JPH0879763A - Moving image coder and moving image decoder - Google Patents

Abstract

PURPOSE: To code a moving quantity of a grating point of an optional rectangular area efficiently by dividing an image into plural rectangular areas and coding each rectangular whose apex movement is '0' altogether. CONSTITUTION: A comparator 113 discriminates whether or not a moving quantity of each grating point of each rectangular of an image is '0'. A counter 114 counts number of gratings whose moving quantity is '0' successively according to the coding sequence of grating points in each frame when the moving quantity is '0'. A variable length coding section 105 processes consecutive numbers of grating points whose moving quantity is '0' and moving quantity of grating points whose moving quantization is not '0' as one code altogether. Through the constitution above, all grating moving quantities are not coded but consecutive grating point moving quantities are coded altogether as to rectangular areas whose grating point moving quantity is '0', then efficient coding is attained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding apparatus and a moving picture decoding apparatus, and more particularly to a moving picture coding apparatus and a moving picture decoding apparatus for coding image data with high efficiency. For example, it is applied to digital image processing.

[0002]

2. Description of the Related Art Conventionally, in a moving picture coding system, there is a method of carrying out coding and decoding in units of regions divided according to the shape of an image. For example, there is a method in which an image is divided into a plurality of regions having uniform properties by a deformation grid as shown in FIG. These are described in the following known documents.
“Very Low Bitrate Video Coder using Warping
"Predicteion" (1993 Image Coding Symposium 8-7, p.
p.167-168), "Adaptive variable block shape KL transform coding of images" (1991 Image Coding Symposium 8-14, p.
p. 235-238). The vertices of this modified grid are hereinafter referred to as grid points, and are represented by ◯ in FIG.

FIG. 11 is a block diagram showing an example of the configuration of a conventional moving picture coding apparatus. In the figure, 1101 is a frame memory, 1102 is a quadrilateral area detection / division unit, 1103 is an orthogonal transformation unit, and 1104 is a quantum. An encoding unit, 1105 is a variable length encoding unit, 1106 is a code buffer, 1107 is an inverse quantization unit, 1108 is an inverse orthogonal transform unit, 1109 is a frame memory, 1110 is a motion compensation unit, 1111 is a lattice point movement amount detection unit, Reference numeral 1112 is a rectangular area detection / division unit, and 1113 is a switch.

A frame memory 110 for storing an input image
1 is capable of storing a plurality of frames because the motion vector is calculated. Square area detection / division units 1102 and 1112
Detects the positions of the grid points in the quadrangle area and creates position information, and divides the image into quadrangle areas surrounded by the grid points as shown in FIG. The orthogonal transformation unit 1103 performs orthogonal transformation on the pixels in the rectangular area created by the rectangular area detection / division unit 1102 or 1112. It should be noted that the orthogonal transformation does not necessarily mean that the rectangular area detecting / dividing units 1102, 11 are
It is not necessary to perform the processing in units of 12 quadrangular areas, and the processing unit may be independently determined.

For example, the orthogonal transformation may be performed in units of square blocks. The quantization unit 1104 quantizes the converted data with an appropriate quantization step size. The variable length coding unit 1105 performs variable length coding on the quantized value or the lattice point position information from the square area detection / division units 1102 and 1112 according to a predetermined code table, and outputs it as a transmission path code. . The code buffer 1106 accumulates and smoothes the data from the variable length coding unit 1105 in order to output the data at a constant rate.

The inverse quantizer 1107 is a quantizer 1104.
Dequantize the output from. The inverse orthogonal transform unit 1108
The output of the inverse quantization unit 1107 is subjected to inverse orthogonal transform. In addition,
The processing unit of the inverse orthogonal transform matches the processing unit of the orthogonal transform. The frame memory 1109 stores an image required for prediction. The motion compensating unit 1110 performs motion compensation prediction using the grid point moving amount output from the grid point moving amount detecting unit 1111 described later. The grid point movement amount detection unit 1111 calculates the movement amount of the grid points when moving each grid point of the quadrangle region created by the quadrangle region detection / division unit 1112 to create a prediction image that approximates the encoding target frame. To do. The switch 1113 switches between processing within the frame and processing between the frames after the orthogonal transformation.

Next, the quadrilateral area detecting / dividing unit will be described. 12A to 12E are diagrams showing an example of a method for detecting lattice points. , The image is divided into rectangular blocks each having a size of N pixels × M pixels (FIG. 12A). 12A, intra-block variances d _A , d _B , d _C and d _D are calculated for each of the four blocks A, B, C and D surrounding the lattice point P. , D1 = d _A + d _B + d _C + d _D.

A square area newly formed by moving the grid point P by one pixel vertically and horizontally (FIGS. 12B to 12E).
Similarly, D2, D3, D4 and D5 are calculated for. , D1 to D5, the minimum value is selected, and the grid point P is moved to the position where the minimum value is obtained. The above-mentioned processes 1 to 3 are performed once for each grid point. , The above is repeated for the entire image until the shape of the quadrangle does not change.

Next, the lattice point movement amount detector will be described. , The reference image already encoded and decoded and stored in the frame memory 1109 is input to the rectangular area detection / division unit 1112 and divided into rectangular areas, and as shown in FIGS. The grid points P _A , P _B , P _C , and P _{D of} are calculated. , C _A , C _B , C _C , and C _D at the same positions as P _A , P _B , P _C , and P _D on the encoding target frame are obtained.

, _A region of L _x pixels × L _y pixels centered on C _A is searched to find the best matching point. For example, the weighted weighted sum of pixel values in the search range on the encoding target frame is calculated, the pixel value having the weighted weighted sum closest to P _A is obtained, and this pixel is designated as C _A ′. , P _A to find the point C _A ′, then P _A to C _A ′
The amount of movement to (mvx _A , mvy _A ) is calculated. The same process is performed on C _B , C _C and C _D.

Next, the motion compensator will be described. The motion compensating unit 1110 calculates the predicted value of each pixel value included in the quadrangle region using the correspondence between the grid points of the reference frame and the frame to be coded, which is checked by the grid point movement amount detecting unit 1111. As shown in FIGS. 14A and 14B, the rectangular area is divided into two triangles, and the pixel position of the rectangular area on the reference frame is changed to the pixel position of the rectangular area on the encoding target frame by affine transformation. A predicted image on the target frame to be associated can be created.

Next, the moving picture decoding apparatus will be described.
FIG. 15 is a block diagram showing a configuration example of a moving picture decoding apparatus that decodes the encoded data created by the moving picture coding apparatus shown in FIG. 11. In the figure, 1501 is a variable length decoding unit, and 1 is a variable length decoding unit.
502 is an inverse quantization unit, 1503 is an inverse orthogonal transform unit, 150
4 is a frame memory, 1505 is a motion compensation unit, 1506.
Is a square area detection / division unit.

The variable length decoding unit 1501 is for decoding variable length encoded data. Inverse quantizer 1502
Is for dequantizing the quantized data decoded by the variable length decoding unit 1501. Inverse orthogonal transform section 1503
Is to perform inverse orthogonal transform on the transform coefficient dequantized by the dequantizer 1502. Frame memory 1504
Stores the image already decoded. The motion compensation unit 1505 is the same as the motion compensation unit 1110 in FIG. The rectangular area detecting / dividing unit 1506 is the same as the rectangular area detecting / dividing unit 1112 in FIG.

[0014]

In the lattice point movement amount detecting section of FIG. 11, the movement amount of each lattice point corresponding between frames is sent to the decoding side. As shown in FIG.
When divided into × K square areas, the grid points at the four corners are fixed, so the grid points with movement amount are indicated by ◯ (L +
1) × (K + 1) −4. Further, since there is information on the amount of movement only in the horizontal direction on the upper and lower boundary lines of the image and only in the vertical direction on the left and right boundary lines, the information of the grid points is actually 2 × (K × L-1) components. Had to send.
However, the information on the grid point movement amount is 2 × (K ×
Since all L-1) components were encoded, a lot of redundant information was included.

The present invention has been made in view of the above circumstances, and divides an image into a plurality of quadrilateral regions, and collectively encodes those whose vertices have a movement amount of "0". It is an object of the present invention to provide a moving picture coding apparatus and a moving picture decoding apparatus that efficiently code the moving amount of grid points in an arbitrary quadrangle region.

[0016]

In order to solve the above-mentioned problems, the present invention (1) divides an image into a plurality of quadrilateral regions, and creates quadrangle regions, the grid points of the quadrilateral are provided between frames. In the moving picture coding apparatus for coding the moving amount of the corresponding position of, the comparator that determines whether the moving amount of the grid point of each quadrangle is "0" and the moving amount is "0". In this case, according to the coding order of the grid points in the frame, a counter for counting the number of grid points followed by the movement amount “0”, the continuous number of grid points with the movement amount “0”, and the movement amount “ A variable-length coding unit that handles the movement amount of the grid point that is not 0 "as one code is provided,
Alternatively, (2) the image is divided into a plurality of quadrilateral regions or the quadrilateral regions into smaller blocks, and the amount of movement of corresponding positions of the grid points of the quadrangle is encoded between frames to create a predicted image. In a moving picture coding apparatus, a rectangular area or a rectangular area is subdivided from a counter that counts the number of transform coefficients “0” in a rectangular area or block that has been orthogonally transformed and quantized, and an output result from the counter. A comparator that determines whether all the conversion coefficients of the block are “0” and also determines whether the movement amount of the grid point of each quadrangle is “0” and each quadrangle region based on these information. For the quadrangle area that needs to determine the state of the quadrangle and the information on the movement amount of the grid point and the conversion coefficient, and after the code indicating the quadrangle state, the grid point transfer is performed. That the amount or transform coefficient information provided a variable length coding unit for encoding, and further,
(3) In (2), the grid point movement amount is “0”.
, And a counter that counts the number of consecutive quadrangles having the above state when all orthogonal transform coefficients in the quadrangle region are “0”, and a variable-length coding unit that encodes the output result of the counter. Furthermore, (4) In (1), (2) or (3) above, instead of the grid point movement amount of each of the quadrangular regions, processing that is adjacent to the grid point in the same frame has already been performed. Using the difference value between the moved amount of the grid point and the moved amount of the grid point, and (5)
In the above (1), (2) or (3), instead of the grid point movement amount of each quadrangle region, the grid point movement amount of the quadrangle region at the same position as the quadrangle region in the already encoded frame In order to use the difference value or (6) to decode the coded data created by the moving picture coding device according to claim 1, a grid point moving amount creating unit for creating a grid point moving amount is provided. Or (7) a grid point shift amount creating unit that creates a grid point shift amount in order to decode the coded data created by the moving picture coding device according to claim 2.
And a coefficient creating unit for creating all orthogonal transform coefficients in a rectangular area or a block obtained by subdividing the rectangular area, or (8) decoding coded data created by the moving picture coding apparatus according to claim 3. In order to do this, a grid point movement amount creation unit that creates a grid point movement amount, a coefficient creation unit that creates all orthogonal transform coefficients in a rectangular area or a subdivided block thereof, and a grid point movement amount and a rectangular area And a zero region reproducing unit for creating all orthogonal transform coefficients, and further,
(9) In (6), (7) or (8) above, instead of the grid point movement amount of each of the quadrangular regions, the movement amount of an already processed grid point adjacent to the grid point in the same frame is , Using a difference value from the grid point movement amount, and further,
(10) In (6), (7) or (8) above, instead of the grid point movement amount of each of the quadrangle areas, the grid point movement of the quadrangle area at the same position as the quadrangle area in the already encoded frame is moved. It is characterized by using a difference value from the quantity.

[0017]

According to the present invention, in a moving picture coding apparatus and a decoding apparatus which divides an image into a plurality of rectangular areas having uniform properties, all the grid point movement amounts are not coded but the grid points belonging thereto are coded. For a quadrangle region in which the movement amounts are all “0”, continuous lattice point movement amounts are collectively encoded, so that efficient encoding is possible. Further, in addition to the grid point movement amount of the quadrangle region, when all the conversion coefficients in the quadrangle region or a block obtained by subdividing this are “0”,
Since the conversion coefficient in this rectangular area or a block obtained by subdividing this rectangular area is encoded only by the flag indicating that, efficient encoding is possible.

Further, the grid point movement amount of the square area is "0".
In addition, when all the transform coefficients in the quadrangle region are “0”, the continuous quadrangle region is collectively encoded, so that efficient encoding is possible. Also, instead of the grid point movement amount of the rectangular area, a difference value between the movement amount of the already processed grid point adjacent to the grid point in the same frame and the grid point movement amount is calculated, and this is encoded. Therefore, efficient encoding is possible. Further, instead of the grid point movement amount of the quadrilateral area, the difference value between the grid point movement amount of the quadrangle area at the same position as the quadrangle area of the encoding target frame on the already encoded frame is calculated, and the difference value is calculated. Therefore, efficient coding is possible.

[0019]

Embodiments will be described below with reference to the drawings. FIG. 1 is a block diagram for explaining an embodiment (embodiment 1) of a moving picture coding apparatus according to the present invention.
Reference numeral 101 is a frame memory, 102 is a quadrilateral region detection / division unit, 103 is an orthogonal transformation unit, 104 is a quantization unit, and 105.
Is a variable length coding unit, 106 is a code buffer, 107 is an inverse quantization unit, 108 is an inverse orthogonal transform unit, 109 is a frame memory, 110 is a motion compensation unit, 111 is a lattice point movement amount detection unit, and 112 is a quadrangle region. A detector / divider, 113 is a comparator, 114 is a counter, and 115 is a switch.

A frame memory 101 for storing an input image
Can store a plurality of frames for calculating a motion vector. The rectangular area detecting / dividing units 102 and 112 detect the positions of the lattice points of the rectangular area and create position information,
The image is divided into rectangular areas surrounded by grid points as shown in FIG. The orthogonal transformation unit 103 performs orthogonal transformation on the pixels in the rectangular area created by the rectangular area detection / division unit 102 or 112. Note that the orthogonal transformation does not necessarily have to be performed in units of the rectangular areas of the rectangular area detecting / dividing units 102 and 112, and processing units may be independently set.

For example, the orthogonal transformation may be performed in units of square blocks. The quantization unit 104 quantizes the converted data with an appropriate quantization step size. The variable length coding unit 105 detects the quantized value or the square area /
The grid point position information from the dividing units 102 and 112 is subjected to variable length coding according to a predetermined code table and output as a transmission path code. Code buffer 106
Stores and smoothes the data in order to output the data from the variable length coding unit 105 at a constant rate.

The inverse quantization unit 107 inversely quantizes the output from the quantization unit 104. The inverse orthogonal transform unit 108 performs an inverse orthogonal transform on the output of the inverse quantization unit 107. The processing unit of the inverse orthogonal transform matches the processing unit of the orthogonal transformation.
The frame memory 109 stores images required for prediction. The motion compensating unit 110 includes a lattice point movement amount detecting unit 1 described later.
Motion compensation prediction is performed using the grid point movement amount that is the output of 11. The grid point movement amount detection unit 111 detects a square area /
Each grid point of the quadrangular area created by the dividing unit 112 is also moved to calculate the moving amount of the grid point when creating the predicted image approximate to the encoding target frame. Switch 115
The processing after the orthogonal transformation is switched between in-frame processing and inter-frame processing.

The difference between the present invention and the prior art is that the comparator 11
3 and the counter 114 and the variable length coding unit 105 are provided. In the comparator 113, the grid point movement amount (mv _x (i), mv _y (i)) of each square area (i is the serial number of the grid point,
i = 1 to (K + 1) × (L + 1) -4) is compared with (0,0), and if the grid point movement amount is (0,0), a flag of “0” is assigned to the grid point, and Otherwise, "1"
Assign the flag. In FIG. 2A, K = 5 and L = 5.
This is an example of the case. The counter 114 counts consecutive “0” s, and the variable length coding unit 115 outputs the number of consecutive “0s” output from the counter 114 and the subsequent grid point movement amount (mv _x (i), mv). _y (i)) is set to create a variable length code. In the example of FIG. 2A, a variable length code is created for each of the sets in Table 1 below.

[0024]

[Table 1]

Here, (end of frame) is a flag indicating that the amount of movement of the lattice points until the end of the image is (0,0). In this way, redundant information having a lattice point movement amount of "0" is not encoded, so that the number of encoded bits can be reduced.

FIG. 3 is a block diagram for explaining another embodiment (embodiment 2) of the moving picture coding apparatus according to the present invention, in which 301 is a frame memory and 302 is a quadrilateral area detecting / dividing unit. , 303 is an orthogonal transformation unit, 304 is a quantization unit, 305 is a variable length coding unit, 306 is a code buffer,
Reference numeral 307 is an inverse quantization unit, 308 is an inverse orthogonal transformation unit, 309 is a frame memory, 310 is a motion compensation unit, 311 is a lattice point movement amount detection unit, 312 is a quadrangular region detection / division unit, 31
3 is a counter, 314 is a comparator, and 315 is a switch.

FIG. 4 is a diagram showing an example in which an image is divided into quadrangular regions of arbitrary shape. From the quadrangle at the upper left, 1, 2, 3, ...
The serial number is attached to KL. The arrows in the figure show an example of a rectangular area to which each grid point belongs. In the example of FIG. 4, three grid points belong to the first square area,
Two lattice points belong to the square region of ~ k-1 or iL + 1 (i = 2 to L-1), the square region of KL has no grid points, and the other square regions are each one. Each has a grid point.

The difference between FIG. 3 and the conventional example is the counter 313.
The point is that the comparator 314 and the variable length coding unit 305 are provided. Further, the processing unit of the orthogonal transformation unit 303 is a quadrilateral region defined by the quadrilateral region detection / division unit 302, or one obtained by further dividing the quadrilateral region into small blocks as shown in FIGS. But it's okay. That is, the processing unit of the orthogonal transformation unit 303 may be each small block. Here, an example in which the rectangular area is divided into small blocks will be described with reference to FIG.

The counter 313 shown in FIG. 3 counts the quantized transform coefficients of each block being "0",
In the comparator 314, from the result of the counter 313, a flag of "0" is set for a block whose transform coefficients belonging to the block are all "0", and a flag of "1" is set for the other blocks. . At the same time, it is determined whether or not the grid point movement amount of each square area is (0, 0), and (0,
If it is 0), the flag "0" is set, and otherwise, the flag "1" is set. In this case, there are eight states in Table 2 below.

[0030]

[Table 2]

The variable length coding unit 305 has a code table for the above states and creates coded data indicating the state of the rectangular area. In addition, the grid point movement amount (mv _x (i), mv
_y (i)), or encode the transform coefficient of the block. If the flag is "0", no coding is performed. In this way,
Coded bits can be reduced by not coding redundant grid point shift amounts or transform coefficients.

FIG. 6 is a block diagram for explaining still another embodiment (third embodiment) of the moving picture coding apparatus according to the present invention. In the figure, 601 is a frame memory, and 602 is square area detection / division. Section, 603 is an orthogonal transformation section, 604 is a quantization section, 605 is a variable length coding section, 606 is a code buffer, 607 is an inverse quantization section, 608 is an inverse orthogonal transformation section, 60
Reference numeral 9 is a frame memory, 610 is a motion compensation unit, 611 is a lattice point movement amount detection unit, 612 is a rectangular area detection / division unit,
Reference numeral 613 is a counter, 614 is a comparator, 615 is a counter, and 616 is a switch.

FIG. 6 is provided with a counter 615 in addition to FIG. 3, in which the above-mentioned (grid point movement amount, B ₁ conversion coefficient, B ₂ conversion coefficient) flag is (0, 0, 0) quadrilateral area, that is, the number of consecutive quadrilaterals in which the grid point movement amount is (0,0) and all transform coefficients in the quadrilateral area are “0” is counted, and this value is counted by the variable length coding unit. 605
And the information in the rectangular area is not encoded at all. For other states, the same processing as in the case of FIG. 3 is performed. In this way, the coded bits can be reduced by not coding the redundant grid point shift amount and the transform coefficient.

As described above, with respect to Examples 1 to 3 described above,
The value of the grid point movement amount was used. As another method, the grid point movement amount of the square area currently being processed and the difference value between the movement amount of the grid point already processed adjacent to the grid point in the same frame and the grid point movement amount are implemented. It may be used as the grid point movement amount in Examples 1 to 3. In addition, as another method, the grid point movement amount of the rectangular area currently being processed,
The difference value from the grid point movement amount of the rectangular area at the same position on the already encoded frame may be used as the grid point movement amount in the first to third embodiments.

Next, the moving picture decoding apparatus will be described.
FIG. 7 is a block diagram for explaining an embodiment (embodiment 4) of a moving picture decoding apparatus for decoding coded data created by the moving picture coding apparatus of the embodiment shown in FIG. In the figure, 701 is a variable length decoding unit, 702 is an inverse quantization unit, 70
3 is an inverse orthogonal transformation unit, 704 is a grid point movement amount creation unit, 70
Reference numeral 5 is a frame memory, 706 is a motion compensation unit, and 707 is a quadrilateral region detection / division unit. The difference from the conventional example shown in FIG. 15 is that the coded data created by the moving picture coding apparatus of FIG. 1 is coded as a set of continuous “zero” and “non-zero” grid point movement amounts. Therefore, the grid point movement amount creation unit 704
Is the point at which the movement amount at each grid point is reconstructed and input to the motion compensation unit 706.

FIG. 8 is for explaining another embodiment (fifth embodiment) of the moving picture decoding apparatus for decoding the coded data produced by the moving picture coding apparatus of the embodiment shown in FIG. 801 is a variable length decoding unit, 802 is an inverse quantization unit, 803 is an inverse orthogonal transform unit, 804 is a lattice point movement amount creation unit, 805 is a frame memory, 806 is a motion compensation unit, and 807 Is a square area detection / division unit, and 808 is a coefficient creation unit.

In the coded data produced by the moving picture coding apparatus of FIG. 3, when the grid point movement amount is (0,0) and all the transform coefficients in the rectangular area or block are "0", ,
Since only the flag indicating this is included in the encoded data,
The coefficient creation unit 808 reconstructs the transform coefficient. For example,
If (grid point movement amount, B ₁ conversion coefficient, B ₂ conversion coefficient) is (0, 1, 0), the grid point movement amount (0, 0) and B ₂ conversion coefficient are created.

FIG. 9 illustrates still another embodiment (sixth embodiment) of the moving picture decoding apparatus for decoding the coded data produced by the moving picture coding apparatus of the embodiment shown in FIG. 902 is a variable length decoding unit, and 902 is a block diagram for
Is an inverse quantization unit, 903 is an inverse orthogonal transform unit, 904 is a lattice point movement amount creation unit, 905 is a frame memory, 906 is a motion compensation unit, 907 is a quadrilateral region detection / division unit, 908 is a coefficient creation unit, and 909 is It is a zero area reproducing unit.

In the coded data created by the moving picture coding apparatus of FIG. 6, information on the rectangular area when the grid point movement amount is (0,0) and the transformation coefficients in the rectangular area are all "0" ( In the zero area, one code is given by combining consecutive zero areas. Therefore, by adding the zero area reproducing unit 909 to the example of FIG. 8, continuous zero areas can be reconstructed.

As described above, in the fourth to sixth embodiments of the moving picture decoding apparatus, the value of the lattice point movement amount itself is used. As another method, the difference value between the movement amount of the lattice point that has already been processed adjacent to the lattice point in the same frame as the movement of the lattice point of the quadrangle region currently being processed and the movement amount of the lattice point is determined by moving image decoding. It may be used as the grid point movement amount in the fourth to sixth embodiments of the apparatus. In addition, as another method, the difference value between the grid point movement amount of the square area currently being processed and the grid point movement amount of the square area at the same position on the already-coded frame is calculated using the fourth to fourth embodiments.
It may be used as the amount of movement of the lattice points in 6.

As described above, the moving picture coding apparatus according to the present invention comprises means for checking the number of consecutive grid points having a movement amount of "0", and the following grid points having a movement amount of not "0". And a means for performing variable length coding with the movement amount set. In addition, means for determining grid points belonging to each quadrilateral area in advance and checking the number of consecutive quadrilaterals in which the movement amounts of the grid points included in the quadrangle area are all “0”,
A means for checking whether or not all the conversion coefficients in the rectangular area or a block obtained by subdividing the rectangular area are “0”, and these 2
Create a variable length code that represents the state of the rectangle from one piece of information,
Means for encoding only the grid point movement amount or transform coefficient which is not "0".

Further, the grid points belonging to each quadrangle region are determined in advance, and the moving amounts of the grid points included in the quadrangle region are all "0", and the transformation coefficients belonging to the quadrangle region are all "0". And a means for performing variable length coding by setting information on the amount of movement or transformation coefficient of grid points of a quadrangle other than the above which follows this quadrangle. Further, instead of the grid point movement amount of the quadrangle region, a means for calculating a difference value between the grid point movement amount and the movement amount of the already processed grid point adjacent to the grid point in the same frame is provided. Further, instead of the grid point movement amount of the quadrangle region, a unit for calculating a difference value with respect to the grid point movement amount of the quadrangle region at the same position as the quadrangle region of the encoding target frame on the already encoded frame is provided.

Since the moving picture decoding apparatus according to the present invention decodes the coded data created by the moving picture coding apparatus, since the decoding apparatus is a quadrangle having no grid point moving quantity, the moving amount of grid point is calculated. Provide a means to create. In addition, since the decoding device decodes the encoded data created by the moving image coding device, since the decoding device is a quadrangle having no lattice point movement amount or a transformation coefficient in a quadrangle region or block, the lattice point movement amount or transformation is performed. Providing means for creating coefficients. Further, in order to decode the coded data created by the moving picture coding device, since the decoding device is a quadrangle that does not have the lattice point movement amount and the transformation coefficient in the quadrangle region, the lattice point movement amount and the transformation coefficient are Provide a means to create.

Further, in order to decode the coded data created by the moving picture coding apparatus, instead of the grid point movement amount of the rectangular area, already processed grid points adjacent to the grid point in the same frame are processed. A means for calculating a difference value between the moving amount of the grid point and the moving amount of the grid point is provided. Further, in order to decode the coded data created by the moving picture coding device, instead of the grid point movement amount of the rectangular area, the same position of the rectangular area of the coding target frame on the already coded frame is detected. A means for calculating a difference value with respect to the grid point movement amount of the rectangular area is provided. As described above, regarding the coding of the grid point movement amount when the image is divided into the quadrangular regions of arbitrary shapes, efficient coding is performed by using the case where the grid point movement amount is (0,0). It can be carried out.

[0045]

As is apparent from the above description, the present invention has the following effects. (1) When the image is divided into quadrilateral regions of arbitrary shape and the movement amount of the lattice points is encoded, the movement amount is encoded for each lattice point for the lattice points having the movement amount of “0”. Instead, since consecutive grid points having a movement amount of “0” are collectively coded, efficient coding is possible as compared with the conventional method in which the movement amounts are coded for all the lattice points. is there. (2) In addition, by using the code indicating the state of the quadrilateral area, when the grid point movement amount of the quadrilateral area is "0" and all the transform coefficients in the quadrilateral area or block are "0", that Since only the flag indicating is encoded, it is possible to perform efficient encoding compared to the case where the lattice point movement amount and the transform coefficient in this rectangular area or block are all encoded in the related art. (3) Further, for a zero area in which the grid point movement amount of the square area is “0” and all the transformation coefficients in the square area are “0”, consecutive zero areas are collectively given one code. Therefore, efficient coding is possible as compared with the case where the grid point movement amount and the transform coefficient of this rectangular area are all coded conventionally. (4) Further, instead of the grid point movement amount of the rectangular area, a difference value between the movement amount of the already processed grid point adjacent to the grid point in the same frame and the grid point movement amount is calculated. Is encoded, it is possible to perform more efficient encoding than encoding the grid point movement amount itself. (5) Further, instead of the grid point movement amount of the quadrilateral region, a difference value with the grid point movement amount of the quadrangle region at the same position as the quadrangle region of the encoding target frame on the already encoded frame is calculated, Since this is encoded, more efficient encoding is possible than encoding the lattice point movement amount itself.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of a moving picture coding apparatus according to the present invention.

[Fig. 2] Fig. 2 is an explanatory diagram of an encoding method of a lattice point movement amount according to the present invention.

FIG. 3 is a block diagram for explaining another embodiment of the moving picture coding apparatus according to the present invention.

FIG. 4 is a diagram showing a relationship between a rectangular area and grid points in the present invention.

FIG. 5 is a diagram showing an example of block division according to the present invention.

FIG. 6 is a block diagram for explaining still another embodiment of the moving picture coding apparatus according to the present invention.

FIG. 7 is a block diagram for explaining an example of a moving image decoding apparatus according to the present invention.

FIG. 8 is a block diagram for explaining another embodiment of the moving image decoding apparatus according to the present invention.

FIG. 9 is a block diagram for explaining still another embodiment of the moving picture decoding apparatus according to the present invention.

FIG. 10 is a diagram showing an example of a modified grid.

FIG. 11 is a block diagram of a conventional moving image encoding device.

FIG. 12 is a diagram showing an example of a method of detecting lattice points.

FIG. 13 is an explanatory diagram of detecting a movement amount of a grid point.

FIG. 14 is a diagram showing an example of affine transformation.

FIG. 15 is a block diagram of a conventional moving image decoding apparatus.

[Explanation of symbols]

101, 109, 301, 309, 601, 609, 7
05, 805, 905, 1101, 1109, 1504
... Frame memory, 102, 112, 302, 312,
602, 612, 707, 807, 907, 1102
1112, 1506 ... Square area detecting / dividing unit, 10
3, 303, 603, 1103 ... Orthogonal transformation unit, 104,
304, 604, 1104 ... Quantizer, 105, 30
5, 605, 1105 ... Variable length coding unit, 106, 30
6,606,1106 ... Code buffer, 107,30
7,607,702,802,902,1107,15
02 ... Inverse quantizer, 108, 308, 608, 703
803, 903, 1108, 1503 ... Inverse orthogonal transform unit,
110, 310, 610, 706, 806, 906, 1
110, 1505 ... Motion compensation unit, 111, 311, 61
1, 1111 ... Lattice point movement amount detection unit, 113, 314
614 ... Comparator, 114, 313, 613, 615 ... Counter, 115, 315, 616, 1113 ... Switch, 701, 801, 901, 1501 ... Variable length decoding section, 704, 804, 904 ... Lattice point movement amount creating section , 8
08, 908 ... Coefficient creating unit, 909 ... Zero area reproducing unit.

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI technical display location H03M 7/40 9382-5K

Claims (10)

[Claims] 1. A moving picture coding apparatus which divides an image into a plurality of rectangular areas and codes a moving amount of corresponding positions of grid points of the rectangular between frames to create a predicted image. Of the grid point of “0” is determined according to a comparator that determines whether or not the grid point of “0” is moved, and when the move amount is “0”, the movement amount of “0” is determined according to the coding order of the grid points in the frame. A variable length that combines a counter that counts the number of subsequent grid points, a continuous number of grid points whose movement amount is "0", and a movement amount of grid points whose movement amount is not "0" into one code A moving picture coding apparatus, comprising: a coding section. 2. An image is divided into a plurality of quadrilateral regions or the quadrilateral regions into smaller blocks, and the amount of movement of corresponding positions of the grid points of the quadrangle is encoded between frames in order to create a prediction image. In a moving picture coding apparatus, a rectangular area or a rectangular area is subdivided from a counter that counts the number of transform coefficients “0” in a rectangular area or block that has been orthogonally transformed and quantized, and an output result from the counter. A comparator that determines whether all the conversion coefficients of the block are “0” and also determines whether the movement amount of the grid point of each quadrangle is “0” and each quadrangle region based on these information. The symbol indicating the state of the rectangle and the symbol indicating the state of the rectangle and the symbol indicating the state of the rectangle are moved only after the symbol indicating the state of the rectangle is required only for the rectangular area in which the information of the moving amount and the conversion coefficient of the lattice point is required. A moving picture coding apparatus, comprising: a variable length coding unit for coding quantity or transform coefficient information. 3. A counter that counts the number of consecutive quadrangles having the above state when the grid point movement amount is “0” and all orthogonal transform coefficients in the quadrangle region are “0”, 3. The moving picture coding apparatus according to claim 2, further comprising a variable length coding unit that codes the output result of the counter. 4. The difference value between the movement amount of the grid point that has already been processed adjacent to the lattice point in the same frame and the movement amount of the lattice point is used instead of the movement amount of the lattice point of each of the quadrangular regions. The moving picture coding device according to claim 1, 2 or 3. 5. A difference value between the grid point movement amount of each rectangular area and the grid point movement amount of a rectangular area at the same position as the rectangular area in an already-coded frame is used. The moving picture coding device according to claim 1, 2 or 3. 6. A moving image comprising a grid point movement amount creation unit for creating a grid point movement amount in order to decode the coded data created by the moving picture coding device according to claim 1. Image decoding device. 7. In order to decode the coded data created by the moving picture coding apparatus according to claim 2, a grid point moving amount creating unit for creating a grid point moving amount, and a quadrangle region or this are subdivided. A moving image decoding apparatus, comprising: a coefficient creating unit that creates all orthogonal transform coefficients in a block. 8. In order to decode the coded data created by the moving picture coding apparatus according to claim 3, a grid point moving amount creating unit for creating a grid point moving amount, and a quadrangle region or this are subdivided. A moving image decoding apparatus comprising: a coefficient creating unit that creates all orthogonal transform coefficients in a block; and a zero region reproducing unit that creates a lattice point movement amount and all orthogonal transform coefficients in a rectangular region. 9. The difference value between the movement amount of the grid point that has already been processed adjacent to the grid point in the same frame and the movement amount of the grid point is used instead of the movement amount of the grid point of each of the quadrangular areas. 9. The moving picture decoding device according to claim 6, 7 or 8. 10. A difference value between the grid point movement amount of each square area and the grid point movement amount of a square area at the same position as the square area in an already-coded frame is used instead of the grid point movement amount. The moving image decoding apparatus according to claim 6, 7, or 8.