WO1998042135A1 - Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system - Google Patents
Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system Download PDFInfo
- Publication number
- WO1998042135A1 WO1998042135A1 PCT/JP1997/003825 JP9703825W WO9842135A1 WO 1998042135 A1 WO1998042135 A1 WO 1998042135A1 JP 9703825 W JP9703825 W JP 9703825W WO 9842135 A1 WO9842135 A1 WO 9842135A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- motion
- unit
- pixel
- motion compensation
- 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
- H04N19/50—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
- H04N19/50—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
- H04N19/523—Motion estimation or motion compensation with sub-pixel accuracy
-
- 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
- H04N19/53—Multi-resolution motion estimation; Hierarchical motion 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
- H04N19/50—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
- H04N19/537—Motion estimation other than block-based
-
- 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
- Image coding apparatus image decoding apparatus, image coding method, image decoding method, and image coding / decoding system
- the present invention provides motion-compensated prediction of an image to be encoded or an image to be decoded from an existing image and encodes a prediction error, or encodes a prediction error and reference image data in high-efficiency encoding or decoding of an image.
- the present invention relates to an apparatus and a system that perform decoding by addition. Background art
- the conventional motion compensation prediction method there is a motion compensation prediction method using block matching by parallel movement.
- ISOZOIEC 11 1 17 22 -2 MPEG 1 Video Standard
- ISOZOIEC 11 1 17 22 -2 MPEG 1 Video Standard
- motion compensation using an affine transformation For example, in “Study on Motion Compensation Prediction Using Affine Transformation” (IEICE Technical Report IE 94-36), the amount of motion in an arbitrary shape region of an image is modeled using affine motion parameters. Explains how to perform motion compensation prediction by detecting parameters.
- Figure 42 shows the concept of motion compensated prediction by block matching.
- f i (x, y, t) is the pixel value at the position (X, y) of the block at screen position i and time t,
- R is the motion vector search range
- V is the motion vector (ER)
- the value of v that minimizes D v is the motion vector.
- the method of searching for a matching block using only the actual pixels that are the actual sample points in the reference image is an integer pixel precision search
- the search method that uses the half pixels in the middle of the integer pixels in addition to the integer pixels is a half pixel Called accuracy search.
- the half-pixel accuracy search has more search points than the integer-pixel accuracy search, and the prediction efficiency increases.
- FIG. 43 is a diagram illustrating a configuration example of a motion compensation prediction unit (block matching unit) of an image encoding device that uses a motion compensation prediction method adopted in, for example, the MPEG1 video standard.
- 207 is a horizontal movement amount counter
- 20 8 is a vertical movement
- a row movement amount counter 211 is a memory read address generation unit
- 213 is a pattern matching unit
- 216 is a minimum prediction error power determination unit.
- 203 is a horizontal parallel movement amount search range instruction signal
- 204 is a vertical parallel movement amount search range instruction signal
- 205 is predicted block data
- 206 is a position signal in the image of the predicted block
- 209 is Horizontal parallel movement amount search point data
- 210 is vertical parallel movement amount search point data
- 2 12 is read address
- 2 14 is read image data
- 2 15 is prediction error power signal
- 2 1 7 Is the motion vector
- 218 is the minimum prediction error power signal.
- Reference numeral 219 denotes a frame memory for storing a reference image.
- FIG. 44 is an operation flowchart showing the operation of the motion compensation prediction unit having the configuration shown in FIG.
- range—h—min is the lower limit of the horizontal translation search range
- range—h—max is the upper limit of the horizontal translation search range
- range—V—min is the lower limit of the vertical translation search range.
- Rang e_v_max is the upper limit of the search range for vertical translation
- D-min is the minimum prediction error power
- (x, y) are coordinates representing the pixel position in the macroblock
- D (dX, dy) is the prediction error power at the time of dx, dy search
- f (x, y) is the value of the pixel (X, y) of the predicted image
- f r (x, y) is the value of pixel (x, y) in the reference image
- D (X, y) is the prediction error at (x, y) when searching for d x, d y, MV—h is the motion vector (parallel movement amount) horizontal component,
- MV V is the motion vector (translation amount) vertical component It is.
- the minimum prediction error power D—min is set to the maximum value MAXINT (for example, OxFFFFFFFF). This is equivalent to S201 in FIG.
- the pixel at the position (x + dx, y + dy) in the reference image that is (dx, dy) away from the pixel position (X, y) of the macroblock to be predicted is extracted from the frame memory.
- the memory read address generator 2 11 in FIG. 4 3 receives the value of dX from the horizontal movement amount counter 2 07 and the value of dy from the vertical parallel movement amount counter 2 08, and Generate an address.
- the prediction error power D (dx, dy) when the motion vector is (dx, dy) is initialized to zero. This corresponds to S202 in FIG.
- the measurement error power D (dx, dy) that is, Dv in equation (1) is obtained.
- This processing is performed by the pattern matching unit 2 13 in FIG. 43.
- the pattern matching unit 2 13 uses D (d X, dy) as the minimum prediction error power determination unit 2 based on the prediction error power signal 2 15. Hand over to 6. This processing corresponds to the processing of S203 to S209 in FIG.
- the minimum prediction error power determination unit 216 compares the value of the internal minimum prediction error power D min with the magnitude of D (dX, dy) passed by the prediction error power signal 215. , D (dX, dy) is updated only when D—min is smaller than D (dx, dy). Also, the value of (d x, d y) at that time is stored as a motion vector candidate (MV-h, MV-V). These updating processes correspond to S211 in FIG.
- Fig. 45 is a diagram showing an outline of the motion compensation prediction method adopted in the MPEG1 video standard.
- each frame of a moving image is called a picture, and the picture is divided into 16 x 16 pixel blocks (color difference signals are 8 x 8 pixels) called macroblocks.
- 16 x 16 pixel blocks color difference signals are 8 x 8 pixels
- macroblocks 16 x 16 pixel blocks
- Block Perform motion compensation prediction by matching.
- the resulting motion vector and prediction error signal are encoded.
- the motion compensation method can be changed for each different picture.
- the I-picture performs closed coding within the picture without performing motion compensation prediction
- the P-picture Forward motion compensation prediction which performs prediction from the image displayed earlier
- forward motion compensation prediction which performs prediction from the image displayed later in time
- interpolation motion compensation prediction in which prediction is performed by averaging two predicted images obtained from forward motion compensation prediction and backward motion compensation prediction, is allowed.
- each motion-compensated prediction in the forward-backward interpolation is basically a motion-compensated prediction based on block matching, only in the difference between reference images used for prediction.
- block matching has been established as the main method for implementing motion-compensated prediction in current video coding methods.
- the process of block matching is based on the assumption of equal luminance, that is, "areas with the same luminance are the same object", and obtains the amount of translation of the object in units of square blocks such as macroblocks. Equivalent to. Therefore, in principle, it is impossible to detect movement other than movement in the direction of the square block shape, and parallel movement such as rotation, enlargement, reduction, camera zooming, and three-dimensional object movement is sufficient. Prediction accuracy drops in areas where unexplained movements occur
- Performing compensation prediction is motion compensation prediction using affine transformation.
- the pixel value (x, y) to be predicted is transformed by the affine transformation shown in the following equation (2).
- Each parameter of the affine transformation is detected as a motion parameter based on the assumption that the pixel value ( ⁇ ', y') in the reference image is converted.
- affine motion parameters are detected for prediction image regions of arbitrary shapes, and motion compensation prediction is performed. A method is proposed. cos ⁇ sln ⁇ C x 0 x X
- FIG. 46 illustrates the concept of motion compensation prediction processing using affine transformation.
- i is the position in the screen of the area that is the unit of motion compensation prediction
- f i (x, y, t) is the pixel value at the position (X, y) of the region as the position i in the screen and the time t,
- Rv is the parallel displacement search range
- R r o t, s c a 1 e is the rotation Z scale amount search range
- r ot is a rotation parameter (two rotation angles ⁇ )
- affine motion parameters consisting of a total of five parameters, a rotation angle ⁇ and a scale (CX, Cy) are detected in addition to the parallel movements and 0 parameters (tX, ty) corresponding to the motion vector.
- the optimal solution is given by a full search of all parameters, but a very enormous performance
- a two-step search algorithm is adopted based on the assumption that the amount of parallel movement is dominant.
- the translation amount (tx, ty) of the region is searched.
- the procedure of searching for the rotation angle ⁇ and the scale (CX, Cy) near (tX, ty) determined in the first stage, and fine-adjusting the translation amount is also performed.
- the difference between the prediction region that gives the minimum prediction error power and the current region is calculated, and the prediction error is encoded.
- the prediction error power of the affine transformation method is expressed by the following equation (3).
- FIG. 47 is a diagram illustrating a configuration example of a motion compensation prediction unit using affine transformation.
- reference numeral 220 denotes a parallel movement fine adjustment amount search range instruction signal
- 2 2 1 Is a rotation amount search range instruction signal
- 2 2 2 is a scale amount search range instruction signal
- 2 2 3 is a translation amount search range instruction signal
- 2 2 4 is a position signal in the predicted area screen
- 2 2 5 is a predicted Area data
- 2 26 is a horizontal parallel movement counter
- 2 27 is a vertical parallel movement counter
- 2 28 is a parallel movement adder
- 2 209 is a first-stage minimum prediction error power determiner
- 2 30 is a memory read address generator
- 231 is an interpolation calculator
- 232 is a half-pixel generator
- 233 is a rotation counter
- 234 is a scale counter
- 23.5 is a parallel movement.
- Rotational no-scaling amount adder 236 is the second-stage minimum prediction error power determiner, 237 is the parallel movement fine adjustment amount counter, 238 is the parallel movement fine adjustment amount adder, and 239 is the final It is a minimum prediction error power determination unit.
- FIG. 48 is a flowchart of the operation of the conventional device.
- FIG. 49 is a flowchart showing details of the affine motion parameter detection process indicated by S224 in FIG.
- MV—h [4] is the horizontal component of the motion vector (4 candidates),
- MV_v [4] is the vertical component of the motion vector (4 signs),
- D__min is the minimum prediction error power
- ⁇ is the amount of rotation [r a d i a n]
- t x and t y are fine adjustments of the motion vector.
- D (6 [i], Cx [i], Cy [i], tx [i], ty [i]) is an affinity when MV-h [ ⁇ , MV-v [i] is selected.
- d C y is the vertical scale amount search point
- dt X is the horizontal translation fine adjustment amount search point
- range—radian—min is the lower limit of the rotation amount search range
- range—radian_max is the upper limit of the rotation amount search range
- range—seale_min is the lower limit of the scale amount search range
- range—sea 1 e _max is , Scale value search range upper limit value
- range—t—h—min are horizontal translation fine adjustment amount search range lower limit value
- r a n g e— t— h— m a X is the horizontal translation fine adjustment amount search range upper limit
- r ange e _ t _ v_min is the vertical translation fine adjustment amount search range lower limit
- r ange e_ t _ v_max is the vertical translation fine-adjustment amount search range upper limit
- D—min is the minimum prediction error power
- (x, y) is the pixel position in the region to be predicted
- f (, y) is the value of the pixel (X, y) of the predicted image
- fr (x, y) is the value of pixel (x, y ) in the reference image
- a x is the horizontal affine transformation value
- a y is the vertical affine transformation value
- D (a x, a y) is the prediction error power at the time of a x, a y search
- D (x, y) is the prediction error at (x, y) when searching for ax, ay.
- a search range is set in the horizontal movement amount counter 226 and the vertical movement amount counter 227 based on the parallel movement amount search range instruction signal 223, and the search point is changed.
- the parallel movement amount adding section 228 adds the current area position in the predicted image area to this count value, and passes the result to the memory read address generating section 230, where the pixel value of the predicted image candidate is calculated. It is read from the frame memory 219.
- the read pixel values are passed to the pattern matching section 21 and are subjected to the same error calculation as in the block matching.
- the matching result is sent to the first-stage minimum prediction error power determination unit 229, and four candidate translation parameters are obtained from the one with the smallest prediction error.
- the operation of the first-stage minimum prediction error power determination unit 229 is the same as that of the minimum prediction error power determination unit 216. This process corresponds to S221 and S222 in FIG.
- FIG. 49 shows a detailed processing procedure. The operation will be described in relation to the operation of the device shown in FIG.
- the rotation amount search range instruction signal 221 First, from the rotation amount search range instruction signal 221, and the scale amount search range instruction signal 222, the rotation amount counter 233 and the scale amount counter 234 are supplied. Set the search range for each. Also, the parallel movement fine adjustment amount search range instruction signal
- the search range is also set in the parallel movement fine adjustment amount counter 2 37.
- the second-stage minimum prediction error power determination unit 236 sets the value of the internal minimum prediction error power D-min to MAXINT. This corresponds to S 2 29 in FIG.
- the above processing is performed by fixing the count values of the scale amount counter 2 34 and the parallel movement fine adjustment amount counter 2 37 in FIG.
- Ax and ay of equation (4) are calculated by the translation / rotation / "scale amount addition unit 2 3 5 according to the count value of 3 3, and ir (ax, ay) is calculated via the memory read address generation unit 230.
- the pixels necessary for the calculation are read out from the frame memory 2 19, and then fr (ax, ay) is calculated from these pixels in the interpolation calculation unit 2 31, and the calculated values are sent to the pattern matching unit 2 13. This is performed by the operation of calculating the absolute value of the difference from the predicted pixel value f (x, y). In FIG. 49, it corresponds to S231 to S234.
- the above processing is performed over the entire rotation amount search range, and the second stage minimum prediction error power determination unit 236 determines the rotation amount ⁇ that gives the minimum prediction error within the rotation amount search range.
- the count value of the parallel movement fine adjustment amount counter 237 is fixed, and the rotation amount ⁇ determined in 2 2) as the rotation amount is substituted into equation (4) to obtain the scale amount.
- the values of C x and C y are changed within the search range to obtain the affine transformation values ax and ay of equation (4).
- the parallel movement fine adjustment amount search point is counted by the parallel movement fine adjustment amount counter 237.
- a half-pixel value is calculated by the half-pixel generation unit 232 as necessary before being sent to the pattern matching unit 21.
- the half-pixel value is calculated as shown in the following equation (5) based on the spatial positional relationship with the integer pixels as shown in FIG.
- both x and y are counted from 0, and the integer pixel position is both EVEN.
- I (x, y) [ I (xp, yp) + I (x P + 1 'yp) + I (x p, y p + 1 I (xp + 1, yp + 1)] / A; x, y: ODD
- the affine motion parameter search requires not only a very large number of processing steps, but also a large computational load in the search.
- FIG. 51 is a diagram showing a method of calculating a non-integer pixel value generated in the process of searching for the rotation amount and the scale amount, that is, a calculation method of fr (ax, ay) in the interpolation calculation unit 231.
- ⁇ is the actual sample point of the image
- ⁇ is the pixel value generated by the operation.
- a motion compensation device disclosed in Japanese Patent Application Laid-Open No. HEI 6-153185 and an encoding device using the same are disclosed. Have been.
- an image in a frame memory serving as a reference image is provided with a thinning-out circuit or an interpolation circuit to reduce or enlarge the image and then detect a motion vector.
- a complicated operation such as the affine transformation is not performed, but a fixed block is extracted from the reference image and interpolation or thinning operation is performed.
- a fixed screen area is cut out, processed in advance, and then compared with the input image, the processing is fixed, making it simple and practical for simple enlargement and reduction. Limited.
- the motion prediction method of the conventional image encoding device is configured and operates as described above.
- the predicted image area is formed by translating the cut-out area of the reference screen, so that only a simple translational motion can be predicted, such as rotation, enlargement, reduction, camera zooming, and the like.
- a simple translational motion such as rotation, enlargement, reduction, camera zooming, and the like.
- the present invention has been made to solve the above-described problem, and differs from an image area to be predicted using existing pixels on a reference screen or pixels obtained by a simple filtering process.
- the purpose of the present invention is to obtain a coding apparatus using a motion-compensated prediction method for an image which can cope with various kinds of motions and temporal changes by relatively simple processing by forming a predicted image area of a shape or a different size. I do.
- An image encoding apparatus is configured to divide an input image into predetermined blocks, and to provide a motion compensation prediction unit based on motion detection between frames of the blocks to compress and encode the input image. Deformation in which only integer pixels, which are actual sample points existing in the corresponding partial area of the reference image for motion detection, are transformed into a predetermined format, specified with coordinates, extracted, and compared with the integer pixels of the above block of the input image.
- a motion detector that outputs a motion vector that gives the minimum error extracted by specifying the coordinates, including the block matching unit, and a reference image according to the motion parameters obtained from the comparison output including the deformed block matching unit.
- a motion compensating unit that outputs a predicted partial image including a corresponding point determining unit that determines the corresponding block by specifying coordinates and deforming the block.
- the modified block matching unit is configured to perform the modification using the integer pixel and the half pixel that is the midpoint of the integer pixel when the partial area of the reference image is modified in a predetermined format.
- a pre-processing unit that separates the input image into a partial region of the image object as a coding target region is added, and each of the separated image objects is divided into a book to perform motion detection and Added motion compensation.
- the modified block matching unit and the corresponding point determining unit are configured to specify an integer pixel or a half pixel as coordinates, and to specify and extract an adjacent point or an adjacent point multiplied by a predetermined number, and perform a comparison.
- a corresponding point determining unit that processes and outputs the reference image is used.
- modified block matching unit and the corresponding point determination unit are configured to extract and compare integer pixels or half pixels by designating coordinates rotated in a predetermined angle direction, and to process the reference image similarly.
- the corresponding point determination unit outputs the data.
- Rotation in a given angle direction is positive or negative 45 degrees, 90 degrees, 135 degrees or 18
- the deformed block matching unit and the corresponding point determination unit search for the region indicated by the partial region of the reference image after the translation, and enlarge or reduce this search region, or combine rotation in a predetermined angular direction. And a corresponding point determination unit that similarly processes and outputs the reference image.
- the modified block matching unit includes a modified pattern table for modifying and comparing the partial area of the reference image, and converts the image of the partial area based on the conversion value extracted from the modified pattern table.
- a modified block matching unit for comparing with an integer pixel or a half pixel of a block of the input image was used, and a corresponding point determining unit was also a corresponding point determining unit for processing and outputting a reference image.
- the modified block matching unit selectively performs specific filtering on the specific pixels of the reference image extracted for the correspondence evaluation and compares the pixels.
- the frame for motion detection is the temporally preceding or succeeding frame
- the reference image is the temporally preceding or succeeding frame stored and compared with the input image
- An image decoding apparatus has a configuration for expanding and reproducing an image compression code of input information provided with a motion compensating means, and extracting a motion parameter in the input information to indicate a direction and an amount of motion.
- a reference image stored in correspondence with a frame by using an edge-to-peak decoding unit that obtains a motion vector and deformation pattern information representing the content of a deformation process instruction, and a motion parameter output from the entropy decoding unit.
- An image encoding method includes a motion compensation prediction unit that stores a reference image for compression encoding of an input digital image, divides the reference image into predetermined blocks, and detects motion between frames.
- a modified block matching step of transforming the integer pixels of the partial area into a predetermined format and specifying and extracting coordinates, generating a predicted partial image and comparing with the block of the input image; And a corresponding point determining step of determining a corresponding point of the partial area by specifying the coordinates from a motion vector giving the minimum error and obtaining a motion compensation output.
- the deformed block matching step includes, as a reference criterion, in addition to the integer pixels of the partial area of the reference image and a half pixel at the midpoint thereof, deforming the image into a predetermined format, specifying coordinates and extracting, and generating a predicted partial image The modified block matching step was compared.
- a deformation pattern table for deforming and adding a partial area of the reference image is provided.
- an image of the partial area based on the conversion value obtained by reading the corresponding address with reference to the deformation pattern table is input.
- a modified block matching step to be compared with the image was performed.
- An image decoding method includes a motion vector representing a direction and an amount of motion by extracting a motion parameter in input information in order to perform motion compensation and expand and reproduce an image compression code of input information;
- the entropy decoding step for obtaining the deformation pattern information representing the instruction content of the deformation processing, and the motion parameters obtained in the entropy decoding step are used to determine the partial area of the reference image stored corresponding to the frame.
- a motion compensation step is used to determine the partial area of the reference image stored corresponding to the frame.
- the coordinate calculation is performed using the coordinate value of the half pixel of the reference image, and the obtained pixel value is transformed into a predetermined format.
- Motion detection that includes a deformed block matching unit that deforms and specifies coordinates to extract and compares with the integer pixel of the block of the input image, and outputs a motion vector that gives the minimum error extracted and specified by coordinates.
- An image encoding device including a motion compensating unit for outputting a predicted partial image including the motion compensation; and a motion compensation predicting device for estimating and reproducing an image compression code of input information by including a motion compensation predicting device based on motion detection between frames.
- an image decoding device configured to add and output.
- FIG. 1 is a diagram showing a basic configuration of an image encoding device according to the present invention.
- FIG. 2 is an internal configuration diagram of the motion detection unit 8 in FIG.
- FIG. 3 is a flowchart showing the operation of the motion detector 8 having the configuration of FIG.
- FIG. 4 is a diagram illustrating an outline of an operation in the modified block matching unit 21 according to the first embodiment.
- FIG. 5 is an internal configuration diagram of the modified block matching unit 21.
- FIG. 6 is a flowchart showing the operation of the modified block matching unit 21.
- FIG. 7 is an internal configuration diagram of the motion compensation unit 9 in FIG.
- FIG. 8 is a flowchart illustrating the operation of the motion compensation unit 9.
- FIG. 9 is a diagram for explaining the operation of separating the image object by the preprocessing unit 2.
- FIG. 10 is another internal configuration diagram of the motion detection unit 8b according to the second embodiment.
- FIG. 11 is an internal configuration diagram of the motion detection unit 8 c according to the third embodiment.
- FIG. 12 is a diagram illustrating an outline of an operation of the modified block matching unit 42.
- FIG. 13 is an internal configuration diagram of the modified block matching unit 42.
- FIG. 14 is a flowchart illustrating the operation of the modified block matching unit 42.
- FIG. 15 is a diagram illustrating an outline of the operation of the modified block matching unit 42b in the fourth embodiment.
- FIG. 16 is an internal configuration diagram of the modified block matching unit 42b in the fourth embodiment.
- Figure 1 7 is a Furochiya one preparative diagram representing the operation of a modified block matching section 4 2 b.
- Figure 1 8 is a diagram for explaining another modification block matching in the fourth embodiment c
- FIG. 19 is a diagram for explaining another modified block matching according to the fourth embodiment.
- FIG. 20 is another internal configuration diagram of corresponding point determining section 34 in the fifth embodiment.
- FIG. 21 is a diagram for explaining modified block matching according to the sixth embodiment.
- FIG. 22 is a diagram illustrating an example of filtering performed on integer pixels forming a predicted image according to the sixth embodiment.
- FIG. 23 is a diagram illustrating an outline of the operation of the modified block matching unit 42c.
- FIG. 24 is an internal configuration diagram of the modified block matching unit 42c.
- FIG. 25 is a flowchart showing the operation of the modified block matching unit 42c.
- FIG. 26 is an internal configuration diagram of the motion compensation unit 9 b according to the sixth embodiment.
- FIG. 27 is a flowchart illustrating the operation of the motion compensation unit 9 b according to the sixth embodiment.
- FIG. 28 is a diagram illustrating a configuration of an image decoding device according to Embodiment 7.
- FIG. 29 is an internal configuration diagram of the motion compensation unit 9 according to the seventh embodiment.
- FIG. 30 is a flowchart showing the operation of the motion compensator 9 in FIG.
- FIG. 31 is a diagram illustrating the coordinate point moving operation performed by the motion compensating unit 9 in FIG. 29.
- FIG. 32 is a diagram for explaining an example of the deformation processing performed by the motion compensation unit 9 in FIG.
- FIG. 33 is a diagram showing an operation for obtaining a half pixel in the coordinate point operation.
- FIG. 34 is a diagram for explaining the operation in the case where the transformation process is rotation enlargement.
- FIG. 35 is a diagram illustrating a configuration of an image decoding apparatus according to Embodiment 8.
- FIG. 36 is an internal configuration diagram of the motion compensation unit 90 according to Embodiment 8.
- FIG. 37 is a flowchart showing the operation of the motion compensation unit 90 of FIG.
- FIG. 38 is a view for explaining an example of the deformation processing performed by the motion compensation unit 90 of FIG.
- FIG. 39 is a diagram illustrating an example of coordinate point calculation performed by the motion compensation unit 90 of FIG.
- FIG. 40 is a flowchart showing the operation of the corresponding point determining unit 37 c in the motion compensating unit according to the ninth embodiment.
- FIG. 41 is a diagram illustrating an example of a deformation process performed by the motion compensation unit according to the ninth embodiment.
- FIG. 42 is a diagram for explaining the concept of motion compensation prediction using block matching in Conventional Example 1.
- FIG. 43 is a diagram illustrating a configuration of a motion compensation prediction unit (block matching unit) of the image encoding device of Conventional Example 1.
- FIG. 44 is a flowchart illustrating the operation of the motion compensation prediction unit of the first conventional example.
- FIG. 45 is a diagram showing an outline of a motion compensation prediction method adopted in the MPEG1 video standard.
- FIG. 46 is a diagram for explaining the concept of motion compensation prediction using the affine transformation of the second conventional example.
- Figure 47 shows the configuration of the motion compensation prediction unit using the affine transformation of Conventional Example 2.
- FIG. 48 is a flowchart showing the operation of the motion compensation prediction unit in Conventional Example 2.
- FIG. 49 is a flowchart showing details of the affine motion parameter detection step in FIG. 48.
- FIG. 50 is a diagram for explaining a method of calculating a half-pixel value in the half-pixel generation unit 232.
- FIG. 51 is a diagram for explaining a method of calculating a non-integer pixel value generated in a rotation Z-scale amount search step in the interpolation calculation unit 231.
- the encoding device and the decoding device according to the present invention are specifically used for a digital image transmission system, a digital image recording device, a digital image storage database and a search / browsing system, etc., via a satellite, a terrestrial wave, or a priority communication network. used.
- FIG. 1 is a basic configuration diagram of an image encoding device.
- 1 is an input digital image signal
- 2 is a preprocessing unit
- 3 and 13 are intra (in a frame) Z-inter (interframe) coding selection unit
- 4 is an orthogonal transformation unit
- 5 is a quantization unit
- 6 is an inverse quantization unit
- 7 is an inverse orthogonal transform unit
- 8 is a motion detection unit
- 9 is a motion compensation unit
- 10 is a frame memory (reference image)
- 11 is a motion parameter including a motion vector
- 1 2 is prediction image data
- 14 is an encoding control unit
- 15 is a forced mode instruction flag
- 16 is an intra Z inter encoding instruction flag
- 17 is a quantization step parameter
- 18 is an entropy signal.
- the encoding unit 19 is compressed image data.
- This apparatus receives an image signal 1 of each frame as a component of a color moving image sequence as an input, and the input image signal 1 is digitized.
- the preprocessing unit 2 performs preprocessing, format conversion, and cutout into block data.
- the block data cut out here is composed of a pair of a luminance signal component and a color difference signal component spatially corresponding to the luminance signal component.
- the luminance component is referred to as a luminance block
- the color difference component is referred to as a chrominance block. Call.
- the intra / inter coding selection unit 3 determines whether each block data is to be intra-frame encoded or inter-frame encoded.
- intra (intra-frame) encoding block data composed of original image data output from the preprocessing unit 2 is input to the orthogonal transform unit 4, and inter (inter-frame) encoding is performed.
- the prediction error block data composed of the difference between the original image data output from the preprocessing unit 2 and the prediction image data 12 output from the motion compensation unit 9 is sent to the orthogonal transformation unit 4. input.
- the selection of the intra z-interframe coding may be forcibly performed by the forced mode instruction flag 15 from the coding control unit 14.
- the selected coding mode is sent to the intra-peak coding unit 18 as an intra Z-inter coding instruction flag 16 and multiplexed into the coding bit stream 19.
- the orthogonal transform unit 4 for example, a discrete cosine transform (DCT) is used.
- the orthogonal transform coefficient is quantized by the quantization unit 5 using the quantization step parameter 17 calculated by the coding control unit 14, and the quantized orthogonal transform coefficient is calculated by the entropy encoding unit 18. After redundancy reduction, they are multiplexed into the encoded bitstream 19.
- inverse quantization is performed by the inverse quantization unit 6, and then inverse orthogonal transformation is performed by the inverse orthogonal transformation unit 7, and the prediction error signal is restored. Will be replaced.
- the prediction image data 12 output from the motion compensation unit 9 is added to this to generate a locally decoded image.
- the 0 signal is selected by the coding selection unit 13 and the addition of the prediction error signal is not performed. Since the locally decoded image is used as a reference image for motion compensation prediction for the next and subsequent frames, the content is written to the frame memory 10.
- a block extracted in the preprocessing unit 2 is a predicted block in the motion compensation prediction.
- the motion compensation prediction process is performed in the motion detection unit 8 and the motion compensation unit 9.
- the motion detection unit 8 detects the motion parameter 11 including the motion vector of the predicted block, and the motion compensation unit 9 sets the motion parameter.
- the predicted image data 12 is extracted from the frame memory 10 using 11.
- the motion detection process is performed using the luminance block, and the motion compensation prediction of the chrominance block uses the motion detection result of the luminance block. In the following, the description will be limited to the motion compensation prediction operation of the luminance block.
- the motion detection processing is performed by the motion detection unit 8.
- the motion detection unit 8 searches for a region that is most similar to the luminance block of the predicted block within a predetermined range in the reference image, and detects a parameter representing a change in the predicted block from the position on the screen. .
- a block that is most similar to the luminance block of the block to be predicted is searched for, and a translation amount from a position in the screen of the block to be predicted is detected as a motion vector.
- the motion detection unit 8 of the present embodiment includes both a block matching based on a conventional square block and a block matching using a deformed block described later. And select the one with higher prediction accuracy.
- FIG. 2 is a detailed configuration diagram of the motion detecting unit 8 in FIG. 1, and FIG. 3 is a flowchart showing an operation thereof.
- reference numeral 20 denotes a block matching unit
- 21 denotes a modified block matching unit
- 22 denotes a motion compensation prediction mode determination unit
- 23 denotes a motion vector obtained by the modified block matching
- 24 denotes a motion vector obtained by the modified block matching.
- the small prediction error value, 25 is the final motion vector
- 26 is the motion compensation prediction mode signal. It is assumed that the motion parameter 11 is a combination of the final motion vector 25 and the motion compensation prediction mode signal 26.
- D—B M represents the value of the minimum prediction error power obtained by block matching
- D—DEF represents the value of the minimum prediction error power obtained by modified block matching.
- FIG. 4 is a schematic explanatory diagram of the operation in the modified block matching unit 21 which is the most important part of the present invention
- FIG. 5 is a detailed internal configuration diagram of the modified block matching unit 21, and
- FIG. 5 is a flowchart showing the operation of the matching unit 21.
- reference numeral 29 denotes a horizontal parallel movement amount search range instruction signal
- 30 denotes a vertical parallel movement amount search range instruction signal
- 31 denotes a horizontal movement amount counter
- 32 denotes a vertical movement amount counter
- 33 denotes Is a rotation amount counter which is a new element
- 34 is a corresponding point determination unit which is also a new element
- 35 is a memory read address generation unit.
- the pattern matching section 2 13 and the minimum prediction error power determination section 2 16 perform the same operations as the corresponding elements of the configuration shown in FIG.
- dx is the horizontal translation amount search point
- dy is the vertical translation amount search point
- r ange e_h_min is the lower limit of the horizontal search range
- r a n g e— V— ma a X is the upper limit of the vertical search range
- D_min is the minimum prediction error power
- D (d X, d y) is the prediction error power when searching d x, d y
- (x, y) is the pixel position in the predicted block
- D (d x, d y) is the prediction error at (x, y) in d x, d y search
- f r (x, y) is the value of pixel (x, y) in the reference image
- MV—h is the horizontal component of the motion vector
- MV—V is the vertical component of the motion vector
- the block matching unit 20 obtains a motion vector for the predicted block according to the procedure and operation described in the conventional example. As a result, the motion vector 2 17 and the minimum prediction error power D—BM 2 18 in the block matching unit 20 are obtained. This corresponds to S 1 in FIG.
- the deformed block matching unit 21 performs a deformed block matching process (S2 in FIG. 3).
- FIG. 4 shows an outline of the processing in the modified block matching unit 21.
- the image to be predicted 27 is encoded by motion compensated prediction.
- a frame (picture) in the pre-processing unit 2 and a reference image 28 are local decoded frame (picture) images that are encoded before the predicted image 27 and stored in the frame memory 10.
- ⁇ in each image indicates an integer pixel which is a real sampling point of a luminance signal actually existing in the frame
- X indicates a half pixel which is a middle pixel between the real sampling points.
- the partial region consisting of 8 is defined as the block to be predicted (the luminance block portion thereof), and the group consisting of the pixels of the mouth of the reference image 28 constitutes the transformed block of the predicted image candidate. That is, a part of the output of the frame memory 10 indicated by () in FIG. 1 and FIG. 2 and a part of the output of the preprocessing unit 2 are cut out and compared by the modified block matching unit 21 in the motion detection unit 8. .
- the brightness block of the reference image is rotated to the right or left 45 degrees, and the scale of each side is doubled.
- the size of the reference image is compared with the predicted image (input image).
- the area that matches the horizontal and vertical sample point distances of the frame's input digital image 1 as a distance of 1 ⁇ 2 is defined as a deformation block, and this area is the integer pixel spacing of the reference image 28
- the modified block matching in the present embodiment is characterized in that the luminance block of the predicted block composed of 8 ⁇ 8 integer pixels shown in FIG. In the up Also corresponds to the process of finding a similar deformed block area in the reference image 28 in FIG.
- each constituent point of the luminance block of the predicted block and each constituent point of the deformed block of the predicted image candidate area are associated one-to-one in advance.
- the predicted image candidate image is a partial image obtained by rotating the deformed block in the reference image 28 by 45 degrees to the right and correcting each side to a half length. Changing this association will change the direction of rotation.
- each of the other constituent points can be uniquely associated. Since each constituent point of the predicted block and each constituent point of the predicted image candidate are associated one-to-one, motion detection can be performed in the same manner as block matching.
- the extraction type of the partial area for comparison of the reference image 28 in FIG. 4 is patterned and specified by addressing (coordinates), and in this case, an instruction is made so that an integer pixel is selected.
- the minimum error power is determined by accumulating the error between the pixel designated by the addressing and the pixel in the original image data corresponding to the predicted image 27. Therefore, high-speed comparisons can be made because there is no operation just by instructing the addressing, and in addition to simple enlargement / reduction by the method of addressing (coordinate designation), rotation as well as rotation and enlargement / reduction can be performed.
- the horizontal movement amount counter 31 and the vertical movement amount counter 32 are transformed from the horizontal parallel movement amount search range instruction signal 29 and the vertical direction parallel movement amount search range instruction signal 30.
- the minimum prediction error power determination unit 211 the minimum prediction error power D—min is set to the maximum value MAXINT (for example, OxFFFFFFFF). This corresponds to S4 in FIG.
- r dx and r dy shown in S6 and S8 in FIG. 6 are used as block deformation parameters.
- the rotation amount counter 33 sets this parameter. In other words, the relationship of rotating the reference image of FIG. 4 clockwise by 45 degrees is defined.
- the values of y are given as initial values, and thereafter, every time X is incremented, rd x is incremented and rd y is decremented.
- the corresponding points rx, ry in the reference image corresponding to the position (x, y) of the predicted block in the luminance block are determined. That is, the correspondence between the first positions in FIG. 4 is performed. This is performed by the corresponding point determination unit 34.
- rx, ry can be obtained by adding rdx, rdy obtained in 2-3) to offset values ix, iy given in advance.
- the pixel in the reference image located at a position (rx + dx, ry + dy) away from the reference image is extracted from the frame memory. In FIG.
- the memory read address generation unit 35 calculates the value of dX from the horizontal movement counter 31, the value of the vertical movement force counter 32, and the value of dy, and the corresponding point determination unit 34 determines rx, ry. And generates an address in the frame memory.
- the prediction error power D (dX, dy) when the motion vector is (dx, dy) is initialized to zero. This corresponds to S5 in FIG.
- the difference between the pixel value read in 2-4) and the pixel value at the corresponding position in the luminance block of the predicted block is calculated, and the absolute value is accumulated in D (dX, dy).
- This processing is performed by the pattern matching unit 213 in FIG. 5, and the pattern matching unit 213 outputs D (dx, dy) to the minimum prediction error power determination unit 216 based on the prediction error power signal 215. Hand over.
- the above processing corresponds to the processing of S9 to S14 in FIG.
- the prediction image most similar to the predicted block in the sense of the minimum error power is found.
- the amount of deviation from the starting point of the selected predicted image is obtained as a motion vector 23 as a result of the modified block matching, and the prediction error power D—DEF 24 at that time is also retained.
- the minimum prediction error power D—BM2 18 obtained by the block matching unit 20 and the minimum prediction error power D comfortablyDEF 24 obtained by the modified block matching unit 21 are calculated. Then, the smaller of block matching and modified block matching is selected as the final motion compensation mode, which corresponds to S3 in FIG.
- the motion compensation prediction mode determination unit 22 uses the finally selected motion compensation prediction mode signal 26 and the final motion vector 25 as the motion parameters 11 and performs the motion compensation unit 9 and the event P code. Sent to chemical department 18.
- the motion compensation processing is performed by the motion compensation unit 9.
- the motion compensation unit 9 extracts a predicted image from the reference image based on the motion parameters 11 obtained by the motion detection unit 8.
- the motion compensating unit 9 of the present embodiment performs both the motion compensation processing of the conventional block matching based on a square block and the motion compensation processing of the block matching using a specific deformed block, and calculates the motion in the motion parameter 11.
- a configuration is adopted in which these processes are switched according to the compensation prediction mode.
- FIG. 7 is a configuration diagram of the motion compensator 9 in FIG. 1, and FIG. 8 is a flowchart showing the operation thereof.
- reference numeral 37 denotes a corresponding element determination unit which is a new element
- reference numeral 38 denotes a memory read address generation unit.
- the predicted image in the reference image 28 is obtained from the position indication signal 206 of the block to be predicted in the screen and the motion parameter 11 sent from the motion detection unit 8. Is determined.
- This processing is performed in the corresponding point determination unit 37 in FIG.
- the motion compensation prediction mode included in the motion parameter 11 indicates block matching
- the corresponding point is translated by an amount indicated by the motion vector from the position signal 206 of the predicted block in the screen.
- the sampling points are included in the region.
- This processing corresponds to the operation of determining the position (x + dx, y + dy) in the reference image 28 when (dx, dy) is the motion vector in S204 in FIG. Hit.
- the motion compensation prediction mode included in the motion parameter 11 indicates the deformed block matching, as described in 2-4) in the description of the motion detection unit 8
- the in-screen position signal 2 of the block to be predicted is used.
- the sample point is moved in parallel by the amount indicated by the motion vector. This processing corresponds to the operation of determining the position (rX + dX, ry + dy) in the reference image 28 when (dx, dy) is the motion vector in S9 in FIG.
- the memory read address generation unit 38 receives the result of the corresponding point determination unit 34 and stores the reference image 28 stored in the frame memory 10. Generates a memory address that specifies the position of the predicted image in the image, and reads the predicted image.
- the half-pixel generation unit 232 generates a half-pixel value before being output from the motion compensation unit 9. This is processing corresponding to S23 and S24 in FIG. 8, and whether or not the predicted image includes pixels with half-pixel accuracy is determined by the corresponding point determination unit 37 by the motion vector in the motion parameter 11. Identify based on the torque value and inform selection switch 36.
- corresponding points of only actual sample points are generated as described in FIG.
- the configuration in the case where there is a half pixel is a modified block matching section 42 having a half pixel generation section 2 32 of FIG. 13 as described later.
- final predicted image data 12 is output.
- the example of the rotation as the modified block matching in the above embodiment is 45 degrees, the rotation is not limited to 90 degrees, 135 degrees, 180 degrees, etc. Rotation can also be realized.
- an image encoding apparatus using an image frame as a unit The preprocessing unit 2 performs processing to separate the input digital image sequence into image objects (partial regions having the same features such as motion and picture, one subject, etc.), and each image object is processed.
- the present invention can be applied to a device that encodes an image object as a unit. For example, as shown in Fig. 9, in a scene where a human image exists in front of a stationary background, the human image is used as an image object, and the area inside the circumscribed rectangle surrounding the image is a small block as shown in the figure. There is a case where a block containing an image object is divided and encoded as an effective block. In this case, the same processing is applied to these effective blocks as to the modified book matching and the motion compensation described in the above embodiment. This is the same in the following embodiments.
- an encoding apparatus based on orthogonal transform encoding has been described.
- the present invention can be applied to an apparatus that encodes a motion compensation prediction error signal using another encoding method. Needless to say. This is the same in the following embodiments.
- the rough movement amount of the partial area to be subjected to the deformed block matching process can be grasped.
- the setting destination of the partial area of the deformed block matching is received from the area information indicated by the motion vector 217 which is the search result of the block matching unit 20, the area is limited to this area, and the area is deformed and compared. Processing time can be reduced. In this embodiment, this configuration will be described. This is the same in other embodiments described below.
- This embodiment shows another embodiment of the motion detection unit 8.
- FIG. 10 is an internal configuration diagram of the motion detection unit 8b in the present embodiment. Is a deformed block matching section, 40 is an addition section, and 41 is a search initial position indication signal.
- the modified block matching section 39 uses only the search initial position indicating signal 41 instead of the input 206, and the other operations are exactly the same as those of the modified block matching section 21 in the first embodiment.
- FIG. 10 shows a specific circuit of a device for obtaining a rough value.
- the motion vector obtained as a result of the block matching unit 20 is added to the predicted block data 205 instead of the in-screen position signal 206 of the predicted block in the modified block matching unit 39.
- the adder 40 adds the torques 2 17, and inputs the result as the search initial position indication signal 41.
- the search range set from the horizontal direction parallel movement amount search range instruction signal 29 and the vertical direction parallel movement amount search range instruction signal 30 is set smaller than in the first embodiment. Thereby, iterative processing in S17 to S20 in FIG. 6 can be shortened.
- the deformed block area is constituted by only the pixel points at the integer pixel interval in the reference image 28 has been described.
- the deformed prog region is configured to include pixel points at half pixel intervals in the reference image 28.
- the internal configurations of the motion detecting unit 8 and the motion compensating unit 9 in FIG. 1 are different from those of the first embodiment. Further, the operation is different only in the deformed block matching unit in the motion detection unit and the corresponding point determination unit in the motion compensation unit, and the other members and operations are exactly the same as those in the first embodiment. Therefore, only the operation of the deformed block matching unit and the corresponding operation of the motion compensation unit will be described in detail below. As in the first embodiment, the operation will be described separately for the motion detecting unit 8c and the motion compensating unit 9.
- FIG. 11 is an internal configuration diagram of the motion detection unit 8c according to the present embodiment. Is a schematic explanatory diagram of the operation in the deformed block matching unit 42, which is one of the most important parts of the present invention, FIG. 13 is a detailed internal configuration diagram of the deformed block matching unit 42, and FIG. 6 is a flowchart showing the operation of a modified block matching unit 42.
- the outline of the processing in the deformed block matching unit 42 is shown in FIG.
- the predicted image 27 and the reference image 28 are as defined in the first embodiment.
- ⁇ in each image indicates the actual sample point (integer pixel) of the luminance signal of the frame
- X indicates the midpoint pixel (half pixel) between the actual sample points.
- the partial area consisting of 8 X 8 (integer pixels) of the predicted image 27 is set as the predicted block (the luminance block portion), and the group consisting of the pixels of the mouth of the reference image 28 constitutes the transformed block of the predicted image candidate. Shall be.
- the luminance block is rotated right or left 45 degrees and the scale of each side is multiplied by 1 to 2 times. That is, the size of the reference image is set to a distance of 2 and the horizontal of the input digital image 1 of the frame
- the area that matches the vertical sample point distance is defined as a deformed block.
- This region is characterized in that it also includes pixel points of the reference image 28 at half pixel intervals. That is, the modified block matching in the present embodiment is performed for a predicted block consisting of 8 ⁇ 8 samples (hereinafter, samples are meanings of integer pixels or half pixels) shown in FIG. 12 within a given search range. This corresponds to the process of finding, from the reference image 28, the deformed block area in the figure most similar to the luminance block.
- each component point of the luminance block of the predicted block and each component point of the deformation block of the predicted image candidate area are made to correspond one-to-one in advance.
- the pixel position of the upper left corner of the predicted block and the left vertex of the deformed block correspond in advance, as indicated by the dotted arrow in FIG.
- the left vertex on the deformed block side is located at a half pixel position, which shows a case where the motion vector includes a half pixel component.
- the predicted image candidate image is a partial image obtained by rotating the deformed block in the reference image 28 by 45 degrees to the right and correcting each side to be twice as long.
- each of the other constituent points can be uniquely addressed. Since each constituent point of the predicted block and each constituent point of the predicted image candidate are associated one-to-one, motion detection can be performed in the same manner as block matching.
- the operation of setting the search range in the actual device is the same as in the first embodiment, and is set using the necessary elements in FIGS. This operation corresponds to step S26 in FIG.
- r dX and r dy shown in FIG. 14 are used as block deformation parameters.
- This parameter is set by the rotation amount counter 45. Given the value of y as their initial value, each time X is incremented by one, the value of 1 ⁇ is incremented by 0.5 and the value of r d y is decremented by 0.5.
- the predicted block corresponds to the position (x, y) in the luminance block.
- rx, ry can be obtained by adding rdX, rdy obtained in 3) to offset values iX, iy given in advance.
- the memory read address generation unit 47 reads the value of dx from the horizontal movement counter 31, the value of dy from the vertical parallel movement counter 32, and reads rx and ry from the corresponding point determination unit 46. Receive and generate the address in the frame memory.
- the data read in S31 of FIG. 14 is used for generating a half-pixel value in the half-pixel generation unit 232 as necessary.
- the prediction error power D (dx, dy) when the motion vector is (dx, dy) is initialized to zero. This corresponds to S27 in FIG.
- the difference between the pixel value read in 4) and the pixel value at the corresponding position of the luminance block of the predicted block is calculated, and the absolute value is accumulated in D (dX, dy).
- This processing is performed by the pattern matching unit 2 13 in FIG. 13, and the pattern matching unit 2 13 converts D (d X, dy) into the minimum prediction error power determination unit 2 16 using the prediction error power signal 2 15.
- Hand over to The processing corresponds to the processing of S32 to S37 in FIG.
- the prediction image most similar to the predicted block in the sense of the minimum error power is found.
- the shift amount from the starting point of the selected predicted image is obtained as a motion vector 43 as a result of the deformed block matching, and the prediction error power D—DEF 44 at that time is also retained.
- the motion vector 43 and the prediction error power D-DEF44 are used for the final motion compensation mode determination, and the final motion compensation mode is determined. This determination method is exactly the same as in the first embodiment.
- the motion compensation processing is performed by the motion compensation unit 9.
- the operation of the corresponding point determination unit 37 is different from that of the first embodiment, and only that part will be described.
- the overall flowchart of the motion compensation follows Figure 8.
- the corresponding point is parallel by the amount indicated by the motion vector from the position signal 206 in the screen of the predicted block. Included in the moved area It is a sampling point. This processing corresponds to the operation of determining the position (x + dx, y + dy) in the reference image 28 when (dx, dy) is the motion vector in S204 in FIG. .
- the motion compensation prediction mode included in the motion parameter 11 indicates the deformed block matching, as described in 4) in the description of the motion detection unit 8
- the position signal in the screen of the predicted block is used. After adding the amount of rotation corresponding to each pixel position to 206, it becomes a sample point included in the area translated by the amount specified by the motion vector.
- This processing corresponds to the operation of determining the position (rx + dx, ry + dy) in the reference image 28 when (dX, dy) is the motion vector in S32 in FIG. .
- the modified block matching unit 42b is assumed to be a variation of the modified block matching unit 42 in FIG. 13 and the input is exactly the same and the output is the motion vector. It is assumed that this is a variation of 4 3 and prediction error power 4 4.
- the corresponding point determining section in the motion compensating section 9 is also a variation of the corresponding point determining section 37 in FIG. Therefore, in the following, the number of the modified block matching unit of the present embodiment will be described. Is described as 4 2 b, and the corresponding point determination unit number is 37.
- FIG. 15 is a schematic explanatory diagram of the operation in the modified block matching unit 42b in the present embodiment
- FIG. 16 is a detailed internal configuration diagram of the modified block matching unit 42b
- FIG. 9 is a flowchart showing the operation of the block matching unit 42b.
- Figure 15 shows an overview of the processing in the modified block matching unit 4 2 b.
- the predicted image 27 and the reference image 28, and the description of the marks in each image are as described above.
- a reduced area in which each side of a luminance block is simply multiplied by 1 Z 2 is defined as a deformation block.
- the deformed block matching in the present embodiment uses the deformed block area of FIG. 15 most similar to the luminance block of the predicted block composed of 8 ⁇ 8 samples shown in FIG. 15 within the given search range. This corresponds to the process of finding from inside.
- each component point of the luminance block of the predicted block and each component point of the deformation block of the predicted image candidate area are made to correspond one-to-one in advance.
- the pixel position at the upper left corner of the predicted block and the pixel position at the upper left corner of the deformed block are associated in advance. Since each constituent point of the predicted block and each constituent point of the predicted image candidate are associated with each other on a one-to-one basis, motion detection can be executed in the same manner as block matching.
- the operation of setting the search range in the actual device is the same as in the first embodiment, and is performed using the necessary elements in FIG. This corresponds to the step S44 in Figure 17. 3) Read predicted image candidate image
- the offset values iX, 1y of the horizontal and vertical components are the values of xZ2, y / 2.
- the corresponding points sx and sy of x and y This corresponding point is performed by the corresponding point determining unit 48.
- the pixel in the reference image located at a position (sx + dx, sy + dy) away from the reference image is extracted from the frame memory.
- the memory read address generation unit 49 obtains the value of dX from the horizontal movement counter 31 and the value of dy from the vertical parallel movement counter 32, and outputs the value of dy from the corresponding point determination unit 48.
- the data read in S48 of FIG. 17 is used for generating a half-pixel value in the half-pixel generation unit 232 as necessary.
- the prediction error power D (dX, dy) when the motion vector is (dx, dy) is initialized to zero. This corresponds to S45 in FIG.
- the difference between the pixel value read in 3) and the pixel value at the corresponding position of the luminance block of the predicted block is calculated, and the absolute value is accumulated in D (dX, dy) by S50. .
- This processing is performed by the pattern matching unit 2 13 in FIG. 16, and the pattern matching unit 2 13 converts D (dX, dy) into the minimum prediction error power determination unit 21 by using the prediction error power signal 2 15. Hand over to 6.
- the processing here corresponds to the processing of S49 to S54 in FIG.
- Update minimum prediction error power value 4) D (dx, dy) i obtained as a result of the search is used to determine whether the minimum error power is given. The determination is performed by the minimum prediction error power determination unit 216 in FIG. S55 in FIG. 17 corresponds to this determination processing. The determination process is exactly the same as in the first embodiment, and the value of (dx, dy) at that time is stored as a motion vector candidate. This update processing corresponds to S56 in FIG.
- the prediction image most similar to the predicted block in the sense of the minimum error power is found.
- the amount of shift from the starting point of the selected predicted image is obtained as a motion vector 43 as a result of the deformed block matching, and the prediction error power D—DEF 44 at that time is also retained.
- the motion vector 43 and the prediction error power D-DEF44 are used for final motion compensation mode determination, and the final motion compensation mode is determined. This determination method is exactly the same as in the first embodiment.
- the motion compensation processing is performed by the motion compensation unit 9.
- the operation of the corresponding point determination unit 37 is different from that of the first embodiment, and only that part will be described.
- the overall flowchart of the motion compensation follows Figure 8.
- the corresponding point is the amount indicated by the motion vector from the position signal 206 in the screen of the predicted block. Included in the area translated only It is a sampling point.
- This processing corresponds to the operation of determining the position (X + dX, y + dy) in the reference image 28 when (dx, dy) is the motion vector in S204 in FIG. Equivalent to.
- the motion compensation prediction mode included in the motion parameter 11 indicates the deformation block matching
- the motion This is the sampling point included in the area translated by the amount specified by the vector.
- This processing corresponds to the operation of determining the position (sx + dx, sy + dy) in the reference image 28 when (dx, dy) is the motion vector in S47 in FIG. .
- the following readout of predicted image data and generation of a predicted image are according to the first embodiment.
- the deformed blocks in each of the above embodiments are: 1) one-to-one correspondence between the predicted pixel and each constituent pixel position of the predicted image; 2) the corresponding pixel point on the reference image side is an integer pixel interval.
- Any shape can be taken if it is based on the two premise that it is composed of For example, shapes as shown in FIGS. 18 and 19 can be considered.
- not only one side can be reduced to half, but if each side is independently reduced and expanded at an arbitrary ratio, it can be transformed into various shapes to perform block matching. In this way, by defining various shapes in advance, it is possible to configure so as to select a deformed block that gives the best prediction result. In this case, the type of the selected deformed block is included in the motion parameter 11 and sent to the entropy encoder 18.
- the fixed point prepared in advance is an integer pixel or a half pixel has been described.
- a pixel at an intermediate point of another ratio such as 1: 3 is used for comparison. You may prepare as.
- the interpolation processing during the comparison processing operation is unnecessary, and the processing scale can be reduced accordingly, and high-speed processing can be performed.
- the block transformation parameter is counted for each pixel, or the coordinate transformation processing corresponding thereto is performed.
- the coordinate transformation value for each pixel is stored in advance in a transformation pattern table such as ROM. It is also possible to adopt a configuration in which a corresponding point is determined based on a conversion value extracted from a table according to each pixel position of a predicted block. In this way, deformation block matching and motion compensation having an arbitrary correspondence that is difficult to express in an arithmetic expression can be effectively performed.
- Embodiment 1 is taken as an example.
- FIG. 20 is another internal configuration diagram of the corresponding point determining unit 34 in FIG. 5, and shows a configuration (corresponding point determining unit 34 b) for realizing the present embodiment.
- the values of rdx and rdy corresponding to x and y are stored as ROM. It can be obtained by extracting the corresponding points rX and ry according to the values of x and y.
- the rotation amount counter 33 in FIG. 5 becomes unnecessary, and as shown in FIG. 20, a configuration in which a ROM table (deformation pattern table 100) is provided in the corresponding point determination unit 34b. Can be realized.
- the corresponding point determination unit 34b changes according to each pixel position (x, y) of the predicted block. 7 25
- the corresponding points are determined by extracting the transformation parameters r d X and r d y from the 48-shaped pattern table 100 and adding them in the adder 110. Then, the data is output to the memory read address generation unit 35.
- This is the same in the other embodiments described above. In this way, by simply adding a small amount to the ROM memory (deformation pattern table 100), the elements for performing the corresponding point arithmetic processing are deleted, thereby simplifying the circuit and reducing the corresponding point arithmetic processing amount. be able to.
- an encoding apparatus that makes frequency characteristics in a predicted image cut out as a deformed block uniform by the method described in each of the above embodiments and reduces mismatches when predicting a predicted block is performed. A description is given below.
- the spatial frequency characteristics are different between the integer pixel and the half pixel.
- the blocks to be predicted are all composed of pixels in the integer pixel space, it is conceivable that this difference in characteristics may cause a mismatch at the time of prediction. Therefore, in the present embodiment, after performing the same definition as the definition of the deformed block described in each of the above embodiments, filtering is performed on pixels in the integer pixel space.
- Pixels in the half-pixel space are generated by performing [12, 1/2] filtering on the surrounding integer pixels. That is, a low-pass filter having a characteristic of cos (cot / 2) is applied.
- the predicted image defined in each of the above embodiments is a mixture of unfiltered integer pixels and half-pixel precision pixels generated by the filtering. And the spatial frequency characteristics in the predicted image vary. If the prediction accuracy drops due to this variation, it is effective to apply a filter with the same characteristics to integer pixels as described below.
- Figure 22 shows an example of filtering.
- an example is shown in which an integer pixel is subjected to the mouth-pass filter F of [1Z8, 68, 1/8] shown in equation (7). ing.
- the present embodiment differs from the previous embodiments in the deformed block matching section and the motion compensation section.
- the modified block is defined as a simple reduced pattern based on Embodiment 4, and the internal configuration of the modified block matching unit is the variation block and the motion compensation of the modified block matching unit 42 in the motion detection unit 8c. Section is also considered as a variation of the motion compensation section 9. Therefore, in the following explanation, the number of the deformed block matching is 4
- FIG. 23 is a schematic explanatory diagram of the operation of the modified block matching unit 42 c in the present embodiment
- FIG. 24 is a detailed internal configuration diagram of the modified block matching unit 42 c
- FIG. 26 is a flowchart showing the operation of the modified block matching unit 42c in the embodiment.
- the definition of the deformed block is exactly the same as in the fourth embodiment, but the present embodiment is different from the first embodiment in that filtering is performed on pixels at integer pixel positions. That is, as shown in FIG. 23, a pixel as a pixel to be subjected to the filtering process is defined in the reference image, and the deformed block includes pixels indicated by ⁇ and a mouth.
- the method of obtaining the corresponding points s x and s y corresponding to the pixels at the positions x and y in the predicted block is exactly the same as in the fourth embodiment.
- a pixel in the reference image located at a position (sX + dX, sy + dy) away from the reference image is extracted from the frame memory.
- it is determined whether the pixel position corresponding to sx + dx, sy + dy is on the integer pixel space or the half-pixel space. This can be determined simply by checking whether s x + d x and s y + d y each have a half-pixel component. This determination is performed in the corresponding point determination unit 48 in FIG. In FIG. In FIG.
- the motion compensation processing is performed by the motion compensation unit 9b.
- FIG. 26 is an internal configuration diagram of the motion compensation unit 9b in the present embodiment
- FIG. 27 is a flowchart showing the operation of the motion compensation unit 9b in the present embodiment.
- the present embodiment is characterized in that a filter unit 50 is added as compared with the motion compensation unit 9 shown in FIG.
- the corresponding point determination unit 37 operates exactly the same as that described in the fourth embodiment.
- the motion compensation prediction mode included in the motion parameter 11 indicates block matching
- the corresponding point is parallel by the amount indicated by the motion vector from the position signal 206 in the screen of the predicted block.
- the sample point is included in the moved area.
- This processing corresponds to the operation of determining the position (x + dx, y + dy) in the reference image 28 when (dx, dy) is the motion vector in S204 in FIG. .
- the following prediction image data The reading of data and the generation of a predicted image are in accordance with the first embodiment.
- the modified block matching unit 42c in the present embodiment performs a search independently for each of the predicted image when the filter is not applied and the predicted image when the filter F is applied, and performs motion compensation on the results.
- the data may be sent to the prediction mode determination unit 22 or a search may be performed only when the filter F is not applied, and a good result may be selected by applying the filter F to only the result.
- the filtering to the integer pixel value is performed. Alone can eliminate the variation in spatial frequency in the predicted image, and the motion vector, which is the amount of translation, cannot minimize the prediction error alone.In other words, even for partial images for which prediction is not successful, Good predictions can be made.
- FIG. 28 shows the configuration of an image decoding apparatus that expands and reproduces a digital image that has been compression-encoded using the image prediction method according to the present embodiment.
- an image decoding apparatus that receives compressed and encoded data (hereinafter, a bitstream) 19 generated by the image encoding apparatus according to Embodiment 1 and performs extension reproduction will be described.
- 51 is an entropy decoding unit
- 6 is an inverse quantization unit
- 7 is an inverse orthogonal transform unit
- 53 is a decoding and adding unit
- 54 is a frame memory
- 56 is a display control unit.
- the decoding device of the present invention is characterized by the configuration and operation of the motion compensation unit 9, and the configuration and operation of each of the above-described elements other than the motion compensation unit 9 are already known, so detailed description will be omitted.
- the motion compensation unit 9 Indicates that it is the same as Part 9. That is, the internal configuration diagram is the same as the internal configuration diagram shown in FIG. 7, and the operation flowchart is the same as the operation flow chart shown in FIG.
- bit stream is analyzed in the entropy decoding unit 51.
- the quantized orthogonal transform coefficient 5 2 is sent to the inverse quantizer 6 and inverse quantized using the inverse quantization step 'parameter 17-the result is inverse orthogonal transformed in the inverse orthogonal transformer 7. , And are sent to the decoding addition section 53.
- the inverse orthogonal transform unit the same one used in the encoding device, such as DCT, is used.
- the following three types of information are sent to the motion compensation unit 9 as the motion parameters 11. That is, the motion vector 25 and the deformation pattern information 26a decoded from the bitstream by the entropy decoding unit 51 and the predicted image area (in this embodiment, fixed-size blocks) Information 27a indicating the position in the screen of is input.
- the motion vector 25 and the position 27 a in the screen of the predicted image area are values unique to each predicted image area, while the deformation pattern information 26 a is Even if the value is specific to the image, it is encoded for each larger image (for example, an image frame or the VOP disclosed in ISOZIECJTC1 / SC29 / WG11) that combines multiple predicted image areas, All the predicted image areas included in the unit may be encoded so as to use the same deformation pattern information.
- the motion compensator 9 extracts the predicted image data 12 from the reference image in the frame memory 54 according to these three types of information. The process of generating a predicted image will be described later in the description of the operation of the motion compensation unit 9.
- the motion compensation unit 9 includes the motion parameters decoded by the entropy decoding unit 51.
- the motion compensator 9 extracts predicted image data 12 from the reference image in the frame memory 54 according to the motion parameters 11.
- the prediction is performed using the motion parameter 11 as in the conventional motion compensation in block matching.
- the image area is uniquely determined.
- the decoding addition unit 53 Based on the value of the intra / inter encoding instruction flag 16, the decoding addition unit 53 outputs the output of the inverse orthogonal transform unit as it is as the decoded image 55 if it is an intra encoded block, and In the case of a coded block, predicted image data 12 is added to the output of the inverse orthogonal transform unit, and the resultant is output as a decoded image 55.
- the decoded image 55 is sent to the display control unit 56, output to a display device (not shown), and written to the frame memory 54 for use as a reference image in the decoding processing of subsequent frames.
- the motion prediction method A predicted image is generated by simple address calculation and interpolation based on the position correction based on the displacement and deformation pattern information 26a due to the torque 25.
- FIG. 29 shows the internal configuration of the motion compensation unit 9.
- reference numeral 37 denotes a corresponding point determination unit
- reference numeral 38 denotes a memory read address generation unit
- FIG. 30 is a flowchart showing the operation.
- Fig. 31 is a diagram illustrating the movement of a specified amount cut out from the reference image by the motion vector and moving to the coordinate position of the predicted image.
- Fig. 32 is a diagram illustrating the movement further specified at the movement destination. Perform dressing with deformation pattern It is a figure explaining an operation.
- ⁇ indicates the position of an integer pixel
- X indicates the position of a half pixel
- the corresponding point determination unit 37 calculates the sample position of the predicted image corresponding to each pixel in the increased area based on the input motion vector 25 and the deformed pattern information 26a.
- FIG. 32 shows an example in which the deformed pattern information 26 a indicates “vertical and horizontal 1/2 reduction”.
- the execution area of the predicted image is one-fourth of the execution area occupied in the screen of the predicted image area.
- the predicted image is reduced in size relative to the predicted image area, and this makes it possible to more efficiently predict a motion involving enlargement on the screen.
- a correction position (i ", J") corresponding to the position (i ', j') in the reference image is obtained. This can be achieved by the following operation (S72 in Fig. 30).
- the coordinate point (i ", j") obtained as described above is output as the predicted image sample position corresponding to the force (i, j).
- the memory read address generation unit 38 Based on the predicted image sample position output from the corresponding point determination unit 37, the memory read address generation unit 38 outputs the image data necessary for generating the predicted image in the reference image stored in the frame memory 54. Generates a memory address that identifies the position and reads the data for predictive image generation.
- the half-pixel generation unit 232 When only coordinate values of integer pixel positions among pixels for generating a predicted image are to be addressd, the data for generating a predicted image is directly used as a predicted image constituent pixel.
- the half-pixel generation unit 232 performs an interpolation process on the data for predictive image generation to generate a half-pixel value. Specifically, the generation of the half pixel value is shown in FIG.
- the method of Fig. 33 is simply a binary operation of addition, and FIG. 8 is a flow chart of the half-pixel generation unit 2 32 described in the state 1, and S 24 of FIG. 8 is described again.
- FIG. 34 The case of FIG. 34 will be described as an example of another deformation pattern.
- (i ", ⁇ ') after the deformation is obtained as follows.
- a deformed pattern is prepared in advance, and simple sample position calculation is performed according to the corresponding mode information. It is possible to obtain a reproduced image from a coded bitstream by efficiently predicting complex movements.
- a coding scheme other than orthogonal transform coding is used. Even in the case of a bit stream obtained by encoding a measurement error signal, the same effect can be obtained by changing elements other than the motion compensation unit 9 for decoding a prediction error signal.
- a decoding device that uses a unit of an arbitrary shaped image object (eg, Video Object P 1 ane disclosed in ISOZIECJTCIZSC29ZWG11 / N1796). Is also applicable.
- an arbitrary shaped image object eg, Video Object P 1 ane disclosed in ISOZIECJTCIZSC29ZWG11 / N1796).
- FIG. 9 described in the first embodiment in a scene where a human image exists in front of a stationary background, the human image is taken as one image object, and the inside of a circumscribed rectangle surrounding it is It is conceivable to divide the region into small blocks and decode the encoded bitstream with the block containing the image object as the effective block. In this case, the same processing may be applied to these valid blocks.
- the image decoding apparatus performs a predetermined deformation process only by performing addressing (coordinate designation) using only integer pixels or half pixels corresponding to the image decoding apparatuses according to the first to sixth embodiments.
- a device for motion compensation has been described.
- an image decoding apparatus that performs operations other than half-pixel generation at the time of addressing and performs more precise motion compensation will be described.
- FIG. 35 shows the configuration of an image decoding apparatus according to the present embodiment that expands and reproduces a compression-encoded digital image.
- 90 is a motion compensation unit
- 25b is 0 to 4 motion vectors
- 60 is interpolation processing accuracy instruction information.
- FIG. 36 is a partial configuration diagram of the motion compensation unit 90.
- 37 b is a motion parameter
- the motion vector 25 b shown in FIG. 35 the deformation pattern information 26 a
- the in-screen position 27 a of the predicted image area and the interpolation processing accuracy
- a corresponding point determination unit that determines a corresponding point by using the instruction information 60 as an input
- 2332b is an interpolation processing unit that obtains a coordinate position interpolated by calculation.
- the in-screen position 27a of the predicted image area is a value unique to each predicted image area, but the motion vector 25b and the deformation pattern information 26a are calculated based on the predicted image area.
- FIG. 37 is a flowchart showing the operation of the motion compensation unit in FIG. 36
- FIG. 38 is a diagram for explaining the operation of the same.
- the coordinate position is obtained by a calculation described later in the operation of determining a corresponding point. Further, the determined coordinate position is rounded by the interpolation processing instruction information to determine the coordinate position.
- the operation of the parts other than the motion compensation unit 90 is the same as that of the device of the seventh embodiment. That is, in the entropy decoding unit 51, the bitstream is analyzed and divided into individual encoded data.
- the quantized orthogonal transformation coefficient 52 is decoded by the inverse quantization unit 6 and the inverse orthogonal transformation unit 7 using the quantization step 'parameter 17, and sent to the decoding addition unit 53.
- the decoding addition unit 53 performs the intra-coding based on the value of the intra-inter-coding instruction flag 16.
- the predicted image data 12 is output as it is or as a decoded image 55 after adding or adding the predicted image data 12.
- the decoded image 55 is sent to the display control unit 56, output to the display device, and written into the frame memory 54 as a reference image.
- the predicted image generation processing in the motion compensation unit 90 will be described.
- a conversion equation necessary for deformation is obtained using the required number of motion vectors 25b, and each pixel in the predicted image area is obtained by the conversion equation.
- the predicted image is generated by a simple interpolation process according to the pixel system defined by the interpolation process accuracy indication information.
- the corresponding point determination unit 37b based on the input motion vector 25b and the deformation pattern information 26a, the coordinate position to be sampled of the predicted image corresponding to each pixel in the predicted image area is determined. calculate. As shown in FIG. 38, here, the motion vector 25 b is four motion vectors at each vertex of the circumscribed rectangle of the predicted image area. First, a conversion formula necessary for deformation is obtained corresponding to the deformation pattern information 26a. For example, the following conversion formula is used.
- the format of the deformation pattern information 26a may be a bit that directly identifies the above-mentioned conversion formulas (9) to (13), and each conversion corresponds to the number of motion vectors. Therefore, it may be a bit that represents the number of motion vectors.
- the point (i, j) in the predicted image area is associated with (i ', j') in the reference image.
- the sample position of the predicted image can take a value up to the accuracy determined by the interpolation processing accuracy instruction information 60. For example, if it is rounded to half-pixel accuracy, (i ', j') obtained by the above conversion formula is rounded to a value of half-pixel accuracy. With up to 14 pixel information, (i ', j,) is rounded to a value of 1/4 pixel precision.
- the information indicating the sample position accuracy is extracted from the bit stream.
- the corresponding point determination rule is determined directly from the motion vector 25b, and the sample position of the predicted image is determined based on this.
- the memory read address generation unit 38b Based on the predicted image sample position output from the corresponding point determination unit 37b, the memory read address generation unit 38b generates image data necessary for generating the predicted image in the reference image stored in the frame memory 54. Generates a memory address that specifies the position of, and reads the predicted image generation data.
- the predicted image generation data becomes the predicted image constituent pixels as they are.
- the position at which the predicted image is addressed and sampled can take on the predetermined accuracy as described above, for example, a value of half a pixel or 1 ⁇ 4 pixels.
- an integer pixel value of a predicted image is generated in the internal processing unit 232b based on an instruction of integer precision defined by the interpolation processing instruction information 60.
- the final sample position is already rounded to the precision specified by the interpolation processing precision indication information 60, but the interpolation processing is performed as shown in FIG.
- the expression (15) is processed. If the position is at half-pixel accuracy, the processing is exactly the same as that of the half-pixel generation unit 232 described in the first embodiment.
- a reproduced image can be obtained from the encoded bitstream.
- the image encoding device according to the first to sixth embodiments and the image decoding device according to the seventh embodiment provide a high-speed and complex image code using motion compensation that is transformed only by integer pixel and half pixel addressing. And decryption.
- the image decoding apparatus uses the same configuration, but performs the operation of determining the corresponding point more closely between the reference image and the target program of the image to be predicted. It has been strengthened in order to obtain a natural movement. Thereby, a smoother motion can be obtained.
- the present embodiment even if a bit stream in which a prediction error signal is encoded by another encoding method other than the orthogonal transform encoding, elements for decoding a prediction error signal other than the motion compensation unit 90 are used. By changing the The result is the same as in the seventh embodiment.
- the present invention is also applicable to a decoding device that uses an arbitrary-shaped image object (such as a Video Object P 1 ane) as a unit.
- an arbitrary-shaped image object such as a Video Object P 1 ane
- FIG. 40 is a flowchart showing the operation of the corresponding point determination unit 37c.
- FIG. 41 is a diagram for explaining the operation of the corresponding point determination unit 37c.
- the corresponding point determination unit 37c in the present embodiment inputs the motion vector 25b, the deformation pattern information 26a, the inner processing accuracy instruction information 91, and the in-screen position 27a of the predicted image area. Then, the sample position of the predicted image corresponding to each pixel in the predicted image region is calculated and output based on the following equation.
- the position 27 a in the screen of the predicted image area is a value unique to each predicted image area, but the motion vector 25 b and the deformation pattern information 26 a are different for each predicted image area. Even if it is a unique value, it is encoded for each larger image (for example, image frames and VOPs disclosed in ISO / IECJ TC1 / SC29 / WG11) that combine multiple predicted image regions. Then, encoding may be performed so that the same motion vector and deformation pattern information are used for all the predicted image areas included in the unit. The following describes an example in which a maximum of three motion vectors are used.
- the format of the deformation pattern information 26a may be a bit string composed of a plurality of bits described to directly identify the above-mentioned conversion formulas (16) to (19). Since each conversion corresponds to the number of motion vectors, it may be a bit representing the number of motion vectors.
- the point (i, j) in the predicted image area is associated with (i ′, j ′) in the reference image.
- the sample position of the predicted image should be able to take a value of a predetermined accuracy. For example, if it is rounded to half-pixel accuracy, (i ′, j ′) obtained by the above conversion formula is a value of half-pixel accuracy, and if it is an instruction to round to 1 ⁇ 4 pixel accuracy, i ', j') are values with 1/4 pixel precision.
- the information indicating the sample position accuracy is extracted from the bitstream.
- the corresponding point determination rule is determined directly from the motion vector 25b, and the sample position of the predicted image is determined based on this.
- the image decoding apparatus of the present embodiment when performing the sample position calculation using zero or more motion vectors, all the division operations by W or H are bit-shifted. Since the calculation can be replaced with an operation, the sample position can be determined more quickly, and a reproduced image can be obtained from a coded bit stream by efficiently predicting motions of different complexity. .
- the present invention is also applicable to a decoding apparatus in units of an arbitrary-shaped image object (such as a Video Object P 1 ane) composed of fixed-size blocks, as in the seventh embodiment. .
- the image encoding device and the image decoding device of the present invention constitute a set to constitute a characteristic image encoding / decoding system.
- a motion vector is predicted by using a modified block obtained only by specifying coordinates using integer pixels of a real sample point or half pixels between the pixels.
- partial images for which prediction is not successful with just the amount of translation, such as torque can be predicted efficiently without complicated operations such as affine transformation.
- deformations that can be described by mathematical expressions such as rotation and scaling but also those that cannot be easily described by mathematical expressions, that is, deformations that are difficult to realize by calculation, can be dealt with. Even with compatible decryption devices There is an effect that can efficiently reproduce excellent images.
- the search range of the deformed block matching can be effectively reduced, and the amount of calculation of the entire motion compensation prediction can be reduced.
- the corresponding point can be determined only by referring to the deformation pattern table, there is an effect that a motion accompanying an arbitrary deformation that cannot be expressed by a simple mathematical expression such as an affine transformation can be predicted well.
- the spatial frequency characteristics in the deformed block can be flattened by using a filter, and there is an effect that prediction mismatch can be reduced. Since the decoding device corresponding to the modified block matching and the motion prediction of the image coding device is configured, there is an effect that the image data subjected to the high-speed optimal motion prediction can be decoded and reproduced.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Description
Claims
Priority Applications (10)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US09/180,188 US6404815B1 (en) | 1997-03-17 | 1997-10-23 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
AU47226/97A AU704916B2 (en) | 1997-03-17 | 1997-10-23 | Video encoder, video decoder, video encoding method, video decoding method and video encoding and decoding system |
KR1019980709254A KR100319944B1 (ko) | 1997-03-17 | 1997-10-23 | 화상부호화장치및화상복호장치 |
EP19970909589 EP0905980B1 (en) | 1997-03-17 | 1997-10-23 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
JP54033398A JP3386142B2 (ja) | 1997-03-17 | 1997-10-23 | 画像復号装置および画像復号方法 |
DE69738792T DE69738792D1 (de) | 1997-03-17 | 1997-10-23 | Vorrichtungen, verfahren und system zur codierung und decodierung von bildern |
US10/086,446 US7110456B2 (en) | 1997-03-17 | 2002-03-04 | Video encoder, video decoder, video encoding method, video decoding method, and video encoding and decoding system |
US11/352,377 US8194742B2 (en) | 1997-03-17 | 2006-02-13 | Video decoder |
US11/386,685 US8170105B2 (en) | 1997-03-17 | 2006-03-23 | Video decoder and video decoding method |
US12/648,837 US8098734B2 (en) | 1997-03-17 | 2009-12-29 | Video encoder, video decoder, video encoding method, video decoding method, and video encoding and decoding system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP1997/000834 WO1998042134A1 (en) | 1997-03-17 | 1997-03-17 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
JPPCT/JP97/00834 | 1997-03-17 |
Related Child Applications (3)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US09180188 A-371-Of-International | 1997-10-23 | ||
US09/180,188 A-371-Of-International US6404815B1 (en) | 1997-03-17 | 1997-10-23 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
US10/086,446 Division US7110456B2 (en) | 1997-03-17 | 2002-03-04 | Video encoder, video decoder, video encoding method, video decoding method, and video encoding and decoding system |
Publications (1)
Publication Number | Publication Date |
---|---|
WO1998042135A1 true WO1998042135A1 (en) | 1998-09-24 |
Family
ID=14180217
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP1997/000834 WO1998042134A1 (en) | 1997-03-17 | 1997-03-17 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
PCT/JP1997/003825 WO1998042135A1 (en) | 1997-03-17 | 1997-10-23 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP1997/000834 WO1998042134A1 (en) | 1997-03-17 | 1997-03-17 | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
Country Status (7)
Country | Link |
---|---|
US (5) | US6404815B1 (ja) |
EP (8) | EP2252066A3 (ja) |
JP (1) | JP3386142B2 (ja) |
KR (1) | KR100319944B1 (ja) |
AU (2) | AU1941797A (ja) |
DE (4) | DE69740005D1 (ja) |
WO (2) | WO1998042134A1 (ja) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008136237A (ja) * | 2008-01-16 | 2008-06-12 | Hitachi Ltd | 画像の符号化/復号化装置、符号化/復号化プログラム及び符号化/復号化方法 |
WO2010021153A1 (ja) | 2008-08-21 | 2010-02-25 | パナソニック株式会社 | 動き検出装置 |
JP2011055556A (ja) * | 2010-12-14 | 2011-03-17 | Hitachi Ltd | 画像の符号化/復号化装置、符号化/復号化プログラム及び符号化/復号化方法 |
CN101218830B (zh) * | 2005-07-22 | 2011-10-12 | 三菱电机株式会社 | 图像编码装置和方法、以及图像解码装置和方法 |
US8917763B2 (en) | 2011-03-07 | 2014-12-23 | Panasonic Corporation | Motion compensation apparatus, video coding apparatus, video decoding apparatus, motion compensation method, program, and integrated circuit |
US10659802B2 (en) | 2016-08-15 | 2020-05-19 | Nokia Technologies Oy | Video encoding and decoding |
Families Citing this family (101)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1998042134A1 (en) * | 1997-03-17 | 1998-09-24 | Mitsubishi Denki Kabushiki Kaisha | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
US6690731B1 (en) * | 1997-12-01 | 2004-02-10 | Neostar, Inc. | Method and apparatus for diagonal processing of video data |
US8179971B1 (en) | 2003-12-08 | 2012-05-15 | G&H Nevada-Tek | Method and apparatus for video data compression |
FR2795589B1 (fr) * | 1999-06-11 | 2001-10-05 | Centre Nat Rech Scient | Decodeur video optimal base sur les standards de type mpeg |
JP2001275110A (ja) * | 2000-03-24 | 2001-10-05 | Matsushita Electric Ind Co Ltd | 動的なループ及びポストフィルタリングのための方法及び装置 |
KR100359115B1 (ko) * | 2000-05-24 | 2002-11-04 | 삼성전자 주식회사 | 영상 코딩 방법 |
US20030021347A1 (en) * | 2001-07-24 | 2003-01-30 | Koninklijke Philips Electronics N.V. | Reduced comlexity video decoding at full resolution using video embedded resizing |
US7181070B2 (en) * | 2001-10-30 | 2007-02-20 | Altera Corporation | Methods and apparatus for multiple stage video decoding |
US7403981B2 (en) * | 2002-01-04 | 2008-07-22 | Quicksilver Technology, Inc. | Apparatus and method for adaptive multimedia reception and transmission in communication environments |
US8284844B2 (en) | 2002-04-01 | 2012-10-09 | Broadcom Corporation | Video decoding system supporting multiple standards |
US20090118019A1 (en) | 2002-12-10 | 2009-05-07 | Onlive, Inc. | System for streaming databases serving real-time applications used through streaming interactive video |
US8711923B2 (en) | 2002-12-10 | 2014-04-29 | Ol2, Inc. | System and method for selecting a video encoding format based on feedback data |
US8549574B2 (en) | 2002-12-10 | 2013-10-01 | Ol2, Inc. | Method of combining linear content and interactive content compressed together as streaming interactive video |
US8366552B2 (en) | 2002-12-10 | 2013-02-05 | Ol2, Inc. | System and method for multi-stream video compression |
US8964830B2 (en) | 2002-12-10 | 2015-02-24 | Ol2, Inc. | System and method for multi-stream video compression using multiple encoding formats |
US9108107B2 (en) | 2002-12-10 | 2015-08-18 | Sony Computer Entertainment America Llc | Hosting and broadcasting virtual events using streaming interactive video |
US10201760B2 (en) | 2002-12-10 | 2019-02-12 | Sony Interactive Entertainment America Llc | System and method for compressing video based on detected intraframe motion |
US9077991B2 (en) | 2002-12-10 | 2015-07-07 | Sony Computer Entertainment America Llc | System and method for utilizing forward error correction with video compression |
US9061207B2 (en) | 2002-12-10 | 2015-06-23 | Sony Computer Entertainment America Llc | Temporary decoder apparatus and method |
US8526490B2 (en) | 2002-12-10 | 2013-09-03 | Ol2, Inc. | System and method for video compression using feedback including data related to the successful receipt of video content |
US9138644B2 (en) | 2002-12-10 | 2015-09-22 | Sony Computer Entertainment America Llc | System and method for accelerated machine switching |
US9314691B2 (en) | 2002-12-10 | 2016-04-19 | Sony Computer Entertainment America Llc | System and method for compressing video frames or portions thereof based on feedback information from a client device |
US9446305B2 (en) | 2002-12-10 | 2016-09-20 | Sony Interactive Entertainment America Llc | System and method for improving the graphics performance of hosted applications |
US9192859B2 (en) | 2002-12-10 | 2015-11-24 | Sony Computer Entertainment America Llc | System and method for compressing video based on latency measurements and other feedback |
JP2004241984A (ja) * | 2003-02-05 | 2004-08-26 | Toshiba Corp | データ伝送装置 |
US7869516B2 (en) * | 2003-03-31 | 2011-01-11 | Hewlett-Packard Development Company, L.P. | Motion estimation using bit-wise block comparisons for video compresssion |
US7620254B2 (en) * | 2003-08-07 | 2009-11-17 | Trident Microsystems (Far East) Ltd. | Apparatus and method for motion-vector-aided interpolation of a pixel of an intermediate image of an image sequence |
KR100703283B1 (ko) * | 2004-03-15 | 2007-04-03 | 삼성전자주식회사 | 회전 매칭을 통해 움직임을 예측하는 영상 부호화 장치 및방법 |
US8018463B2 (en) * | 2004-05-10 | 2011-09-13 | Nvidia Corporation | Processor for video data |
US8130825B2 (en) * | 2004-05-10 | 2012-03-06 | Nvidia Corporation | Processor for video data encoding/decoding |
US7986733B2 (en) * | 2004-07-30 | 2011-07-26 | Broadcom Corporation | Tertiary content addressable memory based motion estimator |
US20060215036A1 (en) * | 2005-03-25 | 2006-09-28 | Multivision Intelligent Surveillance (Hk) Ltd. | Method and apparatus for video stabilization |
US7660354B2 (en) * | 2005-05-11 | 2010-02-09 | Fang Shi | Temporal error concealment for bi-directionally predicted frames |
CN100474929C (zh) * | 2005-09-07 | 2009-04-01 | 深圳市海思半导体有限公司 | 一种运动补偿数据载入装置及方法 |
WO2007063819A1 (ja) * | 2005-11-29 | 2007-06-07 | Pioneer Corporation | 記録装置、再生装置、記録方法、記録プログラムおよびコンピュータに読み取り可能な記録媒体 |
US20080111923A1 (en) * | 2006-11-09 | 2008-05-15 | Scheuermann W James | Processor for video data |
JP4615508B2 (ja) * | 2006-12-27 | 2011-01-19 | シャープ株式会社 | 画像表示装置及び方法、画像処理装置及び方法 |
EP2153660A4 (en) * | 2007-04-09 | 2013-01-23 | Lg Electronics Inc | METHOD AND DEVICE FOR PROCESSING A VIDEO SIGNAL |
US8724708B2 (en) * | 2007-04-27 | 2014-05-13 | Panasonic Corporation | Image decoding device, image decoding system, image decoding method, and integrated circuit |
KR100813989B1 (ko) * | 2007-06-18 | 2008-03-14 | 삼성전자주식회사 | 복수개의 주사 패턴을 이용한 부호화 방법, 복호화 방법,및 그 장치 |
FR2917872A1 (fr) * | 2007-06-25 | 2008-12-26 | France Telecom | Procedes et dispositifs de codage et de decodage d'une sequence d'images representee a l'aide de tubes de mouvement, produits programme d'ordinateur et signal correspondant. |
US9168457B2 (en) | 2010-09-14 | 2015-10-27 | Sony Computer Entertainment America Llc | System and method for retaining system state |
US8443275B2 (en) * | 2008-01-15 | 2013-05-14 | Texas Instruments Incorporated | Lossy compression technique for video encoder bandwidth reduction using compression error data |
KR101456491B1 (ko) * | 2008-05-08 | 2014-11-03 | 삼성전자주식회사 | 복수의 참조 픽처에 기초한 영상 부호화, 복호화 방법 및장치 |
KR101493325B1 (ko) * | 2008-09-03 | 2015-02-16 | 삼성전자주식회사 | 정밀 움직임 예측을 기반으로 한 프레임 보간 장치 및 그 방법 |
KR101624649B1 (ko) * | 2009-08-14 | 2016-05-26 | 삼성전자주식회사 | 계층적인 부호화 블록 패턴 정보를 이용한 비디오 부호화 방법 및 장치, 비디오 복호화 방법 및 장치 |
BR112012019676A2 (pt) | 2010-02-09 | 2016-05-03 | Nippon Telegraph & Telephone | método de codificação preditiva de vetor de movimento, método de decodificação preditiva de vetor de movimento, aparelho de codificação de imagem em movimento, aparelho de decodificação de imagem em movimento e programas destes |
KR20120112723A (ko) | 2010-02-09 | 2012-10-11 | 니폰덴신뎅와 가부시키가이샤 | 움직임 벡터 예측 부호화 방법, 움직임 벡터 예측 복호 방법, 동화상 부호화 장치, 동화상 복호 장치 및 그들의 프로그램 |
US8581974B2 (en) * | 2010-05-06 | 2013-11-12 | Aptina Imaging Corporation | Systems and methods for presence detection |
EP2596636A1 (en) * | 2010-07-21 | 2013-05-29 | Dolby Laboratories Licensing Corporation | Reference processing using advanced motion models for video coding |
US9357229B2 (en) * | 2010-07-28 | 2016-05-31 | Qualcomm Incorporated | Coding motion vectors in video coding |
EP2421266A1 (en) * | 2010-08-19 | 2012-02-22 | Thomson Licensing | Method for reconstructing a current block of an image and corresponding encoding method, corresponding devices as well as storage medium carrying an images encoded in a bit stream |
US9066102B2 (en) | 2010-11-17 | 2015-06-23 | Qualcomm Incorporated | Reference picture list construction for generalized P/B frames in video coding |
EP4250732B1 (en) | 2011-01-07 | 2024-03-20 | Nokia Technologies Oy | Motion prediction in video coding |
CN106713910B (zh) * | 2011-06-14 | 2019-12-10 | 三星电子株式会社 | 对图像进行解码的设备 |
JP2013102297A (ja) | 2011-11-07 | 2013-05-23 | Canon Inc | 画像符号化方法、画像符号化装置及びプログラム、画像復号方法、画像復号装置及びプログラム |
JP5871628B2 (ja) * | 2011-11-07 | 2016-03-01 | キヤノン株式会社 | 画像符号化装置、画像符号化方法及びプログラム、画像復号装置、画像復号方法及びプログラム |
US9769473B2 (en) | 2012-06-08 | 2017-09-19 | Apple Inc. | Predictive video coder with low power reference picture transformation |
CN103929649B (zh) * | 2014-04-28 | 2017-04-26 | 广西大学 | 一种视频压缩感知的重建方法 |
CN107809642B (zh) * | 2015-02-16 | 2020-06-16 | 华为技术有限公司 | 用于视频图像编码和解码的方法、编码设备和解码设备 |
KR102248172B1 (ko) * | 2015-03-16 | 2021-05-04 | 한양대학교 산학협력단 | 영상 분석을 이용한 비디오 부호화/복호화 방법 및 장치 |
CN111526361B (zh) | 2016-02-06 | 2022-05-13 | 华为技术有限公司 | 图像编解码方法及装置 |
ES2692864B1 (es) * | 2016-02-25 | 2019-10-21 | Kt Corp | Metodo y aparato para procesar senales de video |
US10560712B2 (en) * | 2016-05-16 | 2020-02-11 | Qualcomm Incorporated | Affine motion prediction for video coding |
US10979732B2 (en) * | 2016-10-04 | 2021-04-13 | Qualcomm Incorporated | Adaptive motion vector precision for video coding |
US10448010B2 (en) | 2016-10-05 | 2019-10-15 | Qualcomm Incorporated | Motion vector prediction for affine motion models in video coding |
CN108271028A (zh) * | 2016-12-30 | 2018-07-10 | 北京优朋普乐科技有限公司 | 一种子像素全搜索方法及装置 |
US10897269B2 (en) | 2017-09-14 | 2021-01-19 | Apple Inc. | Hierarchical point cloud compression |
US10861196B2 (en) | 2017-09-14 | 2020-12-08 | Apple Inc. | Point cloud compression |
US11818401B2 (en) | 2017-09-14 | 2023-11-14 | Apple Inc. | Point cloud geometry compression using octrees and binary arithmetic encoding with adaptive look-up tables |
US11113845B2 (en) | 2017-09-18 | 2021-09-07 | Apple Inc. | Point cloud compression using non-cubic projections and masks |
US10909725B2 (en) | 2017-09-18 | 2021-02-02 | Apple Inc. | Point cloud compression |
US11877001B2 (en) | 2017-10-10 | 2024-01-16 | Qualcomm Incorporated | Affine prediction in video coding |
US10607373B2 (en) | 2017-11-22 | 2020-03-31 | Apple Inc. | Point cloud compression with closed-loop color conversion |
US10939129B2 (en) | 2018-04-10 | 2021-03-02 | Apple Inc. | Point cloud compression |
US11010928B2 (en) | 2018-04-10 | 2021-05-18 | Apple Inc. | Adaptive distance based point cloud compression |
US10909726B2 (en) | 2018-04-10 | 2021-02-02 | Apple Inc. | Point cloud compression |
US10909727B2 (en) | 2018-04-10 | 2021-02-02 | Apple Inc. | Hierarchical point cloud compression with smoothing |
EP3794827A4 (en) * | 2018-06-28 | 2021-03-31 | Huawei Technologies Co., Ltd. | MEMORY ACCESS WINDOW AND PADDING FOR MOTION VECTOR REFINEMENT AND MOTION COMPENSATION |
US11017566B1 (en) | 2018-07-02 | 2021-05-25 | Apple Inc. | Point cloud compression with adaptive filtering |
US11202098B2 (en) | 2018-07-05 | 2021-12-14 | Apple Inc. | Point cloud compression with multi-resolution video encoding |
US11012713B2 (en) | 2018-07-12 | 2021-05-18 | Apple Inc. | Bit stream structure for compressed point cloud data |
US11367224B2 (en) | 2018-10-02 | 2022-06-21 | Apple Inc. | Occupancy map block-to-patch information compression |
US11430155B2 (en) | 2018-10-05 | 2022-08-30 | Apple Inc. | Quantized depths for projection point cloud compression |
US11025951B2 (en) * | 2019-01-13 | 2021-06-01 | Tencent America LLC | Method and apparatus for video coding |
US11683501B2 (en) | 2019-01-17 | 2023-06-20 | Tencent America LLC | Method and apparatus for video coding |
US11057564B2 (en) | 2019-03-28 | 2021-07-06 | Apple Inc. | Multiple layer flexure for supporting a moving image sensor |
CN115460424B (zh) * | 2019-06-05 | 2024-09-20 | 杭州海康威视数字技术股份有限公司 | 一种编解码方法、装置及其设备 |
US11711544B2 (en) | 2019-07-02 | 2023-07-25 | Apple Inc. | Point cloud compression with supplemental information messages |
US11627314B2 (en) | 2019-09-27 | 2023-04-11 | Apple Inc. | Video-based point cloud compression with non-normative smoothing |
US11562507B2 (en) | 2019-09-27 | 2023-01-24 | Apple Inc. | Point cloud compression using video encoding with time consistent patches |
US11202078B2 (en) * | 2019-09-27 | 2021-12-14 | Apple Inc. | Dynamic point cloud compression using inter-prediction |
US11538196B2 (en) | 2019-10-02 | 2022-12-27 | Apple Inc. | Predictive coding for point cloud compression |
US11895307B2 (en) | 2019-10-04 | 2024-02-06 | Apple Inc. | Block-based predictive coding for point cloud compression |
US11798196B2 (en) | 2020-01-08 | 2023-10-24 | Apple Inc. | Video-based point cloud compression with predicted patches |
US11625866B2 (en) | 2020-01-09 | 2023-04-11 | Apple Inc. | Geometry encoding using octrees and predictive trees |
US11620768B2 (en) | 2020-06-24 | 2023-04-04 | Apple Inc. | Point cloud geometry compression using octrees with multiple scan orders |
US11615557B2 (en) | 2020-06-24 | 2023-03-28 | Apple Inc. | Point cloud compression using octrees with slicing |
US11948338B1 (en) | 2021-03-29 | 2024-04-02 | Apple Inc. | 3D volumetric content encoding using 2D videos and simplified 3D meshes |
CN113658035B (zh) * | 2021-08-17 | 2023-08-08 | 北京百度网讯科技有限公司 | 脸部变换方法、装置、设备、存储介质以及产品 |
US20230134017A1 (en) * | 2021-11-01 | 2023-05-04 | Tencent America LLC | Template-matching based adaptive motion vector resolution (amvr) for bi-prediction and an affine mode |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH048585U (ja) * | 1990-05-14 | 1992-01-27 | ||
JPH05219498A (ja) * | 1992-02-06 | 1993-08-27 | Fujitsu Ltd | 動画像符号化装置 |
JPH05244585A (ja) * | 1992-02-28 | 1993-09-21 | Victor Co Of Japan Ltd | 予測符号化装置及び復号化装置 |
JPH0698314A (ja) * | 1992-06-26 | 1994-04-08 | Sony Corp | 画像信号符号化、復号化方法及び画像信号記録媒体 |
JPH06153185A (ja) * | 1992-11-06 | 1994-05-31 | Toshiba Corp | 動き補償装置及びこれを用いた符号化装置 |
JPH0750773A (ja) * | 1993-03-08 | 1995-02-21 | Sony Electron Inc | ビデオ圧縮処理用装置及び方法 |
Family Cites Families (57)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4661849A (en) * | 1985-06-03 | 1987-04-28 | Pictel Corporation | Method and apparatus for providing motion estimation signals for communicating image sequences |
DE3629472A1 (de) * | 1986-08-29 | 1988-03-03 | Licentia Gmbh | Verfahren zur bewegungskompensierten bild-zu-bild-praediktionscodierung |
US4972260A (en) * | 1988-08-22 | 1990-11-20 | Matsushita Electric Industrial Co., Ltd. | Apparatus for coding a moving-picture signal |
US5236682A (en) * | 1989-10-25 | 1993-08-17 | Hoechst Aktiengesellschaft | Process for producing crystalline sodium silicates having a layered structure |
JP2569219B2 (ja) * | 1990-01-31 | 1997-01-08 | 富士通株式会社 | 動画像予測方式 |
JPH048585A (ja) | 1990-04-26 | 1992-01-13 | Fuji Photo Film Co Ltd | 情報記録媒体の製造方法 |
SE465394B (sv) * | 1990-08-15 | 1991-09-02 | Televerket | Metod foer roerelsekompensering och elastisk deformation i bildsekvenser |
US5467136A (en) * | 1991-05-31 | 1995-11-14 | Kabushiki Kaisha Toshiba | Video decoder for determining a motion vector from a scaled vector and a difference vector |
US5534927A (en) | 1991-11-12 | 1996-07-09 | Nippon Hoso Kyokai | Method of performing high efficiency coding of image signal and system therefor |
JP2549479B2 (ja) * | 1991-12-06 | 1996-10-30 | 日本電信電話株式会社 | 動き補償フレーム間帯域分割符号化処理方法 |
JPH05236466A (ja) * | 1992-02-25 | 1993-09-10 | Nec Corp | 動き補償フレーム間予測画像符号化装置及び方法 |
US5475446A (en) * | 1992-03-09 | 1995-12-12 | Matsushita Electric Industrial Co., Ltd. | Picture signal motion detector employing partial decimation of pixel blocks |
JPH05268594A (ja) * | 1992-03-18 | 1993-10-15 | Sony Corp | 動画像の動き検出装置 |
US5247355A (en) * | 1992-06-11 | 1993-09-21 | Northwest Starscan Limited Partnership | Gridlocked method and system for video motion compensation |
DE4322343C2 (de) * | 1992-07-06 | 1996-10-02 | Mitsubishi Electric Corp | Mittel zum Erfassen eines Bewegungsvektors und Verfahren zum Bestimmen eines Bewegungsvektors |
JPH06113287A (ja) * | 1992-09-30 | 1994-04-22 | Matsushita Electric Ind Co Ltd | 画像符号化装置と画像復号化装置 |
US5398079A (en) | 1993-01-27 | 1995-03-14 | General Instrument Corporation | Half-pixel interpolation for a motion compensated digital video system |
KR100219005B1 (ko) * | 1993-02-22 | 1999-09-01 | 모리시타 요이찌 | 화상부호화방법 및 화상부호화장치 |
TW224553B (en) * | 1993-03-01 | 1994-06-01 | Sony Co Ltd | Method and apparatus for inverse discrete consine transform and coding/decoding of moving picture |
DE69416717T2 (de) * | 1993-05-21 | 1999-10-07 | Nippon Telegraph And Telephone Corp., Tokio/Tokyo | Bewegtbildkoder und -dekoder |
IT1262275B (it) | 1993-05-26 | 1996-06-19 | Metodo e dispositivo per la formazione automatica di pacchi di lamierini metallici, tranciati, per la fabbricazione di armature di macchine elettriche, o simili. | |
FI97008C (fi) | 1993-06-02 | 1996-09-25 | Nokia Oy Ab | Menetelmä videokuvan ennustamiseksi käsittelyjärjestyksessä edellisen kuvan perusteella |
US5473379A (en) * | 1993-11-04 | 1995-12-05 | At&T Corp. | Method and apparatus for improving motion compensation in digital video coding |
US5719630A (en) * | 1993-12-10 | 1998-02-17 | Nec Corporation | Apparatus for compressive coding in moving picture coding device |
JP3266416B2 (ja) * | 1994-04-18 | 2002-03-18 | ケイディーディーアイ株式会社 | 動き補償フレーム間符号化復号装置 |
JP2897649B2 (ja) * | 1994-08-08 | 1999-05-31 | 日本ビクター株式会社 | 動き補償予測符号化装置 |
US5684538A (en) * | 1994-08-18 | 1997-11-04 | Hitachi, Ltd. | System and method for performing video coding/decoding using motion compensation |
EP0708563B1 (en) * | 1994-10-19 | 2004-12-22 | Matsushita Electric Industrial Co., Ltd. | Image decoding device |
TW257924B (en) * | 1995-03-18 | 1995-09-21 | Daewoo Electronics Co Ltd | Method and apparatus for encoding a video signal using feature point based motion estimation |
US5650823A (en) | 1995-03-27 | 1997-07-22 | International Business Machines Corporation | Half pel motion estimation method for B pictures |
US5812791A (en) * | 1995-05-10 | 1998-09-22 | Cagent Technologies, Inc. | Multiple sequence MPEG decoder |
GB2301972B (en) | 1995-06-06 | 1999-10-20 | Sony Uk Ltd | Video compression |
US5774676A (en) * | 1995-10-03 | 1998-06-30 | S3, Incorporated | Method and apparatus for decompression of MPEG compressed data in a computer system |
JPH09102954A (ja) * | 1995-10-04 | 1997-04-15 | Matsushita Electric Ind Co Ltd | 1つまたは2つの予測ブロックからブロックの画素値を計算する方法 |
US5761423A (en) * | 1995-11-14 | 1998-06-02 | Industrial Technology Research Institute | Scalable picture storage access architecture for video decoding |
US6968003B1 (en) * | 1996-01-29 | 2005-11-22 | International Business Machines Corporation | Speed-memory tradeoff for MPEG decoders |
US6049362A (en) * | 1996-02-14 | 2000-04-11 | International Business Machines Corporation | Dual prime motion estimation system and method |
JP3183155B2 (ja) * | 1996-03-18 | 2001-07-03 | 株式会社日立製作所 | 画像復号化装置、及び、画像復号化方法 |
BR9710780A (pt) * | 1996-07-31 | 1999-08-17 | Schionogi & Co Ltd | Novos compostos de para-terefenila |
JP3426092B2 (ja) * | 1996-09-12 | 2003-07-14 | シャープ株式会社 | 動画像符号化装置における動き補償フレーム間予測方法 |
JPH1098729A (ja) * | 1996-09-20 | 1998-04-14 | Hitachi Ltd | フレーム間予測画像の合成方法 |
WO1998017068A1 (en) * | 1996-10-11 | 1998-04-23 | Sarnoff Corporation | Stereoscopic video coding and decoding apparatus and method |
US6005980A (en) * | 1997-03-07 | 1999-12-21 | General Instrument Corporation | Motion estimation and compensation of video object planes for interlaced digital video |
US5991447A (en) * | 1997-03-07 | 1999-11-23 | General Instrument Corporation | Prediction and coding of bi-directionally predicted video object planes for interlaced digital video |
WO1998042134A1 (en) * | 1997-03-17 | 1998-09-24 | Mitsubishi Denki Kabushiki Kaisha | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system |
US6249613B1 (en) * | 1997-03-31 | 2001-06-19 | Sharp Laboratories Of America, Inc. | Mosaic generation and sprite-based coding with automatic foreground and background separation |
US6633611B2 (en) | 1997-04-24 | 2003-10-14 | Mitsubishi Denki Kabushiki Kaisha | Method and apparatus for region-based moving image encoding and decoding |
EP2288164B1 (en) * | 1997-06-09 | 2013-09-11 | Hitachi, Ltd. | Image decoding method and image decoder |
JPH1169345A (ja) * | 1997-06-11 | 1999-03-09 | Fujitsu Ltd | フレーム間予測動画像符号化装置及び復号装置並びにフレーム間予測動画像符号化方法及び復号方法 |
US6249318B1 (en) * | 1997-09-12 | 2001-06-19 | 8×8, Inc. | Video coding/decoding arrangement and method therefor |
KR100404358B1 (ko) * | 1998-01-28 | 2003-11-03 | 시오노기세이야쿠가부시키가이샤 | 신규 삼환식 화합물 |
US6249618B1 (en) * | 1998-12-18 | 2001-06-19 | Syscan, Inc. | Circuit architecture and method for switching sensor resolution |
US6625211B1 (en) * | 1999-02-25 | 2003-09-23 | Matsushita Electric Industrial Co., Ltd. | Method and apparatus for transforming moving picture coding system |
GB9920256D0 (en) * | 1999-08-26 | 1999-10-27 | Wave Limited M | Motion estimation and compensation in video compression |
JP2005503695A (ja) * | 2001-05-29 | 2005-02-03 | 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 | ビデオトランスコード方法及び装置 |
EP2099228B1 (en) | 2001-09-14 | 2014-11-12 | NTT DoCoMo, Inc. | Coding method, decoding method, coding apparatus, decoding apparatus, image processing system, coding program, and decoding program |
JP3824071B2 (ja) * | 2001-12-25 | 2006-09-20 | 信越化学工業株式会社 | 室温硬化性オルガノポリシロキサン組成物 |
-
1997
- 1997-03-17 WO PCT/JP1997/000834 patent/WO1998042134A1/ja active Application Filing
- 1997-03-17 AU AU19417/97A patent/AU1941797A/en not_active Abandoned
- 1997-10-23 WO PCT/JP1997/003825 patent/WO1998042135A1/ja active IP Right Grant
- 1997-10-23 DE DE69740005T patent/DE69740005D1/de not_active Expired - Lifetime
- 1997-10-23 AU AU47226/97A patent/AU704916B2/en not_active Expired
- 1997-10-23 KR KR1019980709254A patent/KR100319944B1/ko not_active IP Right Cessation
- 1997-10-23 JP JP54033398A patent/JP3386142B2/ja not_active Expired - Lifetime
- 1997-10-23 EP EP20100008595 patent/EP2252066A3/en not_active Withdrawn
- 1997-10-23 EP EP20100008597 patent/EP2252068A3/en not_active Withdrawn
- 1997-10-23 DE DE69739999T patent/DE69739999D1/de not_active Expired - Lifetime
- 1997-10-23 EP EP20100008596 patent/EP2252067A3/en not_active Withdrawn
- 1997-10-23 EP EP20060018030 patent/EP1725045B1/en not_active Expired - Lifetime
- 1997-10-23 DE DE69738792T patent/DE69738792D1/de not_active Expired - Lifetime
- 1997-10-23 EP EP20100008594 patent/EP2252065A3/en not_active Withdrawn
- 1997-10-23 DE DE69740004T patent/DE69740004D1/de not_active Expired - Lifetime
- 1997-10-23 EP EP19970909589 patent/EP0905980B1/en not_active Expired - Lifetime
- 1997-10-23 US US09/180,188 patent/US6404815B1/en not_active Expired - Lifetime
- 1997-10-23 EP EP20060018031 patent/EP1725046B1/en not_active Expired - Lifetime
- 1997-10-23 EP EP20060018032 patent/EP1725047B1/en not_active Expired - Lifetime
-
2002
- 2002-03-04 US US10/086,446 patent/US7110456B2/en not_active Expired - Lifetime
-
2006
- 2006-02-13 US US11/352,377 patent/US8194742B2/en not_active Expired - Fee Related
- 2006-03-23 US US11/386,685 patent/US8170105B2/en not_active Expired - Fee Related
-
2009
- 2009-12-29 US US12/648,837 patent/US8098734B2/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH048585U (ja) * | 1990-05-14 | 1992-01-27 | ||
JPH05219498A (ja) * | 1992-02-06 | 1993-08-27 | Fujitsu Ltd | 動画像符号化装置 |
JPH05244585A (ja) * | 1992-02-28 | 1993-09-21 | Victor Co Of Japan Ltd | 予測符号化装置及び復号化装置 |
JPH0698314A (ja) * | 1992-06-26 | 1994-04-08 | Sony Corp | 画像信号符号化、復号化方法及び画像信号記録媒体 |
JPH06153185A (ja) * | 1992-11-06 | 1994-05-31 | Toshiba Corp | 動き補償装置及びこれを用いた符号化装置 |
JPH0750773A (ja) * | 1993-03-08 | 1995-02-21 | Sony Electron Inc | ビデオ圧縮処理用装置及び方法 |
Non-Patent Citations (1)
Title |
---|
JOZAWA H: "MOTION COMPENSATED PREDICTION USING AN AFFINE MOTION MODEL", IEICE TECHNICAL REPORT, DENSHI JOUHOU TSUUSHIN GAKKAI, JP, 26 July 1994 (1994-07-26), JP, pages 01 - 08, XP002905207, ISSN: 0913-5685 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101218830B (zh) * | 2005-07-22 | 2011-10-12 | 三菱电机株式会社 | 图像编码装置和方法、以及图像解码装置和方法 |
JP2008136237A (ja) * | 2008-01-16 | 2008-06-12 | Hitachi Ltd | 画像の符号化/復号化装置、符号化/復号化プログラム及び符号化/復号化方法 |
JP4683050B2 (ja) * | 2008-01-16 | 2011-05-11 | 株式会社日立製作所 | 画像の符号化/復号化装置、符号化/復号化プログラム及び符号化/復号化方法 |
WO2010021153A1 (ja) | 2008-08-21 | 2010-02-25 | パナソニック株式会社 | 動き検出装置 |
JP2011055556A (ja) * | 2010-12-14 | 2011-03-17 | Hitachi Ltd | 画像の符号化/復号化装置、符号化/復号化プログラム及び符号化/復号化方法 |
US8917763B2 (en) | 2011-03-07 | 2014-12-23 | Panasonic Corporation | Motion compensation apparatus, video coding apparatus, video decoding apparatus, motion compensation method, program, and integrated circuit |
US10659802B2 (en) | 2016-08-15 | 2020-05-19 | Nokia Technologies Oy | Video encoding and decoding |
Also Published As
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO1998042135A1 (en) | Image encoder, image decoder, image encoding method, image decoding method and image encoding/decoding system | |
KR100412622B1 (ko) | 화상복호화장치및화상복호화방법 | |
JP2002532026A (ja) | 動き推定とブロックマッチング・パターンの改良 | |
JPH08251583A (ja) | 画像符号化装置及び画像復号装置 | |
KR100246167B1 (ko) | 듀얼프라임 움직임 추정시스템 및 방법 | |
JP2003116053A (ja) | 特殊効果データ符号化方法及び特殊効果表示方法及び特殊効果データ編集方法 | |
JPH10276439A (ja) | 領域統合が可能な動き補償フレーム間予測方式を用いた動画像符号化・復号化装置 | |
EP1819173B1 (en) | Motion vector predictive encoding apparatus and decoding apparatus | |
KR100926752B1 (ko) | 동영상 부호화를 위한 미세 움직임 추정 방법 및 장치 | |
JP2004040575A (ja) | 動き補償装置 | |
JP3235453B2 (ja) | 画像符号化装置 | |
JPH11122619A (ja) | 動画像符号化および復号化方法 | |
JP2002010267A (ja) | 動きベクトル変換装置及び方法 | |
KR100448700B1 (ko) | 프레임속도변환의 크로스 페이드 보간 장치 및 방법 | |
JPH09261661A (ja) | 2つの基準ピクチャから双方向コード化ピクチャを形成するための方法 | |
JPH08228351A (ja) | 動画像の動き補償予測符号化方法 | |
JP4034180B2 (ja) | 動画像圧縮符号化装置、及び動きベクトル検出方法 | |
JPH1023422A (ja) | 画像符号化装置及び画像復号化装置及び画像符号化方法及び画像復号化方法及び画像符号化復号化システム | |
JP2001313948A (ja) | 動きベクトル変換装置及び方法 | |
MXPA98007455A (en) | Video encoding and decoding | |
JPH09116908A (ja) | 画像符号化装置及び画像復号装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AU JP KR US |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1997909589 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1019980709254 Country of ref document: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 09180188 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: 1997909589 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 1019980709254 Country of ref document: KR |
|
WWG | Wipo information: grant in national office |
Ref document number: 1019980709254 Country of ref document: KR |
|
WWG | Wipo information: grant in national office |
Ref document number: 1997909589 Country of ref document: EP |