WO2006049335A1 - 動画像符号化装置および動画像符号化処理プログラム - Google Patents
動画像符号化装置および動画像符号化処理プログラム Download PDFInfo
- Publication number
- WO2006049335A1 WO2006049335A1 PCT/JP2005/020689 JP2005020689W WO2006049335A1 WO 2006049335 A1 WO2006049335 A1 WO 2006049335A1 JP 2005020689 W JP2005020689 W JP 2005020689W WO 2006049335 A1 WO2006049335 A1 WO 2006049335A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- coding
- encoding
- quantization error
- frame
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
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
- H04N19/17—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
- H04N19/10—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
- H04N19/109—Selection of coding mode or of prediction mode among a plurality of temporal 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
- H04N19/103—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
- H04N19/134—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
- H04N19/137—Motion inside a coding unit, e.g. average field, frame or block difference
-
- 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
- H04N19/189—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding
- H04N19/192—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the adaptation method, adaptation tool or adaptation type used for the adaptive coding the adaptation method, adaptation tool or adaptation type being iterative or recursive
Definitions
- the present invention relates to a moving image encoding method that compresses a moving image in consideration of human visual characteristics.
- the present invention relates to an apparatus and a moving image encoding processing program.
- FIG. 9 is a block diagram showing a schematic configuration of the moving picture coding apparatus.
- the subtractor 10 subtracts the pixel value of the prediction block from the pixel value of the current block to generate a prediction error signal representing the luminance difference.
- the transform / quantizer 1 1 performs integer DCT (discrete cosine transform) on the prediction error signal output from the subtractor 10 and quantizes the transform coefficient obtained by this with a predetermined quantization width to obtain coefficient data. Is generated.
- the entropy coding unit 1 2 uses the variable-length coding VL C-based exponential Golomb code and the CA BAG (Context-based Adaptive Binary Arithmetic Coding) applied to the coefficient data generated by the transform quantization unit 1 1. Entropy coding using) and.
- the frame memory 16 forms a local decoding unit.
- the local decoding unit performs inverse quantization and inverse integer DCT on the coefficient data generated by the transform quantization unit 1 1 and then adds the pixel values of the previous prediction block to generate a decoded image.
- Loop noise is applied to reduce block noise, and then temporarily stored in the frame memory 16.
- the intra-frame prediction unit 17 calculates an intra-frame prediction block value using the decoded image read from the frame memory 16.
- the motion detector 18 detects the motion vector of the current block.
- the motion compensation unit 19 calculates the inter-frame prediction block value by performing motion compensation on the reference frame (the decoded image read from the frame memory 16) according to the motion vector detected by the motion detection unit 18. .
- the selector 20 selects either the intra-frame prediction block value calculated by the intra-frame prediction unit 17 or the inter-frame prediction block value calculated by the motion compensation unit 19 according to the instruction of the determination unit 21.
- the determination unit 21 estimates the generated code amount at the time of intraframe predictive encoding and the generated code amount of the interframe predictive encoding, and instructs the selector 20 to select an encoding method with a small generated code amount.
- FIG. 10 is a flowchart showing the operation of the coding determination process executed for each block image input.
- the process proceeds to step SF1 to initialize each part of the apparatus.
- step SF 2 inter prediction processing for estimating the generated code amount at the time of inter-frame predictive coding is executed, and in subsequent step SF 3, the generated code amount at the time of intra-frame predictive coding is estimated and Intra prediction processing that performs encoding (integer D CT, quantization, inverse quantization, and inverse integer D CT) is executed.
- step SF 4 the minimum difference absolute value sum S AD inter (corresponding to the generated code amount at the time of inter prediction) obtained in step SF 2 is the minimum difference absolute value obtained in step SF 3.
- Value sum S AD intra (corresponds to the generated code amount for intra prediction) It is determined whether it is larger.
- the moving image encoding apparatus compresses the moving image by the intra-frame encoding executed in step SF3.
- step SF 4 if the generated code amount at the time of inter prediction is smaller than the generated code amount at the time of intra prediction, the judgment result of the above step SF 4 becomes “N0”, and the process proceeds to step SF 5 where interframe coding (integer D Executes inter D & Q processing that performs CT, quantization, inverse quantization, and inverse integer D CT). Therefore, in this combination, the moving image encoding apparatus compresses a moving image by inter-frame encoding.
- interframe coding integer D Executes inter D & Q processing that performs CT, quantization, inverse quantization, and inverse integer D CT.
- step SG 1 a macro block of 16 ⁇ 16 pixels is divided into 16 ⁇ 4 ⁇ Judge whether all four-pixel blocks have been processed.
- the judgment result is “YE S” and this process is completed. Otherwise, the judgment result is “N ⁇ ”, and the process proceeds to Step SG2.
- the relationship between the 16 X 16 pixel macroblock and the 4 X 4 pixel block is such that a 4 X4 pixel block is specified by the macroblock block number n, as shown in Figure 13 It has become.
- step S G 2 the pixel value 0 rg of the 4 ⁇ 4 pixel block to be processed (hereinafter referred to as the current block) is calculated.
- step S G 3 an MV search process for searching for a motion vector is executed.
- the correlation with the current block is calculated while shifting the pixel around the reference block within the search range of the reference frame. The position of the pixel with the highest correlation is extracted as a motion vector.
- the correlation between the reference block and the current block is evaluated using the sum of absolute differences SAD of both blocks. Therefore, when the position of the pixel with the highest correlation is extracted as a motion vector, the difference absolute value sum SAD is minimized.
- the minimum sum of absolute differences S AD inter is used in step S F 4 (see FIG. 10) for determining whether or not to perform an int D & Q process.
- step SJ 1 shown in FIG. 14 to divide the macro block of 16 ⁇ 16 pixels into 16 parts 4 X Determine if all 4 pixel blocks have been processed.
- the judgment result is “YE S” and this process is completed. Otherwise, the judgment result is “N0” and the process proceeds to step SJ2.
- step SJ2 for example, in the case of intra 4 X4 mode, mode 0 to m
- the prediction block value of each mode is calculated by 9 kinds of prediction methods of ode 8 in total.
- steps SJ 3 to SJ 5 the sum of absolute differences S AD between the current block and the predicted block value of each mode is calculated corresponding to all mo de 0 to mo de 8, and the smallest one of them is calculated.
- the sum of absolute differences S AD intra is obtained.
- the minimum sum of absolute differences S AD intra is used in step S F 4 (see FIG. 10) for determining whether or not to perform an int D D & Q process.
- step SJ 3 When the calculation of the sum of absolute differences S AD is completed for all modes 0 to 8 and the smallest sum of absolute differences S AD intra is determined, the judgment result in step SJ 3 is “YES”, and step SJ Proceed to 6, and encode the current block in frame with mode that generates the minimum sum of absolute differences S AD intra. Thereafter, the above-described steps S J 1 to S J 6 are repeated until all the blocks are processed.
- step SK 1 shown in FIG. 15 and a 16 ⁇ 16 pixel macroblock is divided into 16 4 X Determine if all 4 pixel blocks have been processed. If all the blocks have not been processed, the judgment result is “NO” and the process proceeds to step SK2.
- steps SK 2 to SK 6 conversion processing to a prediction error signal generated by subtracting the inter-frame prediction block value ref (i, j) from the pixel value O rg (i, j) of the current block (integer D CT) Quantization processing Q, inverse quantization Q-1 and inverse transformation processing (integer D CT-1).
- step SK 7 to increment the inter-frame prediction block value ref (i, j).
- steps SK 1 to SK 7 are repeated until all blocks have been processed.
- a moving image coding apparatus that selectively compresses a moving image by using intraframe coding that uses correlation in the spatial domain or interframe coding that uses correlation in the time domain, block image input
- the smallest difference absolute value sum S AD inter corresponding to the generated code amount at the time of inter prediction and the smallest difference absolute value sum SAD intra corresponding to the generated code amount at the time of intra prediction are calculated and compared in magnitude. Therefore, compression encoding is performed by selecting an encoding method with a smaller sum of absolute differences SAD.
- the moving picture coding apparatus that compresses a moving image by selectively using the above-described intraframe coding or interframe coding has the following problems.
- a macro block to be coded is encoded.
- the macroblock image to be coded is divided in units of block images.
- F A first encoding determination means for determining whether or not there is a block image having visually noticeable noise when performing inter-frame encoding, and selecting intra-frame encoding when the corresponding block image exists; The second is to determine whether there is a block image that has visually noticeable noise in consideration of the size of the motion vector obtained in the prediction, and to select intra-frame coding when the corresponding block image exists.
- coding determination means for determining whether or not there is a block image having visually noticeable noise when performing inter-frame encoding, and selecting intra-frame encoding when the corresponding block image exists.
- the evaluation obtained by the motion vector search at the time of inter-frame prediction Error estimation means that estimates the magnitude of the quantization error of the entire macroblock image to be encoded using the value, and inter-frame encoding in units of block images obtained by dividing the macroblock image.
- a first encoding determination means for determining whether or not there is a block image having visually noticeable noise and selecting intra-frame encoding when the corresponding block image exists; When the quantization error estimated by the error estimation means is larger than a predetermined value, the presence or absence of a block image having visually noticeable noise is determined from the estimated quantization error, and the corresponding block image exists. And a second coding determining means for selecting intra-frame coding when performing the above.
- the evaluation obtained by the motion vector search at the time of inter-frame prediction Value is used to estimate the amount of quantization error in the entire macroblock image to be encoded.
- An error estimation unit and when the quantization error estimated by the error estimation unit is smaller than a predetermined value, the flatness of each block image is calculated when inter-frame encoding is performed in units of block images obtained by dividing a macroblock image. Quantization error is detected, and the presence or absence of a block image having visually noticeable noise is judged from the relationship between the detected flatness and the quantization error, and intra-frame coding is performed when the corresponding block image exists.
- a first encoding determination unit to be selected and if the quantization error estimated by the error estimation unit is greater than a predetermined value, the presence / absence of a block image having visually noticeable noise is determined from the estimated quantization error.
- the first quantization determination unit that selects intra-frame coding if the error is conspicuous, and the estimated quantization error To determine whether there is visually noticeable noise in the macroblock image, and if there is no visually noticeable noise, select inter-frame coding. On the other hand, if there is visually noticeable noise, And second encoding determination means for selecting intra-frame encoding.
- the second encoding determination means considers the size of the dynamic vector obtained by inter-frame prediction. The presence or absence of a block image having visually noticeable noise is judged.
- the motion vector having the highest correlation with the evaluation obtained by the motion vector search at the time of inter-frame prediction is detected. It is characterized by the sum of absolute differences.
- inter-frame coding is performed in units of block images obtained by dividing a macroblock image. And detecting the flatness and the quantization error for each block image and detecting the magnitude of the quantization error of the entire macroblock image, and the macroblock image detected by the detection unit. When the total quantization error is smaller than the predetermined value, it is determined whether there is a block image having visually noticeable noise from the relationship between the detected flatness and the quantization error, and the corresponding block image exists.
- a first encoding determination unit that selects intra-frame encoding; and a quantization error of the entire macroblock image detected by the detection unit is greater than a predetermined value. If there is a block image that has visually noticeable noise, considering the size of the motion vector obtained by inter-frame prediction, And a second coding determination means for selecting inner coding.
- a block image obtained by dividing a macroblock image to be encoded When performing inter-frame coding on an image basis, it is determined whether there is a block image having visually noticeable noise, and when the corresponding block image exists, the coding determination process for selecting intra-frame coding is executed by the computer. It is characterized by that.
- a block image obtained by dividing a macroblock image to be encoded First encoding determination process that determines whether there is a block image with visually noticeable noise when inter-frame encoding is performed in units, and selects intra-frame encoding when the corresponding block image exists Considering the size of the motion vector obtained by inter-frame prediction, it is determined whether there is a block image with visually noticeable noise, and intra-frame coding is performed when the corresponding block image exists.
- the second encoding determination process to be selected is executed by a computer.
- a moving image coding processing program for compressing a moving image by selecting either intraframe coding or interframe coding it is obtained by motion vector search at the time of interframe prediction.
- Error estimation processing that estimates the magnitude of the quantization error of the entire macroblock image to be encoded, and inter-frame encoding in units of block images obtained by dividing the macroblock image.
- First coding that determines whether there is a block image with visually noticeable noise from the relationship with the error, and selects intra-frame coding when the corresponding block image exists
- a constant process when the quantization error estimated by the error estimation process is greater than a predetermined value, estimated
- the computer determines whether there is a block image with visually noticeable noise from the determined quantization error, and executes a second encoding determination process that selects intra-frame encoding when the corresponding block image exists. It is characterized by letting
- a moving image coding processing program that compresses a moving image by selecting either intraframe coding or interframe coding, it is obtained by motion vector search at the time of interframe prediction.
- An error estimation process for estimating the magnitude of the quantization error of the entire macroblock image to be encoded, and the quantization error estimated by the error estimation process is smaller than a predetermined value.
- the second encoding determination process to be selected is executed by a computer.
- the quantization error estimated by the estimation process is larger than the predetermined value, it is determined from the estimated quantization error whether a visually noticeable noise exists in the macroblock image, and there is no visually noticeable noise. For example, inter-frame coding is selected. On the other hand, if there is visually noticeable noise, a second coding determination process for selecting intra-frame coding is executed by a computer. In the invention according to claim 14 that depends on any one of claims 1 1 to 1 3, the second encoding determination process considers the size of a motion vector obtained by inter-frame prediction. Thus, the presence or absence of a block image having visually noticeable noise is judged.
- the evaluation value obtained by the motion vector search at the time of inter-frame prediction is the motion having the highest correlation. It is characterized by the sum of absolute differences when a vector is detected.
- a frame is divided in units of block images obtained by dividing a macro block image.
- a detection process for detecting flatness and quantization error for each block image and detecting the magnitude of the quantization error for the entire macroblock image, and a macroblock detected by the detection process
- the quantization error of the entire image is smaller than the predetermined value, it is determined whether there is a block image with visually noticeable noise from the relationship between the detected flatness and the quantization error.
- the second encoding determination process for selecting the intra-frame encoding when the corresponding block image exists is executed by a computer. According to the inventions described in claims 1 and 9, when a block image having visually noticeable noise exists when inter-frame encoding is performed in units of divided block images of a macro block image to be encoded. Since intra-frame coding is selected, the intra-prediction processing (see Fig. 14) with a large amount of computation, which has been performed for coding determination even though the usage frequency is low, can be omitted. The coding mode can be determined at high speed in consideration of human visual characteristics without performing unnecessary calculations.
- intra-frame coding is selected immediately, so that it is possible to quickly cut off the quantization error propagation when falling into the local minimum, and to avoid degradation of image quality.
- intra-frame coding is selected, and intra-frame coding is selected when there is a block image with visually noticeable noise considering the size of the motion vector obtained by inter-frame prediction.
- the intra prediction process see Fig. 14
- the coding mode can be determined at high speed in consideration of visual characteristics.
- intra-frame coding is performed immediately. Since it is selected, the quantization error propagation in the case of falling into the local minimum can be cut off quickly to avoid degradation of image quality.
- the magnitude of the quantization error of the entire macroblock image to be encoded is calculated using the evaluation value obtained by the dynamic vector search at the time of inter-frame prediction. Estimate and detect the flatness and quantization error for each block image when inter-frame coding is performed in units of block images obtained by dividing the macro block image. If the estimated quantization error is smaller than a predetermined value, the presence or absence of a block image having visually noticeable noise is determined from the relationship between the detected flatness and the quantization error, and the corresponding block image is determined. Select intraframe coding when present.
- the estimated quantization error is larger than a predetermined value, the presence of a block image having visually noticeable noise is judged from the estimated quantization error, and intra-frame coding is performed when the corresponding block image exists. select.
- the coding mode can be determined at high speed in consideration of human visual characteristics.
- the intra-frame coding is selected immediately, so that the quantization error propagation in the case of falling into the local minimum can be quickly cut off, and the degradation of image quality can be avoided.
- the magnitude of the quantization error of the entire macro-picture image to be encoded is estimated using the evaluation value obtained by the dynamic vector search at the time of inter-frame prediction. If the estimated quantization error is smaller than the predetermined value, the block image When inter-frame coding is performed, the flatness and quantization error for each block image are detected, and the presence or absence of a block image having visually noticeable noise from the relationship between the detected flatness and the quantization error is detected. Judgment is made and intra-frame coding is selected when the corresponding block image exists.
- the estimated quantization error is larger than a predetermined value, the presence of a block image having visually noticeable noise is judged from the estimated quantization error, and intra-frame coding is performed when the corresponding block image exists. Select.
- the coding mode can be determined at high speed in consideration of human visual characteristics.
- the intra-frame coding is selected immediately, so that the quantization error propagation in the case of falling into the local minimum can be quickly cut off, and the degradation of image quality can be avoided.
- the magnitude of the quantization error of the entire macroblock image to be encoded is calculated using the evaluation value obtained by the dynamic vector search at the time of inter-frame prediction.
- the flatness of the entire macroblock image is calculated to determine whether the flatness is visually inconspicuous, and the noise is visually If it is not noticeable, select inter-frame coding. If it is visually noticeable, select intra-frame coding.
- the estimated quantization error is larger than a predetermined value, it is determined from the estimated quantization error whether visually noticeable noise exists in the macroblock image. If there is no visually noticeable noise, Select interframe coding and select intraframe coding if there is visually noticeable noise. Select.
- the coding mode can be determined at high speed in consideration of human visual characteristics.
- the intra-frame coding is selected immediately, so that the quantization error propagation in the case of falling into the local minimum can be quickly cut off, and the degradation of image quality can be avoided.
- the quantization error of the entire detected macroblock image is larger than a predetermined value, the presence or absence of a block image having visually noticeable noise in consideration of the size of the motion vector obtained by inter-frame prediction.
- the intra-frame coding is selected.
- the intra-frame coding is selected immediately, so that the quantization error propagation in the case of falling into the local minimum can be quickly cut off, and the degradation of image quality can be avoided.
- Human visual characteristics are broadly divided into spatial visual characteristics and temporal visual characteristics.
- spatial visual characteristics it is known that the gradation sensitivity is higher than the resolution in the flat part of the image and the resolution sensitivity is higher than the gradation in the non-flat part. Therefore, when visualizing moving images as a sequence of still images, it is necessary to consider spatial visual characteristics.
- the quantization error that occurs during encoding is added to the predicted image after being subjected to inverse DCT, so that it is finally added to each pixel of the decoded image obtained by adding the decoded differential signal to the predicted image. Will remain. If the difference value for each pixel generated in the encoding / decoding process is added up in units of blocks, the amount of noise generated in that block is obtained.
- the gradation sensitivity is higher than the resolution in the flat part of the image, when comparing blocks with the same amount of noise, the flatter the image, the more visually noticeable noise is.
- noise is conspicuous in flat areas such as “human skin”, “sky” or “ground”, but noise in non-flat areas such as “audience seats” and “forest” is inconspicuous. . Therefore, if the amount of noise and flatness are detected in units of blocks, it can be estimated whether or not the noise generated in the block is visually noticeable.
- the present invention determines whether or not visually noticeable noise exists in the current macroblock to be encoded. Intraframe coding is performed only when there is visually noticeable noise. This makes it possible to determine the coding mode at a high speed in consideration of human visual characteristics without performing useless computation. Also, as will be described later, it is possible to quickly cut off the quantization error propagation when falling into the local minimum, and to avoid degradation of image quality.
- FIG. 1 is a block diagram showing the configuration of the first embodiment according to the present invention.
- Fig. 2 is a diagram illustrating the correspondence between the m i N S A D value (x axis) of a general image such as a broadcast image and the total quantization error ⁇ ⁇ o i s e (y axis).
- Figure 3 is a graph plotting the flatness f 1 at / min S A D value of a macroblock with visually noticeable noise and the flatness f l at / min S A D value of a macroblock without visually noticeable noise.
- FIG. 4 is a flowchart showing the operation of the encoding determination process according to the first embodiment.
- FIG. 5 is a flowchart showing the operation of the int D D & Q process according to the first embodiment.
- FIG. 6 is a flowchart showing the operation of the encoding determination process according to the second embodiment.
- FIG. 7 is a flowchart showing the operation of the encoding determination process according to the third embodiment.
- FIG. 8 is a flowchart showing the operation of the encoding determination process according to the fourth embodiment.
- FIG. 9 is a block diagram showing the configuration of a conventional moving picture coding apparatus.
- FIG. 10 is a flowchart showing the operation of the conventional encoding determination process.
- FIG. 11 is a flowchart showing the operation of the conventional Int er prediction process.
- FIG. 12 is a flowchart showing the operation of the conventional MV search process.
- FIG. 13 is a diagram showing the relationship between a 16 ⁇ 16 pixel macroblock and a 4 ⁇ 4 pixel block.
- FIG. 14 is a flowchart showing the operation of the conventional Intra prediction process.
- FIG. 15 is a flow chart showing the operation of the conventional Int D D & Q process. BEST MODE FOR CARRYING OUT THE INVENTION
- FIG. 1 is a block diagram showing a schematic configuration of a moving picture encoding apparatus according to the first embodiment.
- the video encoding apparatus shown in FIG. 1 is different from the conventional example of FIG. 9 in that a flatness calculation unit 30 and a noise calculation unit 40 are newly provided, while a flatness calculation unit 30 and a noise calculation unit are provided.
- 4 Judgment unit 2 1 selects either intraframe coding or interframe coding according to each output of 0 There is.
- the flatness calculation unit 30 calculates a flatness by performing a known Hadamard transform on a block image (current macroblock) input as an encoding target. Specifically, the absolute sum of the harmonic component coefficients of the Hadamard coefficient obtained by Hadamard transform for each block of 4 x 4 pixels obtained by dividing the current macro block of 16 x 16 pixels into 16 blocks.
- the flatness of each macroblock is calculated as flatness f 1 at (0) to flat (15), and the flatness F of the current macroblock is calculated from the sum of flatness (0) to flat: 15 (5) for each sub-block. Get 1 at.
- the 4 x 4 pixel Hadamard transform has a square wave resolution that is reduced to 1/4 compared to the 16 x 16 pixel Hadamard transform, but sufficient resolution is obtained for the flatness evaluation intended by the present invention. Be done.
- the noise calculation unit 40 extracts the quantization error superimposed on the pixel value and generates noise Noise (0) to (15) for each block.
- the quantization error for each pixel is the difference value An (i, j) between the pixel value 0 rg (i, j) of the current block and the pixel value ref (i, j) of the reference image (motion-compensated local decoded image).
- j) and the difference Z n (i, j) between the difference value A 'n (i, j) after local decoding.
- the difference Z n (i, j) Noise per block whose absolute value sum is 4 x 4 pixels. Noise (0) to (15).
- the decision unit 2 1 uses the minimum absolute difference sum min obtained by the motion vector search during inter prediction as a substitute for the total quantization error ⁇ Noise (the sum of noise Noise (0) to (15)).
- the error situation is judged using S AD (hereinafter abbreviated as min S AD value), and the first or second coding judgment is made according to the result of this error situation judgment.
- the error situation determination is to determine whether or not the min SAD value of the current macroblock to be encoded is larger than a threshold TH 1 (described later). In this error situation determination, if the min S AD value is smaller than the threshold TH 1, the first encoding determination (described later) is performed. On the other hand, if the min SAD value is larger than the threshold TH 1, the second encoding is performed. Judgment (described later) is performed.
- FIG. 1 shows the relationship between the min S AD value (x axis) and the total quantization error ⁇ Noise (y axis) obtained in macroblock units of 16 x 16 pixels in a general image such as a broadcast image.
- FIG. 1 shows the relationship between the min S AD value (x axis) and the total quantization error ⁇ Noise (y axis) obtained in macroblock units of 16 x 16 pixels in a general image such as a broadcast image.
- the m i n S AD value has a linear function K 1 as a boundary, and always takes a value exceeding the total quantization error ⁇ ⁇ 0 i s e. For this reason, it can be seen that even if the total quantization error ⁇ N 0 i s e is substituted with a mi n S A D value, there is no risk of overlooking the actual quantization error, and it can sufficiently withstand rough error determination. In addition, an increase in the amount of computation can be avoided by substituting the total quantization error ⁇ N o i se with a mi n SAD value.
- the decision unit 2 1 performs the first coding decision considering “noise” and “flatness f 1 at” extracted in block units of 4 ⁇ 4 pixels obtained by dividing a macroblock into 16 blocks. I do. The details will be described later.
- the first coding decision the presence or absence of visually noticeable noise is judged from “noise noise” and “flatness f 1 at”, and visually noticeable noise is generated. If it exists, intra-frame coding is selected.
- the determination unit 21 performs the second encoding determination in consideration of “min S A D value”, “flatness f 1 at”, and “motion vector” in units of macro blocks of 16 X I 6 pixels.
- FIG. Figure 3 shows the flatness and flatness of a macroblock (black square in the graph) with visually noticeable noise, and the flatness of the macroblock (white circle in the graph) without visually noticeable noise. It is a graph plotting the degrees f 1 at and min SAD values, respectively.
- the straight line A corresponds to the threshold value T H 1 used for the error situation determination described above.
- the second coding decision it is judged whether the flatness f 1 at and min SAD values of the current macroblock are in the area above line B (the area where visually noticeable noise appears). In the upper area of In some cases, intra-frame coding is selected.
- the motion vector is large due to the movement of the object or camera pan, the current macroblock will be blurred, and in this case, the min SAD value will be smaller despite the quantization error being visually inconspicuous noise. There is a tendency to grow.
- the decision content described above must be adaptively changed according to the motion vector (u, V) of the current macroblock. It is determined whether or not
- FIG. 4 is a flowchart showing the operation of the coding determination process executed every time a block image is input.
- this process is executed in response to block image input, first, each part of the device is initialized in step SA1, and then the process proceeds to step SA2, where inter prediction processing for estimating the amount of generated code at the time of interframe prediction encoding is executed. To do.
- This inter prediction process is the same as the conventional example shown in Fig. 11.
- the motion vector is searched for all 4 X 4 pixel blocks obtained by dividing the current macro block of 16 X 16 pixels into 16 parts.
- MV search processing min S per block Get AD value.
- step S A 3 the i nter D & Q process for performing interframe coding (integer DCT, quantization, inverse quantization, and inverse integer DCT) is executed.
- step S A 4 whether or not the mi n S AD value obtained in the i n ter prediction process in step S A 2 exceeds the threshold value TH 1, that is, the aforementioned error situation determination is executed.
- step S A 5 it is determined whether or not the flag I NTRA—F L G (described later) is “1”.
- the flag I NT R A—F L G is a flag generated by the int D R & Q process (described later), and is set to “1” when visually noticeable noise exists in the current macroblock. Therefore, this step S A 5 corresponds to the first encoding determination described above.
- step SA 5 If there is a visually noticeable noise in the current macroblock, the flag I NT RA—FLG is set to “1”, so the judgment result in step SA 5 is “YES”, and step SA 7 Go ahead and execute intra D & Q processing that encodes within the frame. On the other hand, if no visually noticeable noise exists in the current macroblock, the flag I NTRA—FLG is not set to “1”, and the judgment result in step SA 5 is “NO”. To complete.
- step SA 6 the second coding determination is performed based on the determination condition represented by the above-described equation [1]. In other words, it is determined whether there is visually noticeable noise in the current macroblock after considering the size of the motion vector. Judge that there is visually noticeable noise The result is “YE S”, and the process proceeds to Step SA 7 to execute intra D & Q processing. On the other hand, if there is no visually noticeable noise, the determination result is “NO”, and the process is completed.
- step SB 1 shown in FIG. 5 to divide the current macro block of 16 ⁇ 16 pixels into 16 ⁇ 4 ⁇ 4 Determine if all of the blocks of pixels have been processed.
- the judgment result is “YE S” and this process is completed. Otherwise, the judgment result is “NO” and the process proceeds to step S B 2.
- step S B 2 the flatness f 1 at (n) of the current block is calculated from the absolute sum of the harmonic component coefficients of the Hadamard coefficient obtained by Hadamard transform for each block.
- steps SB 3 to SB 7 a prediction error generated by subtracting the inter-frame prediction block value (reference image pixel value) ref (i, j) from the current block pixel value O rg (i, j)
- the signal (difference value An (i, j)) is subjected to transformation processing (integer DCT), quantization processing Q, inverse quantization Q- 1 and inverse transformation processing (inverse integer DCT).
- transformation processing integer DCT
- quantization processing Q quantization processing
- inverse quantization Q- 1 inverse quantization processing
- inverse integer DCT inverse integer DCT
- step SB8 the maximum quantization error value max No ise (n) in the 4 X4 pixels constituting the current block is detected.
- step SB9 the inter-frame prediction block value ref (i, j) is incremented.
- step SB 1 the noise value Noise (n) of the current block is larger than the threshold TH 2, the flatness flat (n) of the current block is smaller than the threshold TH 3, and the maximum quantization error ma
- the first encoding determination described above is executed to determine whether XNoise (n) is smaller than the threshold TH 10, that is, the presence or absence of visually noticeable noise.
- step S B 11 If visually noticeable noise is present, the judgment result is “Y E S”, the process proceeds to step S B 11, the flag I NTRA—FLG is set to “1”, and the process returns to step S B 1. On the other hand, if there is no visually noticeable noise, the determination result is “NO”, and the process returns to step S B 1.
- the min S AD value obtained by the motion vector search at the time of inter prediction is substituted for the evaluation value representing the total quantization error ⁇ Noise of the current macroblock to be encoded. Then, an error situation determination is performed in which the magnitude of the quantization error is roughly determined based on whether or not the min S AD value exceeds the threshold TH 1. If it is assumed that the quantization error is small in this error situation determination, the flatness f 1 at and noise N noise detected for each block of 4 ⁇ 4 pixels obtained by dividing the current macroblock into 16 Based on this, it is judged whether there is visually noticeable noise in some pixels of the current macro block. If there is visually noticeable noise, intra-frame coding is selected. On the other hand, if the error situation is judged to have a large quantization error, it is determined whether there is any visually noticeable noise in the current macroblock after considering the size of the motion vector. If there is noticeable noise, select intra-frame coding.
- the intra-frame coding is selected immediately. It is also possible to quickly cut the quantization error propagation in the case of a culminating minimum and to avoid image quality degradation.
- each part of the apparatus is initialized in step SC 1 and then proceeds to step SC 2 to generate a code for inter-frame predictive coding.
- the inter prediction process for estimating the quantity is executed, and the min S AD value for each block of 4 ⁇ 4 pixels obtained by dividing the current macroblock of 16 ⁇ 16 pixels by 16 is obtained.
- step SC 3 an error situation determination is performed to determine whether the min S AD value obtained in the inter prediction process in step SC 2 exceeds the threshold TH 1, that is, whether the quantization error is small. Do. If it is determined that the quantization error is small, the determination result is “NO”, and the process proceeds to step S C 4.
- step SC4 while performing interframe coding, visually noticeable noise is generated in some pixels of the current macroblock based on the flatness f 1 at and noise Noise detected for each 4 ⁇ 4 pixel block. If there is any visually noticeable noise, execute the inter D & Q process (see Fig. 5) that sets the flag I NTR A— FLG to “1”.
- step SC 5 whether the flag I NTR A—FLG is “1” No, that is, whether visually noticeable noise is detected in the inter D & Q process of step SC4. If visually noticeable noise is detected, the judgment result will be “YE Sj, and go to step SC 6 to execute intra D & Q processing that performs intra-frame coding. On the other hand, visually noticeable If no noise is detected, the result of step SC 5 is “NO”, and this process is completed.
- step S C 3 determines whether the quantization error is large. If it is determined that the quantization error is large, the determination result of step S C 3 is “YES”, and the process proceeds to step S C 7.
- step SC7 as in the first embodiment, there is visually noticeable noise in the current macroblock, taking into account the size of the motion vector, based on the judgment condition expressed by the above equation [1]. Determine whether to do. If visually noticeable noise does not exist in the current macroblock, the judgment result is “N 0” and the process proceeds to step S C 8.
- step S C 8 the i nt err D & Q process (see Fig. 15) that performs interframe coding is executed.
- the judgment result in step SC7 is “YE S”, and the process proceeds to step SC6 where intra-frame coding is performed. The process is executed to finish the present process.
- the min S AD value obtained by the motion vector search at the time of inter prediction is substituted for the evaluation value representing the total quantization error ⁇ Noise of the current macroblock to be encoded. If the min S AD value exceeds the threshold TH 1, an error situation determination is performed to roughly determine the magnitude of the quantization error. If it is assumed that the quantization error is small, Based on flatness flatness and noise Noise detected for each 4 x 4 pixel block obtained by dividing the current macroblock into 16 Next, it is determined whether there is a visually noticeable noise in some pixels of the current macroblock. Intraframe coding is selected when there is visually noticeable noise.
- the error situation is judged to have a large quantization error, it is determined whether or not there is visually noticeable noise in the current macroblock in consideration of the size of the motion vector. . If there is no visually noticeable noise, choose inter-frame coding, while if there is a visually noticeable noise, choose intra-frame coding.
- intra prediction processing with a large amount of computation (see Fig. 14) is omitted, so that unnecessary computation is not performed and the coding mode is determined at high speed in consideration of human visual characteristics. It is possible to do. Also, based on the flatness f 1 at and the noise noise, it is judged whether there is any noticeable noise in the current macroblock. If there is visually noticeable noise, the intra-frame coding is selected immediately. It is also possible to quickly stop the quantization error propagation in the case of falling into a mum and avoid the deterioration of image quality.
- each part of the device is initialized in step SD 1 and then proceeds to step SD 2 to generate a code for inter-frame predictive coding.
- the inter prediction process to estimate the quantity is executed, and the min SAD value for each block of 4 X 4 pixels obtained by dividing the current macroblock of 16 X 16 pixels into 16 is obtained.
- step SD3 an error status determination is performed to determine whether the min SAD value obtained in the inter prediction process in step SD2 exceeds the threshold TH1, that is, whether the quantization error is small. I do.
- step S D 4 the flatness F 1 at of the current macroblock is calculated from the absolute sum of the harmonic component coefficients of the Hadamard coefficient obtained by Hadamard transform for each block of 4 ⁇ 4 pixels.
- step S D 5 it is determined whether or not the calculated flatness F 1 at is smaller than the threshold TH 3, that is, whether or not the noise is a visually noticeable flatness.
- step SD8 If the flatness is visually inconspicuous, the judgment result will be “NOj, go to step SD8 and perform interframe coding (integer D CT, quantization, inverse quantization and inverse integer D CT) inter. Execute D & Q processing (see Fig. 16) On the other hand, if the level is flat enough for visually noticeable noise, the judgment result in step SD 5 above is “YES”, and step SD 6 Go ahead and execute intra D & Q processing that performs intra-frame coding.
- Step S D 7 as in the first embodiment, there is visually noticeable noise in the current macroblock based on the judgment condition expressed by the above equation [1], taking into account the size of the motion vector. Determine whether to do. If visually noticeable noise does not exist in the current macroblock, the determination result is ⁇ 0 ”, and the process proceeds to step S D 8 to execute the i nter d & Q process (see Fig. 15) that performs interframe coding.
- step SD7 determines whether visually noticeable noise exists in the current macroblock. If visually noticeable noise exists in the current macroblock, the judgment result in step SD7 is “YES”, and the process proceeds to step SD6, where intra-frame coding is performed. & Q processing To finish this processing.
- the min SAD value obtained by the motion vector search at the time of inter prediction is substituted for the evaluation value representing the total quantization error ⁇ Noise of the current macroblock to be encoded, If the min SAD value exceeds the threshold TH1, an error situation determination is performed to roughly determine the magnitude of the quantization error. If it is assumed that the quantization error is small, the current error is determined. The flatness of the macroblock is calculated, and it is judged whether the flatness is visually inconspicuous.
- inter-frame coding is performed. If the noise is visually noticeable, intra-frame coding is performed. On the other hand, if it is considered that the quantization error is large, it is determined whether there is a visually noticeable noise in the current macroblock in consideration of the size of the motion vector, and visually noticeable noise. If there is no, select interframe coding. On the other hand, if there is visually noticeable noise, select intraframe coding.
- intra prediction processing with a large amount of computation (see Fig. 14) is omitted, so that unnecessary computation is not performed and the coding mode is determined at high speed in consideration of human visual characteristics.
- intra-frame coding it immediately selects intra-frame coding, so it quickly cuts off the quantization error propagation when it falls into the local minimum, and avoids image quality degradation. It becomes possible.
- each part of the device is initialized in step SE 1 and then proceeds to step SE 2 to generate a code for inter-frame predictive coding.
- the inter prediction process for estimating the quantity is executed.
- step SE3 while performing interframe coding, some pixels of the current macroblock are visually identified based on the flatness F 1 at and the noise Noise detected for each 4 ⁇ 4 pixel block. Determine if there is noticeable noise, and if there is visually noticeable noise, execute the inter D & Q process (see Fig. 5) that sets the flag I NTR A— FLG to “1”.
- step SE 4 whether or not the total noise Noise obtained by the inter D & Q process in step SE 3 above, that is, whether the total quantization error ⁇ N 0 ise in the current macro block exceeds the threshold TH 1, In other words, it is determined whether the total quantization error ⁇ No ise is small. If it is determined that the total quantization error ⁇ N o i s e is small, the determination result is “NO”, and the process proceeds to step S E 5. Step SE 5 determines whether or not the flag I NTR A— FLG is “1”, that is, whether visually noticeable noise is detected in the inter D & Q process in Step SE 3 above. .
- step S E 7 the process proceeds to step S E 7 to execute intra D & Q processing that performs intraframe coding.
- the judgment result in step S E 5 is “NO”, and this processing is completed.
- step S E 7 based on the determination condition expressed by the following equation [2], it is determined whether or not visually noticeable noise exists in the current macroblock in consideration of the size of the motion vector.
- coefficient a adjusts the contribution of the motion vector.
- the constant and coefficient k are constants corresponding to the slope of the straight line B described above.
- the function f (mv) is the magnitude (u 2 + V 2 ) of the motion vector (u, V) of the current macroblock.
- step S E 6 determines whether there is visually noticeable noise. If there is visually noticeable noise, the determination result in step S E 6 is “YES”, and the process proceeds to step S E 7 to execute int ra D & Q processing for intra-frame coding. On the other hand, if there is no visually noticeable noise, the judgment result will be “NO” and this process will be completed.
- the magnitude of the quantization error of the current macroblock is determined based on whether or not the total quantization error ⁇ Noise exceeds the threshold value TH1, and if the quantization error is small, Noise that is visually noticeable in some pixels of the current macroblock is based on the detected flatness f 1 at and noise of each 4 ⁇ 4 pixel block divided by 16 for the current macroblock. Determine if there is. If there is a visually noticeable noise, select intra-frame coding. On the other hand, if the quantization error is large, it is determined whether there is visually noticeable noise in the current macroblock after considering the size of the motion vector, and if there is visually noticeable noise. Select intra-frame coding.
- intra prediction processing with a large amount of computation (see Fig. 14) is omitted, so that unnecessary computation is not performed and the coding mode is determined at high speed in consideration of human visual characteristics. It is possible to do. Also, based on the flatness of flatness and noise noise, it is judged whether there is any noticeable noise in the current macroblock, and if there is visually noticeable noise, the intra-frame coding is selected immediately. It is also possible to quickly stop the quantization error propagation in the event of a fall, and avoid image quality degradation.
- the processing product of the moving picture coding apparatus Is stored in the memory (eg, ROM) of the video encoding device, but should be protected even when this processing program is stored in a recording medium and only the program is manufactured or sold. is there.
- the method for protecting the program with a patent is realized in the form of a recording medium in which a moving image encoding processing program is stored.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
Claims
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US11/504,021 US8121193B2 (en) | 2004-11-04 | 2006-08-14 | Motion picture encoding device and motion picture encoding processing program |
| US13/351,711 US8824552B2 (en) | 2004-11-04 | 2012-01-17 | Motion picture encoding device and motion picture encoding processing program |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2004-320088 | 2004-11-04 | ||
| JP2004320088A JP4877449B2 (ja) | 2004-11-04 | 2004-11-04 | 動画像符号化装置および動画像符号化処理プログラム |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US11/504,021 Continuation US8121193B2 (en) | 2004-11-04 | 2006-08-14 | Motion picture encoding device and motion picture encoding processing program |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2006049335A1 true WO2006049335A1 (ja) | 2006-05-11 |
Family
ID=36319322
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2005/020689 Ceased WO2006049335A1 (ja) | 2004-11-04 | 2005-11-04 | 動画像符号化装置および動画像符号化処理プログラム |
Country Status (3)
| Country | Link |
|---|---|
| US (2) | US8121193B2 (ja) |
| JP (1) | JP4877449B2 (ja) |
| WO (1) | WO2006049335A1 (ja) |
Families Citing this family (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006254225A (ja) * | 2005-03-11 | 2006-09-21 | Toshiba Corp | 可変長符号の復号装置及び復号方法 |
| PL3306923T3 (pl) * | 2005-11-08 | 2020-04-30 | Sun Patent Trust | Sposób kodowania ruchomych obrazów, sposób dekodowania ruchomych obrazów oraz urządzenia służące do tego celu |
| US7667776B2 (en) * | 2006-02-06 | 2010-02-23 | Vixs Systems, Inc. | Video display device, video encoder, noise level estimation module and methods for use therewith |
| JP4901450B2 (ja) * | 2006-12-19 | 2012-03-21 | 株式会社東芝 | 動画像符号化装置 |
| US8737485B2 (en) * | 2007-01-31 | 2014-05-27 | Sony Corporation | Video coding mode selection system |
| JP4748603B2 (ja) * | 2007-02-28 | 2011-08-17 | 株式会社Kddi研究所 | 動画像符号化装置 |
| JP4739295B2 (ja) * | 2007-08-27 | 2011-08-03 | 日本電信電話株式会社 | 映像信号生成装置、映像信号生成方法、映像信号生成プログラムおよびそのプログラムを記録したコンピュータ読み取り可能な記録媒体 |
| KR101496324B1 (ko) * | 2007-10-17 | 2015-02-26 | 삼성전자주식회사 | 영상의 부호화, 복호화 방법 및 장치 |
| US8363719B2 (en) * | 2007-10-29 | 2013-01-29 | Canon Kabushiki Kaisha | Encoding apparatus, method of controlling thereof, and computer program |
| JP5081729B2 (ja) * | 2008-06-03 | 2012-11-28 | 株式会社日立国際電気 | 画像符号化装置 |
| US8503527B2 (en) | 2008-10-03 | 2013-08-06 | Qualcomm Incorporated | Video coding with large macroblocks |
| JP5339855B2 (ja) * | 2008-10-31 | 2013-11-13 | キヤノン株式会社 | 動きベクトル探索装置、動きベクトル探索方法、画像処理装置、画像処理方法、プログラム及び記憶媒体 |
| KR101538704B1 (ko) * | 2009-01-28 | 2015-07-28 | 삼성전자주식회사 | 보간 필터를 적응적으로 사용하여 영상을 부호화 및 복호화하는 방법 및 장치 |
| 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 |
| JP5215951B2 (ja) * | 2009-07-01 | 2013-06-19 | キヤノン株式会社 | 符号化装置及びその制御方法、コンピュータプログラム |
| US20120019727A1 (en) * | 2010-07-21 | 2012-01-26 | Fan Zhai | Efficient Motion-Adaptive Noise Reduction Scheme for Video Signals |
| US9532059B2 (en) | 2010-10-05 | 2016-12-27 | Google Technology Holdings LLC | Method and apparatus for spatial scalability for video coding |
| US8787454B1 (en) * | 2011-07-13 | 2014-07-22 | Google Inc. | Method and apparatus for data compression using content-based features |
| US9094681B1 (en) | 2012-02-28 | 2015-07-28 | Google Inc. | Adaptive segmentation |
| US11317101B2 (en) * | 2012-06-12 | 2022-04-26 | Google Inc. | Inter frame candidate selection for a video encoder |
| US8396127B1 (en) * | 2012-06-27 | 2013-03-12 | Google Inc. | Segmentation for video coding using predictive benefit |
| JP2012227955A (ja) * | 2012-07-12 | 2012-11-15 | Hitachi Kokusai Electric Inc | 画像符号化装置 |
| US9332276B1 (en) | 2012-08-09 | 2016-05-03 | Google Inc. | Variable-sized super block based direct prediction mode |
| US9380298B1 (en) | 2012-08-10 | 2016-06-28 | Google Inc. | Object-based intra-prediction |
| US9288484B1 (en) | 2012-08-30 | 2016-03-15 | Google Inc. | Sparse coding dictionary priming |
| US9210432B2 (en) * | 2012-10-08 | 2015-12-08 | Google Inc. | Lossless inter-frame video coding |
| US9369732B2 (en) | 2012-10-08 | 2016-06-14 | Google Inc. | Lossless intra-prediction video coding |
| US9756346B2 (en) | 2012-10-08 | 2017-09-05 | Google Inc. | Edge-selective intra coding |
| US9407915B2 (en) | 2012-10-08 | 2016-08-02 | Google Inc. | Lossless video coding with sub-frame level optimal quantization values |
| US9300906B2 (en) * | 2013-03-29 | 2016-03-29 | Google Inc. | Pull frame interpolation |
| US9106888B2 (en) * | 2013-09-25 | 2015-08-11 | Apple Inc. | Reducing quantization artifacts using neighbor-based weighted dithering |
| US9286653B2 (en) | 2014-08-06 | 2016-03-15 | Google Inc. | System and method for increasing the bit depth of images |
| JP6052319B2 (ja) * | 2015-03-25 | 2016-12-27 | Nttエレクトロニクス株式会社 | 動画像符号化装置 |
| DE102016207017A1 (de) | 2016-04-26 | 2017-10-26 | Trumpf Laser Gmbh | Pumplichtanordnung, Scheibenlaser damit und Verfahren zum Pumpen eines laseraktiven Mediums |
| EP3370419B1 (en) | 2017-03-02 | 2019-02-13 | Axis AB | A video encoder and a method in a video encoder |
| CN116996693A (zh) * | 2019-06-21 | 2023-11-03 | 华为技术有限公司 | 编码器、解码器及对应方法 |
| SE2351198A1 (en) * | 2021-04-19 | 2023-10-18 | Hanwha Vision Co Ltd | Noise removal for surveillance camera image by means of ai-based object recognition |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63121372A (ja) * | 1986-11-10 | 1988-05-25 | Kokusai Denshin Denwa Co Ltd <Kdd> | 動画像信号のハイブリツド符号化方式 |
| JPH05236462A (ja) * | 1992-02-17 | 1993-09-10 | Sony Corp | 画像データ処理装置 |
| JPH1079948A (ja) * | 1996-09-03 | 1998-03-24 | Mitsubishi Electric Corp | 画像符号化装置 |
| JP2000138937A (ja) * | 1998-10-30 | 2000-05-16 | Sony Corp | 画像圧縮符号化方法及び装置 |
| JP2003037844A (ja) * | 2001-07-23 | 2003-02-07 | Sharp Corp | 動画像符号化装置 |
Family Cites Families (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0557007A2 (en) | 1992-02-15 | 1993-08-25 | Sony Corporation | Picture processing apparatus |
| US5436665A (en) * | 1992-03-03 | 1995-07-25 | Kabushiki Kaisha Toshiba | Motion picture coding apparatus |
| US5387938A (en) * | 1992-10-08 | 1995-02-07 | Matsushita Electric Industrial Co., Ltd. | Adaptive interframe/intraframe block coding method and apparatus |
| US5724097A (en) * | 1993-10-18 | 1998-03-03 | Mitsubishi Denki Kabushiki Kaisha | Adaptive quantization of video based on edge detection |
| JP3954656B2 (ja) * | 1994-09-29 | 2007-08-08 | ソニー株式会社 | 画像符号化装置及び方法 |
| US6415057B1 (en) * | 1995-04-07 | 2002-07-02 | Sony Corporation | Method and apparatus for selective control of degree of picture compression |
| US6151360A (en) * | 1995-04-28 | 2000-11-21 | Sony Corporation | Method for encoding video signal using statistical information |
| JPH09163373A (ja) * | 1995-12-08 | 1997-06-20 | Toshiba Corp | 雑音低減装置 |
| JPH09214975A (ja) * | 1996-02-02 | 1997-08-15 | Graphics Commun Lab:Kk | シーンチェンジ適応動画像符号化方法および装置 |
| GB2311184A (en) * | 1996-03-13 | 1997-09-17 | Innovision Plc | Motion vector field error estimation |
| JPH09294262A (ja) * | 1996-04-26 | 1997-11-11 | Hitachi Ltd | 画像符号化装置 |
| JP3633159B2 (ja) * | 1996-12-18 | 2005-03-30 | ソニー株式会社 | 動画像信号符号化方法及び装置、並びに動画像信号伝送方法 |
| JP3633204B2 (ja) * | 1997-05-14 | 2005-03-30 | ソニー株式会社 | 信号符号化装置、信号符号化方法、信号記録媒体及び信号伝送方法 |
| EP0921497B1 (en) * | 1997-12-02 | 2004-03-24 | Daewoo Electronics Corporation | Interlaced binary shape coding apparatus |
| JP2000134631A (ja) * | 1998-10-23 | 2000-05-12 | Canon Inc | 画像符号化装置、方法、画像復号化装置、方法及びコンピュータ読み取り可能な記憶媒体 |
| US6625211B1 (en) * | 1999-02-25 | 2003-09-23 | Matsushita Electric Industrial Co., Ltd. | Method and apparatus for transforming moving picture coding system |
| US6295089B1 (en) * | 1999-03-30 | 2001-09-25 | Sony Corporation | Unsampled hd MPEG video and half-pel motion compensation |
| JP3630590B2 (ja) * | 1999-08-25 | 2005-03-16 | 沖電気工業株式会社 | 復号化装置及び伝送システム |
| US7266148B2 (en) * | 2001-01-05 | 2007-09-04 | Lg Electronics Inc. | Video transcoding apparatus |
| JP4644939B2 (ja) * | 2001-01-17 | 2011-03-09 | 日本電気株式会社 | 動画像符号化装置および動画像符号化方法 |
| ATE324013T1 (de) * | 2001-03-12 | 2006-05-15 | Koninkl Philips Electronics Nv | Vorrichtung zur videokodierung und -aufzeichnung |
| US6831947B2 (en) * | 2001-03-23 | 2004-12-14 | Sharp Laboratories Of America, Inc. | Adaptive quantization based on bit rate prediction and prediction error energy |
| WO2002102086A2 (en) * | 2001-06-12 | 2002-12-19 | Miranda Technologies Inc. | Apparatus and method for adaptive spatial segmentation-based noise reducing for encoded image signal |
| US7944971B1 (en) * | 2002-07-14 | 2011-05-17 | Apple Inc. | Encoding video |
| US7173971B2 (en) * | 2002-11-20 | 2007-02-06 | Ub Video Inc. | Trailing artifact avoidance system and method |
| JP2006519564A (ja) * | 2003-03-03 | 2006-08-24 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | ビデオ符号化 |
| US20060188014A1 (en) * | 2005-02-23 | 2006-08-24 | Civanlar M R | Video coding and adaptation by semantics-driven resolution control for transport and storage |
| US8031946B2 (en) * | 2008-03-27 | 2011-10-04 | Texas Instruments Incorporated | Reduced calculations in determining intra-prediction type method and system |
| JP5156704B2 (ja) * | 2008-07-29 | 2013-03-06 | パナソニック株式会社 | 画像符号化装置、画像符号化方法、集積回路及びカメラ |
-
2004
- 2004-11-04 JP JP2004320088A patent/JP4877449B2/ja not_active Expired - Fee Related
-
2005
- 2005-11-04 WO PCT/JP2005/020689 patent/WO2006049335A1/ja not_active Ceased
-
2006
- 2006-08-14 US US11/504,021 patent/US8121193B2/en active Active
-
2012
- 2012-01-17 US US13/351,711 patent/US8824552B2/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63121372A (ja) * | 1986-11-10 | 1988-05-25 | Kokusai Denshin Denwa Co Ltd <Kdd> | 動画像信号のハイブリツド符号化方式 |
| JPH05236462A (ja) * | 1992-02-17 | 1993-09-10 | Sony Corp | 画像データ処理装置 |
| JPH1079948A (ja) * | 1996-09-03 | 1998-03-24 | Mitsubishi Electric Corp | 画像符号化装置 |
| JP2000138937A (ja) * | 1998-10-30 | 2000-05-16 | Sony Corp | 画像圧縮符号化方法及び装置 |
| JP2003037844A (ja) * | 2001-07-23 | 2003-02-07 | Sharp Corp | 動画像符号化装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| US8121193B2 (en) | 2012-02-21 |
| US8824552B2 (en) | 2014-09-02 |
| US20120114035A1 (en) | 2012-05-10 |
| US20070019729A1 (en) | 2007-01-25 |
| JP2006135461A (ja) | 2006-05-25 |
| JP4877449B2 (ja) | 2012-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP4877449B2 (ja) | 動画像符号化装置および動画像符号化処理プログラム | |
| KR100391027B1 (ko) | 비디오정보를예측부호화하는방법및장치 | |
| JP5497169B2 (ja) | ビデオコーディングにおける単方向予測および双方向予測のための異なる重み付け | |
| US7983493B2 (en) | Adaptive overlapped block matching for accurate motion compensation | |
| KR101375892B1 (ko) | 영상 캡처 파라미터들에 기초한 비디오 인코딩 제어 | |
| JP2006519565A (ja) | ビデオ符号化 | |
| JP5133290B2 (ja) | 動画像符号化装置および復号装置 | |
| CN103609122B (zh) | 用于视频编码的率失真优化 | |
| JP5640979B2 (ja) | 動画像符号化装置、動画像符号化方法および動画像符号化プログラム | |
| JP2006211152A (ja) | 画像符号化装置、画像復号装置、画像符号化方法、画像復号方法、画像符号化用プログラム、画像復号用プログラム | |
| JP2005286645A (ja) | 画像符号化装置、画像符号化方法および画像符号化プログラム | |
| JP5173409B2 (ja) | 符号化装置および符号化装置を備えた動画像記録システム | |
| JP4257655B2 (ja) | 動画像符号化装置 | |
| JP3982615B2 (ja) | 符号化動画像データの方式変換装置 | |
| JP5708490B2 (ja) | 動画像符号化装置、動画像符号化方法および動画像符号化プログラム | |
| JP2010239423A (ja) | 撮影解像度予測型動画像符号化装置および復号装置 | |
| JP2001238215A (ja) | 動画像符号化装置及びその方法 | |
| JP4375143B2 (ja) | 動画像符号化装置 | |
| JP2019087985A (ja) | 画像符号化装置及び画像符号化プログラム | |
| JP2005303555A (ja) | 動画像符号化装置および動画像符号化方法 | |
| JP2009130864A (ja) | 画像処理装置、およびそれを搭載した撮像装置 | |
| JP2020102798A (ja) | 動画像符号化装置、イントラリフレッシュ制御装置およびそれらのプログラム | |
| JP4100067B2 (ja) | 画像情報変換方法及び画像情報変換装置 | |
| WO2010116708A1 (ja) | 動画像符号化装置、動画像復号装置、動画像符号化方法及び動画像復号方法 | |
| JP4000581B2 (ja) | 画像符号化装置および方法 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BW BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE EG ES FI GB GD GE GH GM HR HU ID IL IN IS KE KG KM KN KP KR KZ LC LK LR LS LT LU LV LY MA MD MG MK MN MW MX MZ NA NG NI NO NZ OM PG PH PL PT RO RU SC SD SE SG SK SL SM SY TJ TM TN TR TT TZ UA UG US UZ VC VN YU ZA ZM ZW |
|
| AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): BW GH GM KE LS MW MZ NA SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LT LU LV MC NL PL PT RO SE SI SK TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 11504021 Country of ref document: US |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
| WWP | Wipo information: published in national office |
Ref document number: 11504021 Country of ref document: US |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 05805972 Country of ref document: EP Kind code of ref document: A1 |