.
- the flag When the flag is TRUE, Use pixel adaptive prediction.
- the flag When the flag is FALSE, direction prediction is used. This flag is used when at least one of ex_adaptive-intra-in-seq-flag, ex-adaptive-intra-in-pic-flag, ex-adaptive-intra-in-slice-flag is TRUE Only valid. Available only when the macroblock prediction type is intraframe prediction.
- ex_direct—intra—in—seq—flag shown in the sequence parameter set syntax of FIG. 18 is a flag indicating whether or not the fixed mode prediction is changed for each sequence.
- the flag is TRUE It is possible to switch the availability of fixed mode prediction on a sequence basis.
- the flag is FALSE, fixed mode prediction cannot be used in the sequence.
- ex-direct-intra-in-pic-flag shown in the picture parameter set syntax of FIG. 19 is a flag indicating whether or not the fixed mode prediction is changed for each picture, and the flag is TRUE.
- the availability of fixed mode prediction can be switched on a picture-by-picture basis.
- the flag is FALSE, fixed mode prediction cannot be used in the picture.
- ex-direct-intra-in-slice-f lag shown in the slice header syntax in Figure 20 is a flag that indicates whether to change the fixed mode prediction for each slice. In some cases, the availability of fixed mode prediction can be switched on a slice-by-slice basis. Is possible. When the flag is FALSE, fixed mode prediction cannot be used within a slice.
- the ex-direct-intra-flag shown in the macroblock layer syntax in Fig. 21A is a flag that indicates whether or not to use fixed mode prediction in the macroblock. When the flag is TRUE Uses fixed mode prediction. When the flag is FALSE, fixed mode prediction is not available. This flag is used when at least one of ex_direct- intra- in- seq-flag, ex- direct- intra- in- pic- flag, ex- direct- intra- in- slice- flag is TRUE Valid only for. Only available when the macroblock prediction type is intraframe prediction.
- FIG. 21B shows the macroblock prediction syntax.
- prediction mode information in the corresponding macroblock is stored.
- the corresponding prediction mode information is set by this syntax.
- the prev- intra4x4- pred- mode- flag shown in this syntax is not used when ex- direct -intra- flag is TRUE!
- the prev—intra4x4—pred—mode—flag is based on the prediction mode (current—mode) estimated from the adjacent prediction block and the prediction mode (pred—mode) actually selected in the prediction target block. This flag indicates whether or not. If they match (TRUE), rem_intra4x4_pred-mode is not used. On the other hand, if they do not match, rem—intra4x4—pred one mode) ⁇ lj is used.
- the rem 1 mtra4x4 1 pred 1 mode represents the amount of deviation between the current 1 mode and the prediction mode (pred-mode) actually selected in the prediction target block.
- the value of is stored.
- ex-direct-intra-flag and ex-adaptive-intra-flag are TRUE, it becomes a code mode that does not send mode information in pixel adaptive prediction.
- the prediction accuracy for a prediction pixel far from the reference pixel is improved, and the prediction error is improved.
- the moving image coding unit 1000 according to the second embodiment shown in FIG. 10 adds the provisional coding Z code amount measuring unit 1001, the coding distortion measuring unit 1002, and the coding.
- a switching switch 1003 is further provided. Since the functions of the intra-frame prediction unit and the mode selection unit are different from those of the first embodiment, reference numbers different from those of the first embodiment are assigned. In the present embodiment, the same components as those in FIG.
- the moving image signal input to the moving image encoding device 1000 is divided into a plurality of small pixel blocks by the screen dividing unit 101, and each block is subjected to intraframe prediction as the input image signal 115.
- Part 1004 is input.
- the intra-frame prediction unit 1004 uses the reference image temporarily stored in the reference image memory 106 to generate the predicted image signal 114 in all prediction modes that can be selected in the macroblock.
- H.264 intra-frame prediction (4x4 pixel prediction (Fig. 4C) or 8x8 pixel prediction (Fig. 4D)
- the intra-frame prediction unit 1004 may perform transformation and quantization, inverse quantization, and inverse transformation.
- the predicted image signal 114 generated by the intra-frame prediction unit 1004 is sent to the mode selection unit 105.
- the mode selection unit 1005 generates a prediction error signal 116 obtained by subtracting the prediction image signal 114 from the input image signal 115.
- the mode selection unit 1005 receives, as inputs, the code amount 1006 accumulated by the provisional encoding Z code amount measurement unit 1001 and the code distortion 1007 calculated by the code distortion measurement unit 1002. Calculate the cost.
- the coding cost is expressed by the following equation (24), where R is the code amount and D is the code distortion.
- ⁇ is given as a constant and is determined based on the quantization width and the value of the quantization parameter.
- a mode is selected on the basis of the code cost thus obtained. Cost J force S smallest The mode giving the value is selected as the optimum mode.
- code amount 1006 and coding distortion 1007 are used.
- a mode may be selected using only the code amount or only code distortion.
- a cost may be created using the activity of the input image signal 115, or a cost function may be created using a quantization width and a quantization meter.
- the sign frequency distortion is weighted by visual frequency characteristics and sensitivity, etc., and the sign key cost is also good.
- the mode selection unit 1005 is connected to the transform and quantization unit 104, and the mode information selected by the mode selection unit 105 and the prediction error signal 116 are input to the transform and quantization unit 104.
- the transform and quantization unit 104 transforms and quantizes the prediction error signal 116 to output a quantized transform coefficient 117.
- the conversion coefficient 117 is sent to the temporary code key changeover switch 1003.
- the temporary code key switching switch 1003 sends the output of the transform and quantization unit 104 to the temporary encoding Z code amount measuring unit 1001. Connecting.
- the conversion coefficient 117 is input to the temporary code Z code amount measuring unit 1001, and temporary entropy encoding is performed.
- the provisional encoding Z code amount measuring unit 1001 cumulatively adds the code amounts, estimates the total code amount generated in the actual code, and does not output the code data.
- the code amount 1006 measured by the temporary code Z code amount measuring unit 1001 is sent to the code amount distortion measuring unit 1002.
- the sign distortion measurement unit 1002 receives as input the decoded image signal 119 generated by the sum of the prediction error signal 116 decoded by the inverse quantization and inverse conversion unit 105 and the prediction image signal 114.
- the input image signal 115 is simply received, and a square error (sign distortion 1007) between the decoded image signal 119 and the input image signal 115 is calculated.
- Coding distortion measuring section 1002 sends code amount 1006 received as input and code distortion 1007 obtained by calculation to mode determining section 1005.
- Temporary coding loop (intra-frame prediction unit 1004 in FIG. 10 ⁇ mode selection unit 1005 ⁇ transformation and quantization unit 104 ⁇ temporary code key switching switch 1003 ⁇ temporary code key Z code amount measurement unit 1001 ⁇ code key (Processing that flows in the order of ⁇ distortion measurement unit 1002) is a single loop when processing is performed for one mode that can be selected by the macroblock. mode If there are 10 types of force S, the provisional coding loop is repeated 10 times. When the temporary encoding loop is completed for all modes selectable in the macro block, the mode selection unit 1005 sets the temporary code key flag of the code key control unit 108 to FALSE.
- the temporary code switching switch 1003 connects the output of the transform and quantization unit 104 to the code processing unit 107.
- the mode determination unit 1005 calculates the code cost for all modes, and among them, the transform coefficient 117 and the mode information of the mode that gives the smallest code cost are sent to the encoding processing unit 107.
- the encoding processing unit 107 actually encodes the quantized transform coefficient 117 by the method specified in the input mode information.
- the code processing unit 107 may perform the t processing, in which the code data stored in the memory is called and copied without executing the encoding.
- the input image signal 115 of the next block is input to the intra-frame prediction unit 1004, and the coding of the next block is performed.
- the code control unit 108 performs feedback control of the generated code amount, quantization characteristic control, mode control, and the like, and performs rate control for controlling the generated code amount, control of the prediction unit, and control of the entire encoding.
- each unit described above can be realized by a program stored in a computer.
- a moving picture coding method that works in the present embodiment will be described with reference to the intra-frame prediction unit 1004 shown in FIG. In the present embodiment, the same components as those in FIG.
- the signal is input to the pixel adaptation prediction unit 201, the direction prediction unit 202, and the fixed mode prediction unit 203.
- a prediction image signal 114 corresponding to the prediction mode and the prediction block shape is generated by the pixel adaptive prediction unit 201, the direction prediction unit 202, and the fixed mode prediction unit 203.
- Predicted image signals 114 predicted by the prediction units 201, 202, and 203, respectively.
- the prediction mode information is input to the internal mode selection unit 1104.
- the internal mode selection unit 1104 generates a prediction error signal 116 obtained by subtracting the prediction image signal 114 from the input image signal 115.
- the internal mode selection unit 1104 is calculated by the internal temporary code amount Z code amount measuring unit 1101 and is calculated by the internal code amount 1105 and the internal code amount distortion measuring unit 1102, and the internal coding distortion 1106 is calculated. Based on the above, the inner code key cost is calculated using equation (24), and the mode that gives the smallest coding cost J is selected as the optimum mode.
- the mode may be determined using only the force code amount using the inner code amount 1105 and the inner code distortion 1006 or using only the code distortion.
- the cost may be created using the activity of the input image signal, or the cost function may be created using the quantization width and the quantization parameter.
- the coding distortion may be weighted by adding visual frequency characteristics and sensitivity to the coding distortion.
- the selected mode force is a mode predicted by the pixel adaptive prediction unit 201 or a mode predicted by the force fixed mode prediction unit 203 that is a mode predicted by the direction prediction unit 202.
- the internal mode selection unit 1104 transmits mode information indicating the force to the sign control unit 108.
- the sign key control unit 108 controls the prediction switching switch 205 according to the given mode information.
- the mode information and prediction error signal selected by internal mode selection section 1104 are input to internal conversion and quantization section 206.
- the transform / quantization unit 206 performs orthogonal transform on the input prediction error signal 116 to generate transform coefficient data 117.
- the transform and quantization unit 206 quantizes the transform coefficient.
- the quantization parameter required for quantization is set by the code control unit 108.
- the sign key control unit 108 refers to the temporary sign key flag. If this flag is TRUE, the sign key control unit 108 switches the internal temporary code key switch 1103 to the internal temporary code key Z code amount measuring unit 1101. Connecting. As a result, the conversion coefficient and mode information of the mode are input to the internal temporary code Z code amount measurement unit 1101. The internal tentative encoding Z code amount measurement unit 1101 performs entropy encoding of the input data. In this case, the internal temporary encoding Z code amount measurement unit HOI
- the code amount 1105 is sent to the internal coding distortion measurement unit 1102.
- the inner code distortion measurement unit 1102 adds the prediction error signal 116 decoded by the inner inverse quantization and inverse transformation unit 207 and the predicted image signal 114, and generates a decoded image signal 1 19 Is received as an input, and a square error (internal code distortion 1106) between the decoded image signal 119 and the input image signal 115 is calculated.
- the internal code distortion measurement unit 1102 sends the internal code amount 1105 received as input and the internal code distortion 1106 obtained by calculation to the internal mode determination unit 1104.
- Internal tentative coding loop (respective prediction units in FIG. 11 ⁇ internal mode selection unit 1104 ⁇ prediction switching switch 205 ⁇ internal conversion and quantization unit 206 ⁇ internal tentative code switching switch 1103 ⁇ internal tentative coding
- the processing that flows in the order of the Z code amount measurement unit 1101 ⁇ the internal coding distortion measurement unit 1102) is performed once when processing is performed for one mode that can be selected in a small pixel block below the macro block. It becomes the loop of. If there are 10 modes, the inner temporary coding loop is repeated 10 times.
- the internal mode selection unit 1104 sets the temporary encoding flag of the encoding control unit 108 to FALSE.
- the internal temporary code switch 1103 When the temporary code flag is set to FALSE in the code controller 108, the internal temporary code switch 1103 outputs the output of the internal transformer / quantizer 206 to the intra-frame predictor 1004. Connect to outside. At this time, the coding cost for all modes is calculated in the internal mode determination unit 1104, and the transform coefficient 117 and mode information of the mode that gives the smallest coding cost among them are stored in the intra-frame prediction unit 1004. Sent outside.
- the image dividing unit 101 divides one frame of the moving image into a plurality of macro blocks. One that is divided The macroblock is input to the intraframe prediction unit 102 as the input image signal 115 (step S002).
- the mode selection unit 1005 and the internal mode selection unit 1104 initialize the index indicating the mode and the cost (step S003).
- the intra-frame prediction unit 1004 uses the input image signal 115 to generate a predicted image signal 114 for one mode that can be selected by the prediction block (step S004).
- the mode selection unit 1005 calculates a difference between the predicted image signal 114 and the input image signal 115 and generates a prediction error signal 116.
- the transform and quantization unit 104 transforms and quantizes the prediction error signal 116, and the quantized transform coefficient 117 is input to the temporary code Z code amount measurement unit 1001.
- the temporary code Z Z code amount measuring unit 1001 temporarily encodes the conversion coefficient (step S005), and cumulatively adds the code amount.
- the transform coefficient is locally decoded, and a sign error measurement unit 1002 calculates a square error between the input image signal and the decoded image signal. Based on the obtained code amount R and code distortion D, the encoding cost is calculated (step S006).
- the mode selection unit 1005 determines whether or not the calculated coding cost cost is smaller than the minimum coding cost min_cos (step S007).
- the minimum code key cost is updated, the corresponding code key mode is held as the best-mode index, and the temporary code key data is stored in the temporary memory (step S008).
- the mode selection unit 1005 increments the index indicating the mode number and increments it. It is determined whether the subsequent index is the last mode (step S009).
- index is greater than MAX, which is the last number of the mode! /, (YES)
- index is the best-mode code key mode information, prediction information 109, prediction switching information 110 and code key processing unit 107 Is entropy-encoded (step S010).
- the index is smaller than MAX, which is the last number of the mode (NO)
- the prediction image signal 115 of the encoding mode indicated by the next index is generated (step S004).
- the quantized transform coefficient 117 is input to the inverse quantization and inverse transform unit 105, and inverse quantization and inverse transform are performed.
- the decoded prediction error signal 116 is added by the adder 118 with the best-mode prediction image signal 114 provided from the mode selection unit 1004, and the decoded image signal 119 is added to the reference image memory 10. Saved to 6.
- step SI 1 It is determined whether or not one frame has been encoded.
- the input image signal of the next frame is input to the screen dividing unit 101, and the encoding process of the next frame is performed. If the encoding process for one frame has not been completed (NO), the input signal of the next macroblock is input and the encoding process is continued.
- the moving image encoding method according to the present embodiment is executed by the moving image encoding apparatus 1000.
- the code amount can be appropriately allocated for each block by using the temporary code and the index, and the code efficiency is improved.
- a first intra-frame prediction unit 2202 and a second intra-frame prediction unit 2203 are further provided. Since the function of the screen dividing unit 2201 is different from that of the second embodiment, different reference numerals are given. In the embodiment of FIG. 10, the same reference numerals are given to the same components as those of the embodiment of FIG. 1, and description thereof is omitted.
- first intra-frame prediction unit 2202 and second intra-frame prediction unit 2203 differ only in the prediction block size, and the prediction method is shown in Fig. 11. This is the same as the intra-frame prediction unit 1004.
- the screen dividing unit 2201 divides the input moving image signal into small pixel blocks. At this time, the macroblock is divided into different block shapes. One is divided into a block shape composed of 16 4x4 pixel blocks, and the other is divided into a block shape composed of 4 8x8 pixel blocks.
- the divided blocks are input as input image signals 115 to the first intra-frame prediction unit 2202 and the second intra-frame prediction unit 2203.
- the first intra-frame prediction unit 2202 performs 4 ⁇ 4 pixel prediction.
- the second intra-frame prediction unit 2203 performs 8 ⁇ 8 pixel prediction. That is, the second intra-frame prediction unit 2203 performs one 8 ⁇ 8 pixel prediction four times.
- a prediction image signal based on 4x4 pixel prediction and a prediction image signal based on 8x8 pixel prediction are input to the mode selection unit 1005.
- the mode selection unit 1005 is a code encoder according to a predetermined method. The prediction mode with the lowest encoding cost is selected from the calculated encoding costs.
- the decoded image corresponding to the selected mode is stored in the reference image memory 106 via the transform / quantization unit 104 and the inverse quantization / inverse transform unit 105.
- the encoding control unit 108 connects the temporary encoding switching switch 1003 to the encoding processing unit 107 and performs encoding.
- the information of the force encoded by the 8x8 pixel block force 4x4 pixel prediction and the force encoded by the 8x8 pixel prediction is also encoded at the same time. This information is encoded by the encoding processing unit 107 as 4-bit data for the macroblock.
- the ex_adaptive—blocksiz e— in— seq— flag shown in the sequence parameter set syntax of Figure 23 is a flag that indicates whether the block size switching prediction is changed for each sequence. When this flag is TRUE, Whether block size switching prediction can be used can be switched in sequence units. When the flag is FALSE, block size switching prediction cannot be used in the sequence.
- ex-adaptive- blocksi ze- in-pic-flag shown in the picture parameter set syntax in Fig. 24 is a flag indicating whether or not to change block size switching prediction for each picture.
- TRUE the availability of block size switching prediction can be switched on a picture-by-picture basis.
- FALSE block size switching prediction cannot be used in the picture.
- ex—adaptive—blocksize—in—slice—flag shown in the slice header syntax of FIG. 25 is a flag indicating whether or not to change the block size switching prediction for each slice, and this flag is TRUE. At this time, it is possible to switch the availability of block size switching prediction in units of slices. When the flag is FALSE, block size switching prediction cannot be used in a slice! /.
- ex-adaptive-blocksize-mode shown in the macroblock layer syntax in FIG. 26 is 4-bit data and corresponds to block size switching information.
- FIG. 28A shows 8 ⁇ 8 pixel block power in the order of A, B, C, and D ⁇ power that is ⁇ 4 pixel prediction, and whether it is 8 ⁇ 8 pixel prediction.
- ex-adaptive-blocksize-mode indicates 1100
- blocks A and B are encoded with 8x8 pixel prediction and blocks C and D are encoded with 4x4 pixel prediction as shown in Fig. 28B.
- FIG. 27 shows an example in which block size switching information is added to the macroblock type.
- macroblock prediction is performed only with 4 ⁇ 4 pixel prediction.
- mb_type 1
- macroblock prediction is performed only with 8x8 pixel prediction.
- mb If the type is 1 to 5, block size switching is predicted.
- Intra-A-B-C-D shows which block force S8x8 pixel prediction is used.
- Figure 28A shows the block positions of A, B, C, and D, respectively. For example, when mb-type is 3, it indicates that blocks A and B are predicted with 8x8 pixel prediction.
- provisional code processing is actually performed in units of 8x8 pixels.
- the mode with the lowest coding cost that also obtains the code amount and code distortion power of the key data is selected, and the coded data is output based on the selected mode.
- mode selection is performed by switching the prediction block size for each 8x8 pixel block in the macro block! /, So that code efficiency can be increased with high code efficiency according to the prediction method.
- the lock shape can be selected, that is, suitable encoding can be performed according to the contents of the pixel block.
- the decoded image signal is generated only in the selected mode in the loop for prediction mode determination. It does not necessarily have to be executed.
- a video decoding device corresponding to the above video encoding device will be described.
- the code data transmitted from the video encoding apparatus and sent via the transmission system or the storage system is once stored in the input buffer 901 and multiplexed and separated. After being separated into a plurality of pixel blocks based on the syntax for each frame by the device 902, they are input to the code string decoding unit 903.
- the code sequence decoding unit 903 sequentially decodes the code sequence of each status of the code data for each of the high level syntax, slice level syntax, and macroblock level syntax according to the syntax structure shown in FIG. .
- the quantized transform coefficient, quantization matrix, quantization meter, prediction mode information, prediction switching information, etc. are restored.
- the quantized transform coefficient is restored and the inverse quantization and inverse transform unit 904 performs inverse quantization using the quantization matrix and quantization parameter of each piece of information.
- inverse orthogonal transform for example, inverse discrete cosine transform.
- the inverse orthogonal transform has been described.
- the inverse quantization and inverse transform unit 904 performs the corresponding inverse quantization and inverse wavelet transform or the like. Also good.
- the coefficient converted by the inverse quantization and inverse conversion unit 904 is sent to the adder 905 as an error signal 911.
- the adder 905 adds the prediction signal 916 output from the intra-frame prediction unit 907 and the error signal 911, and inputs the addition signal to the reference image memory 906 as a decoded signal 912.
- the decoded signal 912 is further sent to the output buffer 913 and output at a timing managed by the decoding control unit 908.
- Prediction mode information 909 and prediction switching information 910 decoded by code string decoding section 903 Input to the intra-frame prediction unit 907.
- the reference signal 914 that has already been encoded is read from the reference image memory 906 to the intra-frame prediction unit 907.
- the intra-frame prediction unit 907 generates a prediction signal 916 based on the input information and inputs it to the adder 905.
- the decoding key control unit 908 controls the input buffer 901 and the output buffer 913, controls the decoding timing, and the like.
- the moving picture decoding method according to the present embodiment will be described with reference to FIG. 30 by the intra-frame prediction unit 907 of the moving picture decoding apparatus 900.
- the same components as those of the embodiment of FIG. 29 are denoted by the same reference numerals and description thereof is omitted.
- the prediction switching switch 3001 is operated based on the prediction switching information 910. Specifically, ex-adaptive-intra-flag and ex-direct-intra-flag are input to the prediction control unit 3007 as prediction switching information.
- ex—adaptive—intra-flag is a flag indicating whether the image is predicted by pixel adaptive prediction or directional prediction.
- the prediction switching switch 3001 is set to the pixel adaptive prediction unit. Connect to 300-2.
- the flag is FALSE, the prediction switching switch 3001 is connected to the direction prediction unit 3003.
- ex- direct- intra- flag is a flag that indicates whether fixed mode prediction has been performed! /, and has not been performed. If this flag is TRUE, the ex-adaptive-intra-flag state is set. Regardless, the prediction switching switch 3001 is connected to the fixed mode prediction unit. When the prediction switching switch 3001 is switched according to the respective information, the reference signal 914 is input to the intra-frame prediction unit 907.
- the reference signal 914 is input to the prediction unit corresponding to the flag via the prediction switching switch 3001.
- the prediction control unit 3007 inputs prediction mode information 909 to the pixel adaptive prediction unit 3002. Based on this information, the pixel adaptive prediction unit 3002 performs pixel adaptive prediction and generates a prediction signal 913.
- the prediction control unit 3007 inputs the prediction mode information 909 to the direction prediction unit 3003. Based on this information, the direction prediction unit 3003 Measurement is performed and a prediction signal 913 is generated.
- prediction control unit 3007 inputs prediction mode information 909 to fixed mode prediction unit 3004.
- the fixed mode prediction 3004 performs a fixed mode prediction and generates a prediction signal 913.
- the prediction signal 916 that has been predicted by the prediction U ⁇ 3002, 3003, 3004 is output to the outside of the intra-frame prediction unit 907.
- each prediction unit 3002, 3003, 3004 will be described. These prediction units are stored in the reference image memory 906 and perform prediction of a block to be predicted using the already decoded reference signal 914. There are nine prediction modes, each with a different prediction direction of 22.5 degrees as shown in Figure 5A. Mode 0 to mode 8 are specified, and mode 2 is DC prediction in the direction prediction unit 3003. The names of the direction prediction mode performed in the direction prediction unit 3003 and the pixel adaptive prediction mode performed in the pixel adaptive prediction unit 3002 are shown in FIG. The relationship between the prediction block of 4x4 pixel prediction and the reference pixel is shown in FIG. 5B. Pixels from uppercase A to M are reference pixels, and pixels from lowercase a to p are prediction pixels.
- the direction prediction unit 3003 uses a prediction method that copies the prediction value interpolated from the reference pixel in the prediction direction shown in FIG. 5A.
- Equation (3) is used as the prediction value generation method when mode 0 (vertical prediction) is selected. This mode can only be selected when reference pixels A through D are available. In this prediction method, as shown in FIG. 5C, the luminance values of the reference pixels A to D are copied as they are in the vertical direction and compensated as predicted values.
- Prediction modes For prediction methods other than 0 and 2, almost the same framework is used. Interpolated values are generated from reference pixels that can be used in the prediction direction, and the values are copied according to the prediction direction. Make a prediction that [0153]
- the pixel adaptive prediction unit 3002 will be described. The prediction mode is the same as described in Fig. 5A.
- the pixel adaptive prediction unit 3002 performs prediction in which the number of reference pixels to be used changes according to the distance between the prediction pixel and the reference pixel, and the prediction value changes in units of pixels. By changing the weighting table of the reference pixels to be used according to the distance, it is possible to generate more accurate predicted values.
- FIG. 6A shows the relationship between the reference pixel and the prediction block.
- the reference pixels and prediction pixels shown in Fig. 6A and Fig. 5B have a one-to-one correspondence, but different indexes are given to these pixels to facilitate explanation of the prediction formula.
- the reference pixels are xOO force and up to x08, and 13 evens of x09, xl 8, x27, and x36.
- the total number of pixels is 16 even from xlO force xl3, xl9 force x22, x28 force x31, x37 force x40.
- the 16 pixels from xl4 force to xl7, x23 force to x26, x32 force to x35, and x41 force to x44 are prediction auxiliary pixels, and are used to improve prediction accuracy.
- FIG. 6B shows the prediction method for pixel adaptive vertical prediction. As can be seen from FIG. 6B, as the distance between the reference pixel and the target prediction pixel increases, the predicted value is determined using more reference pixels.
- the predicted value generation method will be specifically described.
- the prediction pixel is calculated using Eq. (5).
- n indicates an index corresponding to the predicted pixel position (xlO to xl3, xl9 force, etc. x22, x28 force, etc. x31, x37 force, etc. x40) shown in FIG. 6A.
- equation (5) is expanded according to the distance between the reference pixel and the prediction pixel, the prediction equation becomes equation (7).
- L represents the distance between the reference pixel and the prediction pixel.
- Vi is an index determined according to the corresponding prediction mode.
- hi represents the filter coefficient, and the number of taps varies with L.
- the prediction pixel that is one distance away from the reference pixel in the prediction direction uses three reference pixels.
- a prediction pixel that is two distances away from the reference pixel in the prediction direction is predicted by Equation (9) using five reference pixels.
- a prediction pixel that is three distances away from the reference pixel in the prediction direction is predicted by Equation (10) using seven reference pixels.
- Prediction pixels that are four distances away from the reference pixel in the prediction direction are predicted by Equation (10) using nine reference pixels.
- Pixel adaptive horizontal prediction (mode 1) is predicted using Eq. (14).
- Pixel adaptive upper left prediction (mode 2) is predicted using equation (15).
- Pixel adaptive orthogonal lower left prediction (mode 3) performs prediction using equation (16).
- Pixel adaptive orthogonal lower right prediction (mode 4) is predicted using Eq. (17).
- Pixel-adapted vertical left prediction (mode 5) is predicted using equation (18).
- Pixel adaptive horizontal prediction (mode 6) is predicted using Eq. (19).
- Pixel-adapted vertical left prediction (mode 7) is predicted using equation (20).
- Pixel-adapted horizontal prediction (mode 8) is predicted using Eq. (21).
- the number of reference pixels, and a weighting table can be determined by substituting and developing necessary predicted images.
- the reference pixel is compensated as in the pixel adaptive vertical prediction. For example, when x37 is predicted by pixel adaptive horizontal prediction, ⁇ 45 reference pixels cannot be used. Therefore, this reference pixel is filled with ⁇ 36.
- the predicted image generation method is executed by the pixel adaptive prediction unit 3002.
- the fixed mode prediction unit 3004 will be described in detail.
- the fixed mode prediction unit 3004 performs prediction of mode information, and performs pixel adaptive prediction or direction prediction in a predetermined prediction mode.
- mode information of a prediction block adjacent to the prediction target block is used.
- Fig. 8 shows the relationship between a prediction target block and adjacent blocks in the case of 4x4 pixel prediction.
- prediction The left adjacent block of the target block C is A, and the upper adjacent block is B.
- ⁇ prev-left-mode and prev-upper-mode the prediction mode of the block is determined by Equation (23).
- current—mode is the prediction mode of the prediction block.
- the function min (A, B) returns the smaller of A and B. Since the prediction mode information of the prediction block is predicted from the prediction mode information of the surrounding blocks in this way, it is one of the code modes that can greatly reduce the amount of code for encoding the prediction mode information. It is.
- Whether to generate a prediction image using the pixel adaptive prediction unit 201 or to generate a prediction image using the direction prediction unit 202 is determined by the ex-direct-intra flag, and the encoding control unit 108 It is determined based on ex-direct-intra flag information given by.
- the prediction mode information of the current prediction block is determined based on the prediction mode information of adjacent prediction blocks A and B.
- the determination may be made based on prediction mode information of surrounding prediction blocks. For example, using the prediction mode information of the right block of prediction block B, the upper block of prediction block B, the left block of prediction block A, and the upper block of prediction block A, The prediction mode of the prediction block may be determined. For example, the prediction mode with the highest frequency among the prediction modes of the peripheral prediction block, the median value of the prediction mode of the peripheral prediction block, or the average value of the prediction modes of the peripheral prediction block is determined as the prediction mode of the current prediction block. You can do it.
- a decryption apparatus 3100 according to the second embodiment will be described with reference to FIG.
- the intra-frame prediction unit 907 is changed to a first intra-frame prediction unit 3102 and a second intra-frame prediction unit 3103 with respect to the fourth embodiment.
- a new block size switch 3101 has been introduced.
- the same components as those in FIG. 29 are denoted by the same reference numerals and description thereof is omitted.
- Prediction mode information 909, prediction switching information 910, block size switching information 915, and the like decoded by the code string decoding unit 903 are sent to the block size switching switch 3101 and the decoding key control unit 908.
- the decoding control unit 908 connects the block size switching switch 3101 to an appropriate end point based on the block size switching information.
- the block size is determined based on the ex-adaptive-block size-mode power given from the code string decoding unit 903.
- This information is composed of 4 bits, and indicates whether 4x4 pixel prediction or 8x8 pixel prediction is performed within 8x8 pixels in the macroblock. Bits are assigned in the order of upper left force and lower right, in order of A, B, C, D, as shown in Figure 28A. ex—adaptive —1) 10.1 ⁇ 126—1110 (When the value of 16 is 1100, as shown in Figure 28B, 8x8 pixel prediction is assigned to blocks A and B, and blocks C and D Is assigned a 4x4 pixel prediction!
- decoding control section 903 connects the output end of block size switching switch 3101 to first intra-frame prediction section 3102.
- the first intra-frame prediction unit 3102 performs 4 ⁇ 4 pixel prediction on the four 4 ⁇ 4 pixel blocks, and sends the predicted image to the adder 905.
- the decoding key control unit 908 connects the output terminal of the block size switching switch 3101 to the second intra-frame prediction unit 3103.
- Second intra-frame prediction section 3103 performs 8 ⁇ 8 pixel prediction on one 8 ⁇ 8 pixel block, and sends the predicted image to adder 905.
- the first intra-frame prediction unit 3102 and the second intra-frame prediction unit 3103 perform prediction of the determined block size based on the prediction switching information input in the internal structure equal to that in FIG. !, Output a prediction signal.
- the decoding key control unit 908 controls the block size switching switch in addition to the control of the input buffer 901 and the output buffer 913 and the control of the decoding timing.
- the syntax of the block size switching information required when switching the prediction information in the present embodiment is the same as in FIGS. You can get the information from the macroblock type that doesn't get the block size switching information from the macroblock layer.
- FIG. 27 is an example of obtaining block size switching information from the macroblock type.
- mb_type When mb_type is 0, macroblock prediction is performed only with 4x4 pixel prediction.
- mb—type macroblock prediction is performed with 8x8 pixel prediction only.
- Intra-A-B-C-D shows which block uses 8x8 pixel prediction.
- Figure 28A shows the block positions of A, B, C, and D, respectively.
- the prediction mode to be used is set for each sequence, each slice, or each macroblock. Because it can be changed, it is possible to generate a predicted image that also considers subjective image quality.
- the present invention can be applied to encoding and decoding of moving images, still images, voices, and the like in various fields such as video, audio devices, mopile devices, broadcasting, information terminals, and networks.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Description
Claims
Priority Applications (6)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
BRPI0619193-2A BRPI0619193A2 (pt) | 2005-11-30 | 2006-11-27 | método de codificação de imagem / decodificação de imagem, aparelho de codificação de imagem / decodificação de imagem |
EP06833410A EP1965589A1 (en) | 2005-11-30 | 2006-11-27 | Image encoding/image decoding method and image encoding/image decoding apparatus |
CA 2631336 CA2631336A1 (en) | 2005-11-30 | 2006-11-27 | Image encoding/image decoding method, image encoding/image decoding apparatus |
CN2006800512728A CN101361370B (zh) | 2005-11-30 | 2006-11-27 | 图像编码/图像解码方法以及图像编码/图像解码装置 |
AU2006320064A AU2006320064B2 (en) | 2005-11-30 | 2006-11-27 | Image encoding/image decoding method and image encoding/image decoding apparatus |
JP2007547928A JP5044415B2 (ja) | 2005-11-30 | 2006-11-27 | 画像符号化/画像復号化方法及び画像符号化/画像復号化装置 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2005-347121 | 2005-11-30 | ||
JP2005347121 | 2005-11-30 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2007063808A1 true WO2007063808A1 (ja) | 2007-06-07 |
Family
ID=38092141
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2006/323606 WO2007063808A1 (ja) | 2005-11-30 | 2006-11-27 | 画像符号化/画像復号化方法及び画像符号化/画像復号化装置 |
Country Status (10)
Country | Link |
---|---|
US (1) | US8149910B2 (ja) |
EP (1) | EP1965589A1 (ja) |
JP (1) | JP5044415B2 (ja) |
KR (1) | KR100977101B1 (ja) |
CN (1) | CN101361370B (ja) |
AU (1) | AU2006320064B2 (ja) |
BR (1) | BRPI0619193A2 (ja) |
CA (1) | CA2631336A1 (ja) |
RU (2) | RU2369038C1 (ja) |
WO (1) | WO2007063808A1 (ja) |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2009246975A (ja) * | 2008-03-28 | 2009-10-22 | Samsung Electronics Co Ltd | 画像のイントラ予測モード情報の符号化及び/又は復号化方法及び装置 |
WO2010084692A1 (ja) * | 2009-01-22 | 2010-07-29 | 株式会社エヌ・ティ・ティ・ドコモ | 画像予測符号化装置、方法及びプログラム、画像予測復号装置、方法及びプログラム、並びに、符号化・復号システム及び方法 |
JP2010529811A (ja) * | 2007-06-12 | 2010-08-26 | トムソン ライセンシング | スライスデータ用のマルチパスビデオのシンタックス構造をサポートする方法及び装置 |
JP2010283495A (ja) * | 2009-06-03 | 2010-12-16 | Fujitsu Ltd | 動画像符号化装置及び動画像符号化方法 |
JP2011166218A (ja) * | 2010-02-04 | 2011-08-25 | Victor Co Of Japan Ltd | 画像信号符号化装置、画像信号符号化方法、画像信号復号装置、及び、画像信号復号方法 |
WO2012011340A1 (ja) * | 2010-07-20 | 2012-01-26 | ソニー株式会社 | 画像処理装置及び画像処理方法 |
JP2012034091A (ja) * | 2010-07-29 | 2012-02-16 | Ntt Docomo Inc | 画像予測符号化装置、方法およびプログラム、並びに、画像予測復号装置、方法およびプログラム |
WO2012035640A1 (ja) * | 2010-09-16 | 2012-03-22 | 株式会社 東芝 | 動画像符号化方法及び動画像復号化方法 |
US8204136B2 (en) | 2008-03-07 | 2012-06-19 | Canon Kabushiki Kaisha | Image encoding apparatus |
JP2012521735A (ja) * | 2009-04-20 | 2012-09-13 | ドルビー ラボラトリーズ ライセンシング コーポレイション | 多層映像配信のための適応補間フィルタ |
JP2013507087A (ja) * | 2009-10-05 | 2013-02-28 | トムソン ライセンシング | 映像符号化及び復号化における色度成分の予測画素の適応フィルタリング |
RU2488235C2 (ru) * | 2008-10-22 | 2013-07-20 | Ниппон Телеграф Энд Телефон Корпорейшн | Способ масштабируемого кодирования видео, устройство масштабируемого кодирования видео, программа масштабируемого кодирования видео и машиночитаемый носитель записи, сохраняющий программу |
JP2013223149A (ja) * | 2012-04-17 | 2013-10-28 | Nippon Hoso Kyokai <Nhk> | 画像符号化装置、画像復号装置、画像符号化プログラム及び画像復号プログラム |
WO2014050971A1 (ja) * | 2012-09-28 | 2014-04-03 | 日本電信電話株式会社 | イントラ予測符号化方法、イントラ予測復号方法、イントラ予測符号化装置、イントラ予測復号装置、それらのプログラム並びにプログラムを記録した記録媒体 |
US9369712B2 (en) | 2010-01-14 | 2016-06-14 | Dolby Laboratories Licensing Corporation | Buffered adaptive filters |
JP2016519898A (ja) * | 2013-04-08 | 2016-07-07 | ジーイー ビデオ コンプレッション エルエルシー | コンポーネント間予測 |
US9554154B2 (en) | 2010-07-31 | 2017-01-24 | M & K Holding Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9774882B2 (en) | 2009-07-04 | 2017-09-26 | Dolby Laboratories Licensing Corporation | Encoding and decoding architectures for format compatible 3D video delivery |
JP2020509714A (ja) * | 2017-04-04 | 2020-03-26 | ホアウェイ・テクノロジーズ・カンパニー・リミテッド | ピクチャブロックのイントラ予測に使用される参照サンプルの処理 |
JPWO2019189904A1 (ja) * | 2018-03-30 | 2021-02-12 | 日本放送協会 | イントラ予測装置、画像符号化装置、画像復号装置、及びプログラム |
Families Citing this family (94)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101222641B (zh) * | 2007-01-11 | 2011-08-24 | 华为技术有限公司 | 帧内预测编解码方法及其装置 |
KR101365569B1 (ko) * | 2007-01-18 | 2014-02-21 | 삼성전자주식회사 | 인트라 예측 부호화, 복호화 방법 및 장치 |
US8472522B2 (en) | 2007-02-23 | 2013-06-25 | Nippon Telegraph And Telephone Corporation | Video encoding method and decoding method, apparatuses therefor, programs therefor, and storage media which store the programs |
AU2012204067B2 (en) * | 2007-06-04 | 2015-04-09 | Ntt Docomo, Inc. | Image prediction encoding device, image prediction decoding device, image prediction encoding method, image prediction decoding method, image prediction encoding program, and image prediction decoding program |
JP5188875B2 (ja) * | 2007-06-04 | 2013-04-24 | 株式会社エヌ・ティ・ティ・ドコモ | 画像予測符号化装置、画像予測復号装置、画像予測符号化方法、画像予測復号方法、画像予測符号化プログラム、及び画像予測復号プログラム |
WO2009004985A1 (ja) * | 2007-06-29 | 2009-01-08 | Sharp Kabushiki Kaisha | 画像符号化装置、画像符号化方法、画像復号装置、画像復号方法、プログラム、および記録媒体 |
JP2009094828A (ja) * | 2007-10-10 | 2009-04-30 | Hitachi Ltd | 画像符号化装置及び画像符号化方法、画像復号化装置及び画像復号化方法 |
BRPI0818649A2 (pt) * | 2007-10-16 | 2015-04-07 | Thomson Licensing | Métodos e aparelho para codificação e decodificação de vídeo em superblocos geometricamente particionados. |
KR101946376B1 (ko) * | 2007-10-16 | 2019-02-11 | 엘지전자 주식회사 | 비디오 신호 처리 방법 및 장치 |
KR101375664B1 (ko) * | 2007-10-29 | 2014-03-20 | 삼성전자주식회사 | 영상의 디퓨전 특성을 이용한 영상 부호화/복호화 방법 및장치 |
KR101439847B1 (ko) | 2008-01-02 | 2014-09-16 | 삼성전자주식회사 | 부호화 정보의 압축률 향상을 이용한 영상 부호화, 복호화방법 및 장치 |
KR101456279B1 (ko) * | 2008-01-03 | 2014-11-04 | 한국전자통신연구원 | 참조 블록 파형 정보에 기반하여 인트라 영상을 부호화하는 장치 및 복호화 하는 장치 |
US8295356B2 (en) * | 2008-03-07 | 2012-10-23 | International Business Machines Corporation | Method and system for coding mode selection in video compression systems |
JP5206070B2 (ja) * | 2008-03-28 | 2013-06-12 | 富士通株式会社 | 復号装置および復号方法 |
WO2009122463A1 (ja) * | 2008-03-31 | 2009-10-08 | 富士通株式会社 | 画像データ圧縮装置、復元装置、圧縮方法、復元方法及びプログラム |
WO2009136066A1 (fr) * | 2008-04-15 | 2009-11-12 | France Telecom | Prediction d'images par determination prealable d'une famille de pixels de reference, codage et decodage utilisant une telle prediction |
CN102017605A (zh) * | 2008-04-22 | 2011-04-13 | 韩国科亚电子股份有限公司 | 活动影像抖动修正装置及方法 |
KR101379187B1 (ko) * | 2008-06-23 | 2014-04-15 | 에스케이 텔레콤주식회사 | 블록 변환을 이용한 인트라 예측 방법 및 장치와 그를이용한 영상 부호화/복호화 방법 및 장치 |
CN102197650A (zh) * | 2008-10-27 | 2011-09-21 | 松下电器产业株式会社 | 图像编码方法、图像编码装置以及摄像系统 |
JP5394212B2 (ja) * | 2008-12-19 | 2014-01-22 | トムソン ライセンシング | データを挿入する方法、挿入されたデータを読み出す方法 |
KR101567974B1 (ko) * | 2009-01-05 | 2015-11-10 | 에스케이 텔레콤주식회사 | 블록 모드 부호화/복호화 방법 및 장치와 그를 이용한 영상부호화/복호화 방법 및 장치 |
JP5170800B2 (ja) * | 2009-01-15 | 2013-03-27 | ルネサスエレクトロニクス株式会社 | 画像処理装置、デコード方法、フレーム内復号装置、フレーム内復号方法、及びフレーム内符号化装置 |
KR20100095992A (ko) | 2009-02-23 | 2010-09-01 | 한국과학기술원 | 비디오 부호화에서의 분할 블록 부호화 방법, 비디오 복호화에서의 분할 블록 복호화 방법 및 이를 구현하는 기록매체 |
WO2010110126A1 (ja) | 2009-03-23 | 2010-09-30 | 株式会社エヌ・ティ・ティ・ドコモ | 画像予測符号化装置、画像予測符号化方法、画像予測符号化プログラム、画像予測復号装置、画像予測復号方法、及び画像予測復号プログラム |
KR101527085B1 (ko) * | 2009-06-30 | 2015-06-10 | 한국전자통신연구원 | 인트라 부호화/복호화 방법 및 장치 |
US8767816B2 (en) * | 2009-07-10 | 2014-07-01 | Samsung Electronics Co., Ltd. | Spatial prediction method and apparatus in layered video coding |
KR101479129B1 (ko) * | 2009-10-01 | 2015-01-06 | 에스케이텔레콤 주식회사 | 가변 크기의 매크로블록을 이용한 영상 부호화/복호화 방법 및 장치 |
KR101484280B1 (ko) * | 2009-12-08 | 2015-01-20 | 삼성전자주식회사 | 임의적인 파티션을 이용한 움직임 예측에 따른 비디오 부호화 방법 및 장치, 임의적인 파티션을 이용한 움직임 보상에 따른 비디오 복호화 방법 및 장치 |
CN101827270B (zh) * | 2009-12-16 | 2011-12-14 | 香港应用科技研究院有限公司 | 一种帧内预测中获取像素预测值的方法 |
CN102118613B (zh) * | 2009-12-31 | 2012-11-21 | 华为技术有限公司 | 视频解码和编码方法及装置 |
EP3595316B1 (en) | 2010-01-08 | 2022-03-30 | Nokia Technologies Oy | Apparatus and method for video coding and decoding |
JP2011199396A (ja) * | 2010-03-17 | 2011-10-06 | Ntt Docomo Inc | 動画像予測符号化装置、動画像予測符号化方法、動画像予測符号化プログラム、動画像予測復号装置、動画像予測復号方法、及び動画像予測復号プログラム |
CN102215384B (zh) * | 2010-04-02 | 2014-03-26 | 富士通株式会社 | 图像压缩方法和系统 |
CN102823248B (zh) | 2010-04-08 | 2015-06-24 | 株式会社东芝 | 图像编码方法以及图像编码装置 |
CN103813163B (zh) * | 2010-04-08 | 2017-03-01 | 株式会社东芝 | 图像解码方法以及图像解码装置 |
US8306343B2 (en) * | 2010-04-09 | 2012-11-06 | Newport Media, Inc. | Optimized prediction based image compression |
US20110249735A1 (en) * | 2010-04-09 | 2011-10-13 | Jie Zhao | Methods and Systems for Intra Prediction |
US20110249734A1 (en) * | 2010-04-09 | 2011-10-13 | Segall Christopher A | Methods and Systems for Intra Prediction |
US8644375B2 (en) | 2010-04-09 | 2014-02-04 | Sharp Laboratories Of America, Inc. | Methods and systems for intra prediction |
US8619857B2 (en) * | 2010-04-09 | 2013-12-31 | Sharp Laboratories Of America, Inc. | Methods and systems for intra prediction |
EP2375751A1 (en) * | 2010-04-12 | 2011-10-12 | Panasonic Corporation | Complexity reduction of edge-detection based spatial interpolation |
US8463059B2 (en) * | 2010-04-23 | 2013-06-11 | Futurewei Technologies, Inc. | Two-layer prediction method for multiple predictor-set intra coding |
US20110310967A1 (en) * | 2010-06-22 | 2011-12-22 | Lei Zhang | Method and System for Video and Image Coding Using Pattern Matching for Intra-Prediction |
CN103493491A (zh) | 2010-07-15 | 2014-01-01 | 株式会社东芝 | 图像编码方法以及图像解码方法 |
SG187094A1 (en) | 2010-07-20 | 2013-03-28 | Ntt Docomo Inc | Image prediction encoding device, image prediction encoding method, image prediction encoding program, image prediction decoding device, image prediction decoding method, and image prediction decoding program |
KR101663764B1 (ko) * | 2010-08-26 | 2016-10-07 | 에스케이 텔레콤주식회사 | 인트라 예측을 이용한 부호화 및 복호화 장치와 방법 |
EP2624567B1 (en) * | 2010-09-27 | 2020-03-18 | LG Electronics Inc. | Method for partitioning block and decoding device |
SG10201507816SA (en) * | 2010-09-30 | 2015-10-29 | Mitsubishi Electric Corp | Moving image encoding device, moving image decoding device, moving image coding method, and moving image decoding method |
KR20240128140A (ko) * | 2010-12-08 | 2024-08-23 | 엘지전자 주식회사 | 인트라 예측 방법과 이를 이용한 부호화 장치 및 복호화 장치 |
US10045046B2 (en) * | 2010-12-10 | 2018-08-07 | Qualcomm Incorporated | Adaptive support for interpolating values of sub-pixels for video coding |
FR2969456A1 (fr) * | 2010-12-15 | 2012-06-22 | France Telecom | Procedes et dispositifs de codage et de decodage d'au moins une image mettant en œuvre une selection des pixels a predire, programme d'ordinateur correspondant |
PL2661892T3 (pl) | 2011-01-07 | 2022-08-16 | Nokia Technologies Oy | Przewidywanie ruchu w kodowaniu wideo |
JP5781313B2 (ja) * | 2011-01-12 | 2015-09-16 | 株式会社Nttドコモ | 画像予測符号化方法、画像予測符号化装置、画像予測符号化プログラム、画像予測復号方法、画像予測復号装置及び画像予測復号プログラム |
KR101604959B1 (ko) * | 2011-01-12 | 2016-03-18 | 미쓰비시덴키 가부시키가이샤 | 화상 부호화 장치, 화상 복호 장치, 화상 부호화 방법 및 화상 복호 방법 |
JP6056122B2 (ja) * | 2011-01-24 | 2017-01-11 | ソニー株式会社 | 画像符号化装置と画像復号装置およびその方法とプログラム |
US9930366B2 (en) * | 2011-01-28 | 2018-03-27 | Qualcomm Incorporated | Pixel level adaptive intra-smoothing |
CN103780912B (zh) * | 2011-02-10 | 2017-11-24 | 威勒斯媒体国际有限公司 | 图像处理装置和图像处理方法 |
US20130336398A1 (en) * | 2011-03-10 | 2013-12-19 | Electronics And Telecommunications Research Institute | Method and device for intra-prediction |
US9135724B2 (en) | 2011-03-11 | 2015-09-15 | Sony Corporation | Image processing apparatus and method |
HUE035495T2 (en) * | 2011-06-28 | 2018-05-02 | Samsung Electronics Co Ltd | A method and apparatus for video coding and a method and apparatus for decoding video, associated with intra prediction |
MX368350B (es) | 2011-06-28 | 2019-09-30 | Samsung Electronics Co Ltd | Metodo y aparato para codificar y decodificar imagenes utilizando intra-prediccion. |
WO2013001727A1 (ja) * | 2011-06-30 | 2013-01-03 | 三菱電機株式会社 | 画像符号化装置、画像復号装置、画像符号化方法及び画像復号方法 |
GB2493755B (en) | 2011-08-17 | 2016-10-19 | Canon Kk | Method and device for encoding a sequence of images and method and device for decoding a sequence of images |
SI3217667T1 (sl) * | 2011-10-17 | 2020-12-31 | Kabushiki Kaisha Toshiba | Kodirni postopek in dekodirni postopek |
JP5698644B2 (ja) * | 2011-10-18 | 2015-04-08 | 株式会社Nttドコモ | 動画像予測符号化方法、動画像予測符号化装置、動画像予測符号化プログラム、動画像予測復号方法、動画像予測復号装置及び動画像予測復号プログラム |
SI2773118T1 (sl) * | 2011-10-24 | 2021-02-26 | Innotive Ltd | Postopek in naprava za dekodiranje slike |
KR20130050149A (ko) | 2011-11-07 | 2013-05-15 | 오수미 | 인터 모드에서의 예측 블록 생성 방법 |
BR122020011581B1 (pt) | 2011-11-08 | 2022-05-03 | Kabushiki Kaisha Toshiba | Método de decodificação realizado pelo menos parcialmente por um circuito eletrônico e método de codificação realizado pelo menos parcialmente por um circuito eletrônico |
EP2618309A1 (en) * | 2012-01-19 | 2013-07-24 | Siemens Aktiengesellschaft | Methods and devices for pixel-prediction for compression of visual data |
US9210438B2 (en) * | 2012-01-20 | 2015-12-08 | Sony Corporation | Logical intra mode naming in HEVC video coding |
RU2582057C2 (ru) * | 2012-02-29 | 2016-04-20 | Сони Корпорейшн | Устройство и способ обработки изображений |
US9510019B2 (en) | 2012-08-09 | 2016-11-29 | Google Inc. | Two-step quantization and coding method and apparatus |
US9332266B2 (en) | 2012-08-24 | 2016-05-03 | Industrial Technology Research Institute | Method for prediction in image encoding and image encoding apparatus applying the same |
JP5798539B2 (ja) * | 2012-09-24 | 2015-10-21 | 株式会社Nttドコモ | 動画像予測符号化装置、動画像予測符号化方法、動画像予測復号装置及び動画像予測復号方法 |
US9253483B2 (en) | 2012-09-25 | 2016-02-02 | Google Technology Holdings LLC | Signaling of scaling list |
EP2920969A4 (en) * | 2012-11-13 | 2016-07-13 | Intel Corp | VIDEO CODEC ARCHITECTURE FOR VIDEOS OF THE NEXT GENERATION |
CN104704827B (zh) | 2012-11-13 | 2019-04-12 | 英特尔公司 | 用于下一代视频的内容自适应变换译码 |
CN102984523B (zh) * | 2012-12-14 | 2015-12-09 | 北京大学 | 一种多方向的帧内预测编解码方法及装置 |
US9794569B2 (en) | 2013-01-30 | 2017-10-17 | Intel Corporation | Content adaptive partitioning for prediction and coding for next generation video |
CN105122807B (zh) * | 2013-03-29 | 2018-11-27 | Jvc建伍株式会社 | 图像解码装置及图像解码方法 |
US20140328406A1 (en) * | 2013-05-01 | 2014-11-06 | Raymond John Westwater | Method and Apparatus to Perform Optimal Visually-Weighed Quantization of Time-Varying Visual Sequences in Transform Space |
US9565440B2 (en) * | 2013-06-25 | 2017-02-07 | Vixs Systems Inc. | Quantization parameter adjustment based on sum of variance and estimated picture encoding cost |
GB2516425B (en) * | 2013-07-17 | 2015-12-30 | Gurulogic Microsystems Oy | Encoder and decoder, and method of operation |
CN113784124B (zh) * | 2014-01-01 | 2024-03-19 | 上海天荷电子信息有限公司 | 使用多种形状子块进行精细划分的块匹配编码和解码方法 |
CN107925771B (zh) * | 2015-05-29 | 2022-01-11 | 深圳市大疆创新科技有限公司 | 视频处理的方法、系统、存储介质和成像装置 |
EP3393126A4 (en) * | 2016-02-16 | 2019-04-17 | Samsung Electronics Co., Ltd. | INTRA PREDICTION METHOD FOR REDUCING INTRA PREDICTION ERRORS AND DEVICE THEREFOR |
RU2631992C2 (ru) * | 2016-02-25 | 2017-09-29 | Кабусики Кайся Тосиба | Способ кодирования изображения и способ декодирования изображения |
JP2019525577A (ja) | 2016-07-18 | 2019-09-05 | エレクトロニクス アンド テレコミュニケーションズ リサーチ インスチチュートElectronics And Telecommunications Research Institute | 画像符号化/復号方法、装置、及び、ビットストリームを保存した記録媒体 |
WO2018026166A1 (ko) | 2016-08-01 | 2018-02-08 | 한국전자통신연구원 | 영상 부호화/복호화 방법, 장치 및 비트스트림을 저장한 기록 매체 |
CN118474339A (zh) | 2016-10-04 | 2024-08-09 | 有限公司B1影像技术研究所 | 图像数据编码/解码方法、介质和传输数据的方法 |
KR102228474B1 (ko) | 2016-10-14 | 2021-03-16 | 후아웨이 테크놀러지 컴퍼니 리미티드 | 비디오 코딩을 위한 디바이스들 및 방법들 |
CN110710205B (zh) | 2017-05-19 | 2023-05-05 | 松下电器(美国)知识产权公司 | 编码装置、解码装置、编码方法和解码方法 |
WO2019047883A1 (en) * | 2017-09-08 | 2019-03-14 | Fg Innovation Ip Company Limited | DEVICE AND METHOD FOR ENCODING VIDEO DATA BASED ON MULTIPLE REFERENCE LINES |
US20230158653A1 (en) * | 2021-03-11 | 2023-05-25 | Illinois Tool Works Inc. | Fastener-driving tool with chamber member retaining assembly |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH02143679A (ja) * | 1988-11-24 | 1990-06-01 | Canon Inc | 画像符号化方法 |
JPH06181523A (ja) * | 1992-12-14 | 1994-06-28 | Ricoh Co Ltd | 予測符号化方式の符号化装置および復号化装置 |
JP2005198310A (ja) * | 2004-01-06 | 2005-07-21 | Samsung Electronics Co Ltd | 予測符号化装置、予測符号化方法及び前記予測符号化方法を行うプログラムが記録されたコンピュータ読み取り可能な記録媒体 |
JP2005318497A (ja) * | 2003-12-26 | 2005-11-10 | Ntt Docomo Inc | 画像符号化装置、画像符号化方法、画像符号化プログラム、画像復号装置、画像復号方法、及び画像復号プログラム。 |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5086487A (en) * | 1988-11-24 | 1992-02-04 | Canon Kabushiki Kaisha | Method and apparatus for image encoding in which reference pixels for predictive encoding can be selected based on image size |
MXPA05007449A (es) * | 2003-01-10 | 2005-09-12 | Thomson Licensing Sa | Definicion de filtros de interpolacion para ocultamiento de error en una imagen codificada. |
CN1213613C (zh) * | 2003-09-12 | 2005-08-03 | 浙江大学 | 视频编解码中运动矢量的预测方法和装置 |
JP2005159947A (ja) | 2003-11-28 | 2005-06-16 | Matsushita Electric Ind Co Ltd | 予測画像生成方法、画像符号化方法および画像復号化方法 |
JP2005311473A (ja) * | 2004-04-16 | 2005-11-04 | Sharp Corp | 撮像装置および信号処理方法ならびにそのプログラムと記録媒体 |
US8467447B2 (en) * | 2004-05-07 | 2013-06-18 | International Business Machines Corporation | Method and apparatus to determine prediction modes to achieve fast video encoding |
-
2006
- 2006-11-27 JP JP2007547928A patent/JP5044415B2/ja not_active Expired - Fee Related
- 2006-11-27 KR KR20087013004A patent/KR100977101B1/ko not_active IP Right Cessation
- 2006-11-27 EP EP06833410A patent/EP1965589A1/en not_active Withdrawn
- 2006-11-27 BR BRPI0619193-2A patent/BRPI0619193A2/pt not_active IP Right Cessation
- 2006-11-27 RU RU2008123371A patent/RU2369038C1/ru not_active IP Right Cessation
- 2006-11-27 AU AU2006320064A patent/AU2006320064B2/en not_active Ceased
- 2006-11-27 CN CN2006800512728A patent/CN101361370B/zh not_active Expired - Fee Related
- 2006-11-27 WO PCT/JP2006/323606 patent/WO2007063808A1/ja active Application Filing
- 2006-11-27 CA CA 2631336 patent/CA2631336A1/en not_active Abandoned
- 2006-11-29 US US11/564,641 patent/US8149910B2/en not_active Expired - Fee Related
-
2009
- 2009-05-21 RU RU2009119262A patent/RU2408161C1/ru not_active IP Right Cessation
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH02143679A (ja) * | 1988-11-24 | 1990-06-01 | Canon Inc | 画像符号化方法 |
JPH06181523A (ja) * | 1992-12-14 | 1994-06-28 | Ricoh Co Ltd | 予測符号化方式の符号化装置および復号化装置 |
JP2005318497A (ja) * | 2003-12-26 | 2005-11-10 | Ntt Docomo Inc | 画像符号化装置、画像符号化方法、画像符号化プログラム、画像復号装置、画像復号方法、及び画像復号プログラム。 |
JP2005198310A (ja) * | 2004-01-06 | 2005-07-21 | Samsung Electronics Co Ltd | 予測符号化装置、予測符号化方法及び前記予測符号化方法を行うプログラムが記録されたコンピュータ読み取り可能な記録媒体 |
Cited By (43)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9516340B2 (en) | 2007-06-12 | 2016-12-06 | Thomson Licensing | Methods and apparatus supporting multi-pass video syntax structure for slice data |
JP2010529811A (ja) * | 2007-06-12 | 2010-08-26 | トムソン ライセンシング | スライスデータ用のマルチパスビデオのシンタックス構造をサポートする方法及び装置 |
US8204136B2 (en) | 2008-03-07 | 2012-06-19 | Canon Kabushiki Kaisha | Image encoding apparatus |
CN101527848B (zh) * | 2008-03-07 | 2012-09-05 | 佳能株式会社 | 图像编码设备 |
JP2009246975A (ja) * | 2008-03-28 | 2009-10-22 | Samsung Electronics Co Ltd | 画像のイントラ予測モード情報の符号化及び/又は復号化方法及び装置 |
US8509302B2 (en) | 2008-10-22 | 2013-08-13 | Nippon Telegraph And Telephone Corporation | Scalable video encoding method, scalable video encoding apparatus, scalable video encoding program, and computer readable recording medium storing the program |
RU2488235C2 (ru) * | 2008-10-22 | 2013-07-20 | Ниппон Телеграф Энд Телефон Корпорейшн | Способ масштабируемого кодирования видео, устройство масштабируемого кодирования видео, программа масштабируемого кодирования видео и машиночитаемый носитель записи, сохраняющий программу |
US8483496B2 (en) | 2009-01-22 | 2013-07-09 | Ntt Docomo, Inc. | Image encoding/decoding system |
WO2010084692A1 (ja) * | 2009-01-22 | 2010-07-29 | 株式会社エヌ・ティ・ティ・ドコモ | 画像予測符号化装置、方法及びプログラム、画像予測復号装置、方法及びプログラム、並びに、符号化・復号システム及び方法 |
JP2010171729A (ja) * | 2009-01-22 | 2010-08-05 | Ntt Docomo Inc | 画像予測符号化装置、方法及びプログラム、画像予測復号装置、方法及びプログラム、並びに、符号化・復号システム及び方法 |
US9078008B2 (en) | 2009-04-20 | 2015-07-07 | Dolby Laboratories Licensing Corporation | Adaptive inter-layer interpolation filters for multi-layered video delivery |
JP2012521735A (ja) * | 2009-04-20 | 2012-09-13 | ドルビー ラボラトリーズ ライセンシング コーポレイション | 多層映像配信のための適応補間フィルタ |
JP2010283495A (ja) * | 2009-06-03 | 2010-12-16 | Fujitsu Ltd | 動画像符号化装置及び動画像符号化方法 |
US9106917B2 (en) | 2009-06-03 | 2015-08-11 | Fujitsu Limited | Video encoding apparatus and video encoding method |
US10798412B2 (en) | 2009-07-04 | 2020-10-06 | Dolby Laboratories Licensing Corporation | Encoding and decoding architectures for format compatible 3D video delivery |
US10038916B2 (en) | 2009-07-04 | 2018-07-31 | Dolby Laboratories Licensing Corporation | Encoding and decoding architectures for format compatible 3D video delivery |
US9774882B2 (en) | 2009-07-04 | 2017-09-26 | Dolby Laboratories Licensing Corporation | Encoding and decoding architectures for format compatible 3D video delivery |
JP2013507087A (ja) * | 2009-10-05 | 2013-02-28 | トムソン ライセンシング | 映像符号化及び復号化における色度成分の予測画素の適応フィルタリング |
US10291938B2 (en) | 2009-10-05 | 2019-05-14 | Interdigital Madison Patent Holdings | Methods and apparatus for adaptive filtering of prediction pixels for chroma components in video encoding and decoding |
US9369712B2 (en) | 2010-01-14 | 2016-06-14 | Dolby Laboratories Licensing Corporation | Buffered adaptive filters |
JP2011166218A (ja) * | 2010-02-04 | 2011-08-25 | Victor Co Of Japan Ltd | 画像信号符号化装置、画像信号符号化方法、画像信号復号装置、及び、画像信号復号方法 |
WO2012011340A1 (ja) * | 2010-07-20 | 2012-01-26 | ソニー株式会社 | 画像処理装置及び画像処理方法 |
JP2012034091A (ja) * | 2010-07-29 | 2012-02-16 | Ntt Docomo Inc | 画像予測符号化装置、方法およびプログラム、並びに、画像予測復号装置、方法およびプログラム |
US9584829B2 (en) | 2010-07-31 | 2017-02-28 | M&K Holdings Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9609360B2 (en) | 2010-07-31 | 2017-03-28 | M&K Holdings Inc. | Image decoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9554154B2 (en) | 2010-07-31 | 2017-01-24 | M & K Holding Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9578349B2 (en) | 2010-07-31 | 2017-02-21 | M&K Holdings Inc. | Image decoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9584827B2 (en) | 2010-07-31 | 2017-02-28 | M&K Holdings Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9584828B2 (en) | 2010-07-31 | 2017-02-28 | M&K Holdings Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9584830B2 (en) | 2010-07-31 | 2017-02-28 | M&K Holdings Inc. | Image decoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
USRE48564E1 (en) | 2010-07-31 | 2021-05-18 | M&K Holdings Inc. | Image decoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9584826B2 (en) | 2010-07-31 | 2017-02-28 | M&K Holdings Inc. | Image encoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
US9609359B2 (en) | 2010-07-31 | 2017-03-28 | M&K Holdings Inc. | Image decoding apparatus adaptively determining a scan pattern according to an intra prediction mode |
WO2012035640A1 (ja) * | 2010-09-16 | 2012-03-22 | 株式会社 東芝 | 動画像符号化方法及び動画像復号化方法 |
JP2013223149A (ja) * | 2012-04-17 | 2013-10-28 | Nippon Hoso Kyokai <Nhk> | 画像符号化装置、画像復号装置、画像符号化プログラム及び画像復号プログラム |
JP5841670B2 (ja) * | 2012-09-28 | 2016-01-13 | 日本電信電話株式会社 | イントラ予測符号化方法、イントラ予測復号方法、イントラ予測符号化装置、イントラ予測復号装置、それらのプログラム並びにプログラムを記録した記録媒体 |
US9813709B2 (en) | 2012-09-28 | 2017-11-07 | Nippon Telegraph And Telephone Corporation | Intra-prediction encoding method, intra-prediction decoding method, intra-prediction encoding apparatus, intra-prediction decoding apparatus, program therefor and recording medium having program recorded thereon |
WO2014050971A1 (ja) * | 2012-09-28 | 2014-04-03 | 日本電信電話株式会社 | イントラ予測符号化方法、イントラ予測復号方法、イントラ予測符号化装置、イントラ予測復号装置、それらのプログラム並びにプログラムを記録した記録媒体 |
JP2016027756A (ja) * | 2012-09-28 | 2016-02-18 | 日本電信電話株式会社 | イントラ予測符号化方法、イントラ予測復号方法、イントラ予測符号化装置、イントラ予測復号装置、それらのプログラム並びにプログラムを記録した記録媒体 |
JP2016519898A (ja) * | 2013-04-08 | 2016-07-07 | ジーイー ビデオ コンプレッション エルエルシー | コンポーネント間予測 |
US10237567B2 (en) | 2013-04-08 | 2019-03-19 | Ge Video Compression Llc | Inter-component prediction |
JP2020509714A (ja) * | 2017-04-04 | 2020-03-26 | ホアウェイ・テクノロジーズ・カンパニー・リミテッド | ピクチャブロックのイントラ予測に使用される参照サンプルの処理 |
JPWO2019189904A1 (ja) * | 2018-03-30 | 2021-02-12 | 日本放送協会 | イントラ予測装置、画像符号化装置、画像復号装置、及びプログラム |
Also Published As
Publication number | Publication date |
---|---|
KR20080067363A (ko) | 2008-07-18 |
RU2369038C1 (ru) | 2009-09-27 |
KR100977101B1 (ko) | 2010-08-23 |
US8149910B2 (en) | 2012-04-03 |
BRPI0619193A2 (pt) | 2011-09-20 |
JPWO2007063808A1 (ja) | 2009-05-07 |
AU2006320064A1 (en) | 2007-06-07 |
US20070121731A1 (en) | 2007-05-31 |
EP1965589A1 (en) | 2008-09-03 |
RU2408161C1 (ru) | 2010-12-27 |
JP5044415B2 (ja) | 2012-10-10 |
CN101361370A (zh) | 2009-02-04 |
CA2631336A1 (en) | 2007-06-07 |
CN101361370B (zh) | 2010-12-01 |
AU2006320064B2 (en) | 2010-09-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2007063808A1 (ja) | 画像符号化/画像復号化方法及び画像符号化/画像復号化装置 | |
AU2006346583B2 (en) | Image encoding and decoding method and apparatus | |
RU2414093C2 (ru) | Способ и устройство, и программа кодирования/декодирования видео | |
KR101538704B1 (ko) | 보간 필터를 적응적으로 사용하여 영상을 부호화 및 복호화하는 방법 및 장치 | |
KR101838124B1 (ko) | 평면 표현을 이용한 향상된 인트라-예측 부호화 | |
US20100118945A1 (en) | Method and apparatus for video encoding and decoding | |
JP2010135864A (ja) | 画像符号化方法及び装置並びに画像復号化方法及び装置 | |
CN104054338A (zh) | 位深和颜色可伸缩视频编码 | |
JPWO2008084817A1 (ja) | 画像符号化と復号化の方法及び装置 | |
KR20110000438A (ko) | 다차원 정수 변환을 이용한 영상 부호화/복호화 장치 및 방법 | |
US20060198439A1 (en) | Method and system for mode decision in a video encoder | |
RU2427976C2 (ru) | Способ и устройство для кодирования и декодирования изображения | |
KR101739580B1 (ko) | 적응적 스캔 장치 및 그 스캔 방법 | |
KR20100088034A (ko) | 인트라 예측 모드 결정 방법 및 장치 | |
Rao et al. | H. 264/mpeg-4 advanced video coding | |
MX2008006909A (en) | Image encoding/image decoding method and image encoding/image decoding apparatus | |
MX2009000832A (es) | Metodo y aparato de codificacion y decodificacion de imagenes. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 2007547928 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2631336 Country of ref document: CA Ref document number: 4594/DELNP/2008 Country of ref document: IN Ref document number: 2006833410 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: MX/a/2008/006909 Country of ref document: MX Ref document number: 1020087013004 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 12008501303 Country of ref document: PH |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2008123371 Country of ref document: RU |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2006320064 Country of ref document: AU |
|
ENP | Entry into the national phase |
Ref document number: 2006320064 Country of ref document: AU Date of ref document: 20061127 Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 200680051272.8 Country of ref document: CN |
|
ENP | Entry into the national phase |
Ref document number: PI0619193 Country of ref document: BR Kind code of ref document: A2 Effective date: 20080529 |