JP3204824B2 - Moving image processing method and processing apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a processing method and apparatus suitable for recording, transmitting and reproducing moving images. More specifically, the present invention provides a processing method and a processing apparatus for encoding and restoring moving image information.
It is related to the location .

[0002]

2. Description of the Related Art Transmission and reproduction of moving images have been conventionally performed by television, but as a method of image encoding (compression), mainly band compression technology has been conventionally known. In recent years, with the advance of digital signal processing technology and digital communication technology, data compression by digital processing has been performed.

As is well known, a moving image is composed of a plurality of frames, for example, 30 frames per second.
When encoding a moving image, it can be classified into intra-frame encoding in which each frame is independently encoded and inter-frame encoding in which processing is performed on a plurality of frames. Furthermore, in the case of a television, since one frame is usually composed of two fields, intra-frame encoding is divided into intra-field encoding and inter-field encoding.

In the intra-frame coding and the intra-field coding, the coding technique of a still picture can be directly applied to a moving picture, but inter-frame coding is required for more effective compression. Conventionally, this technique includes (i) conditional pixel replenishment, (ii) motion compensation prediction and background prediction, and (iii) inter-frame adaptive prediction combining intra-frame motion correction prediction and intra-field prediction.

The conditional pixel replenishment method (i) is a method of extracting and transmitting only data of a moving part of an image. A difference between frames is obtained for each pixel, and the value is determined by a predetermined threshold value. This is a method of sending a larger pixel with an address specifying the pixel.

The motion compensation prediction of (ii) is for detecting the position and speed of the target object in the previous frame and predicting the position of the target object in the current frame. Since motion-compensated prediction has no effect on a background image that appears after the object moves, only a background image separated from the moving object is recorded by having a second frame memory in the encoder, This is used for the background image.

The inter-frame adaptive prediction (iii) is a combination of inter-frame motion compensation prediction and intra-field prediction, and can be applied to severe motion.

[0008]

However, each of the above-mentioned inter-frame codings has its own problems.

First, the conditional pixel replenishment method (i) has a drawback that the amount of coded data increases rapidly with an increase in the image motion.

In the motion compensation prediction of the above (ii), there are restrictions on the accuracy of motion detection and the amount of motion that can be detected. further,
Background prediction has the disadvantage of having to have two frame memories.

In the above (iii) inter-frame adaptive prediction,
There is a problem that the configuration is complicated as compared with the above (i) and (ii).

In recent years, moving image recording / transmission /
Compression needs are not limited to television, but also
It has spread to a wide range of media, including video conferencing systems and computer graphics.

In view of the above, a first object of the present invention is a moving image processing method and processing apparatus which can be applied to a wide range of media in common, has a simple configuration, and has both high compression and high fidelity. Is to provide.

A second object of the present invention is to provide a processing method and a processing apparatus capable of recording, transmitting and reproducing even a severe movement.

Further, a third object of the present invention is to provide a processing method and apparatus for efficiently recording / transmitting / reproducing a binary image, a grayscale image, a monochrome image and a color image.
And a processing device .

[0016]

In order to achieve the above object, the present invention uses the following methods (A) and (B) for binary and multivalued images, respectively.

(A) Binary image: At the time of recording or transmitting a moving image, a set of points where the image data of the moving image represented by the binary data is 1 is represented by:
Expressed as one or a plurality of solids in a three-dimensional space having xy orthogonal coordinates as an image plane and a time axis orthogonal to each of the coordinate axes, with the image plane as the bottom surface and the height from the image start time to the end time. Is divided into a plurality of rectangular parallelepipeds, and each of the divided rectangular parallelepipeds
The ratio of the inner volume of the stereo record or transmitted as Rv value, at the time of reproduction of the moving image, Rv value of the target rectangular Contact
From the Rv value of the rectangular parallelepiped adjacent to the rectangular parallelepiped
Calculate binary image data of the image plane at each time of the cube
Thus, a moving image represented by binary data is reproduced.

(B) Multi-valued image: At the time of recording or transmitting a moving image, a moving image represented by multi-valued data has an xy orthogonal coordinate which is an image plane, a time axis, and an axis representing a density level. represented by one or more four-dimensional region within the four-dimensional Euclidean space, and dividing the four-dimensional space into a plurality of four-dimensional cuboid, divided respectively
Of the lower volume of the four-dimensional region in the four-dimensional rectangular solid of FIG.
The rate was recorded or transmitted as Rv value, at the time of reproduction of the moving image, attention four-dimensional rectangular Rv value and interest 4D
From the Rv value of the four-dimensional rectangular parallelepiped adjacent to the rectangular parallelepiped,
Multi-level image data of the image plane at each time of the original cuboid
To reproduce the moving image represented by the multi-value data .

[0019]

According to the above-mentioned method of the present invention, the image plane, the gray level and the time are treated without discrimination, and the sudden change of the contour and the gray level of the shape and the sudden appearance and disappearance of the object in the image can be achieved. Appropriate measures can be taken.

[0020]

Embodiments of the present invention will be described below in detail with reference to the drawings.

Embodiment 1 FIGS. 1 to 5 are drawings showing a first embodiment of the present invention.

FIG. 1 shows a typical state of a binary moving image at several times. At a time of 0 seconds, the two circles separated from each other approach each other while accelerating, and after a head-on collision, they merge into one circle, then suddenly disappear at 1.9 seconds, and at 2.0 seconds, The image ends.

FIGS. 2 and 3 conceptually show a method of coding an image. The xy plane is the image plane, x
A time axis (t axis) perpendicular to both the axis and the y axis is taken. In FIG. 2, reference numeral 1 denotes a three-dimensional surface formed by connecting the shapes of the original image in the xyt space. Next, the pixel area (648 × 488 pixels) in the xy plane is equally divided into 8 × 8 pixel cells.

[0024]

[Outside 1]

Next, ten rectangular parallelepipeds are stacked on each cell (in the t-axis direction) (not shown). Therefore, the number of cuboids is 49410. The height of the rectangular parallelepiped is equivalent to 0.2 seconds (six frames).

FIG. 3 is an enlarged view of the vicinity of the point P shown in FIG. 2. Reference numeral 2 denotes one of the rectangular parallelepipeds.

For each of the above rectangular solids, the ratio R of the volume inside the surface 1 in the rectangular solid to the volume of the rectangular solid is calculated and recorded. Specifically, the total value of the pixel data (0 or 1) in the cell corresponding to the rectangular parallelepiped is further summed over six frames, and this value R _V is recorded or transmitted.

FIG. 4 shows an apparatus for performing this encoding process. First, a clock signal corresponding to a pixel is generated by a clock generator 11, and binarization is performed by a binarizer 12 in synchronization with the clock signal. At this time, a predetermined appropriate threshold is given.

The binarized signal is sent to the counter 14 and is counted only when the value is 1. Counter 1
4 has only 81 cells in the horizontal direction (14-
1, 14-2, ..., 14-81). The counter is switched every eight pixels, and the next counter 14-81 returns to the first counter 14-1. When this cycle is repeated eight times, the values of the 81 counters are sent to the memory.

In the eighth cycle, the data is transferred to the memory in the order in which the data has been sent to the counter 14. That is, when the data from the binarizer 12 is sent to the counter 14-1, the data of the counter 14-1 is transferred to the memory while the data from the binarizer 12 is transferred to the counter 14-2. I do. Then, the memory group is switched every six frames.

This is because data processing and transmission for the previous six frames are performed while data input and storage processing for the current six frames are being performed.

Since the value stored in the counter 14 is 0 to 64, each memory group is designed to be able to simultaneously access 8 bits. Since the recording capacity is the number of cells × the number of frames / 1 rectangular parallelepiped, it is 4941 × 6 = 29646 bytes. In this embodiment, 64 kilobit RAM is used for a total of 8
16 pieces × 2 groups = 16 pieces.

Next, the data stored in the memory is added for all the frames for each cell and sent to the transmission line. For example, addresses 1, 1 + 4941, 1 + 4941 × 2
1 + 4941 × 3, 1 + 4941 × 4, 1 + 4941 ×
The six sets of data at address 5 are called up bit by bit and summed by adder 17. Since this value is 0 to 384, it is 9 bits / rectangular solid. In the case of this embodiment, 9 bits × 4941 = 444 during 6 frames (0.2 seconds).
What is necessary is just to be able to transmit 69 bits. That is, about 220 Kbit / sec.

In order to use the transmission path more effectively,
By dividing 9-bit data by about 255 or 127, it is possible to reduce the data to about 8 or 7 bits. Further, by using a buffer between the adder and the transmission path output and a communication path coding such as a Huffman encoder, the transmission path can be more effectively utilized.

[0035] FIG. 5 is a block diagram of a device for reproducing an image from data of the transmitted R _V. The process starts when the first two frame groups (one frame group = 6 frames) are transmitted to the buffer memory 31 from the transmission path. That is, the previous frame group, the current frame group, and the next frame group are used. The main purpose of this reproducing apparatus is to reproduce the three-dimensional surface (1 in FIG. 2) of the original image.

For this purpose, first, a normal vector of the plane is obtained. This is the part of the "calculation of f _x, f _y, f _z", calculated by the following equation.

[0037]

(Equation 1)

Here, k of R _V (i, j, k) indicates a frame group number, and i, j indicates a cell position. Also, C is a difference operator, and most simply, when described in a matrix format,

[0039]

(Equation 2)

However, in order to cope with an abrupt change in image space and time,

[0041]

(Equation 3)

Is desirable. Since the operation of the above equation can be performed by bit shift and addition / subtraction, it is performed by a random logic circuit (not shown). When k is the first frame group, that is, when k = 1, R _V (i, j, k−1) = R _V (i,
j, 0) is regarded as 0 for all i, j. When k is the last frame group, for all i and j, R
_{It is} assumed that _V (i, j, k + 1) = 0. Also, i, for even j, when (i ± 1, j) or (i, j ± 1) indicates a cell location that does not exist, the processing at the R _V value zero.

When the normal vector of the plane is determined in this way, the position of the plane is uniquely determined by calculation from the R _V value of the rectangular parallelepiped. However, since this operation includes a square root and division, it is difficult to make a random logic, and even if this operation can be performed, the amount of operation is large, so that the delay in image reproduction becomes significant.

Therefore, in this embodiment, the ROM ₁ and the ROM ₂
F _x, f _y, the value of f _z theta, seeking phi, theta by,
The position (x ₀ , y ₀ , t) of the point through which the plane passes is obtained from φ, R _V (θ and φ are shown in FIG. 5).

In the image data developing unit 35, 384 points (x, y, t) (1 ≦ x ≦ 8, 1 ≦ y ≦ 8, 1)
≦ t ≦ 6)

[0046]

[Number 4] _{(xx 0) · f x +} (yy 0) · f y + (tt 0) · f z ≧ 0 ... (4) if the pixels in the frame t x, and 1 the value of y.
Otherwise, it is set to 0 and stored in the image memory. When the storage is completed for all the rectangular parallelepipeds, the data is converted into a TV signal by a D / A converter, and a synchronizing signal is added and output.

As described above, in the present embodiment, the cell size (Δx = Δy = 8 pixels) and the size of the frame group (6 frames) do not specify the spatial and temporal resolution, but the cell size. The major feature is that it can cope with the contour line of the object shape in the middle and the disappearance / generation of the image in the middle of the frame group.

However, it is not possible to cope with a periodic change of the shape in the same cell (for example, a line passing through the cell).
In such a case, a method of separately reducing the cell size locally may be used together.

In the above description, an image transmission and reproduction method using a random logic circuit has been described. However, the present invention is not limited to this, and may be executed using a microprocessor. In particular, when real-time performance is required for image reproduction, that is, when a moving image is stored in a database as a database, the moving image is compressed and recorded in an external storage device of the computer.
For reproduction at a later date, a method using a microprocessor is effective.

In recent years, since particularly high-speed microprocessors have been developed, the ROM ₁ to ROM shown in FIG.
X ₀ , y ₀ , t ₀ by direct calculation instead of using ₅
You may ask.

In the method described above with reference to the present embodiment, the density distribution curved surface is approximated by a plane in a rectangular parallelepiped, and the contour of the image is also approximated by a straight line. In order to reproduce the density change of the original image more faithfully, it is necessary to reproduce the second-order differential information of the density and a method of reproducing the contour of the image by a curve.

For example, in addition to the aforementioned f _x , f _y , f _z , 2
Second derivative

[0053]

[Number 5] _{^{∂ 2 R V / ∂x 2 ≡f}} xx, ∂ 2 R V / ∂y 2 ≡f yy, ∂ 2 R V / ∂z 2 ≡f zz, ∂ 2 R V / ∂x∂y≡ f _xy , ∂ ² R _V / ∂y∂z f _yz , ∂ ² R _V / ∂z∂x≡f _zx (5) are obtained by a similar method, and a rectangular parallelepiped is obtained from the first- and second-order differential coefficients. ∂z / ∂x, ∂z / ∂y, ∂ ² z at the center
/ ∂x ² , ∂ ² z / ∂x∂y, ∂ ² z / ∂y ² , f (x, y, z) = const. (Z = z (x,
y)) is obtained by differentiating x and y. In particular,

[0054]

(Equation 6)

Is used.

[0056] _{If, | f z | «| f} x | and | f _z |«
If | f _y |, the density surface crossing the rectangular parallelepiped is almost parallel to the z-axis, that is, the density changes stepwise in the cell.
Instead of obtaining the first derivative, the first and second derivatives of x with respect to y and z are obtained, or the first and second derivatives of y with respect to z and x are obtained.

Here, if | f _y | ≦ | f _x |, the former is selected, and if | f _x | ≦ | f _y |, the latter is selected. In this way, the first and second derivatives of the curved surface at the center of the rectangular parallelepiped are obtained. From these derivatives, the curved surface is reproduced in the rectangular solid.

[0058] In this way, the proportion lower volume of curved surface obtained is greater than this rectangular parallelepiped R _V (if the order of 5% or more large), the curve _{(f x, f y, f} z) direction to then translate slightly, while smaller when (if the order of 5% or more small) is _{(-f x, -f y, -f} z) by only translated, may be brought close to the actual R _V.

However, when the R _V value of the rectangular parallelepiped of interest is extremely small (about 10 ⁻² or less), the curved surface may not be reproduced with high accuracy even by the above-described method. In such a case, f _x, f _y, prior to calculating the f _z, etc. R _V
It is good to apply a diffusion process to the value. In particular,

[0060]

(Equation 7)

[0061]

(Equation 8)

[0062] Here, D _I is the diffusion coefficient, consider the stability take 1/6 or less.

By repeating this operation several times, R _V
Can reproduce a curved surface with good reproducibility even in the case of a small rectangular parallelepiped.

According to the above-described embodiment, the recording information amount naturally increases in an area where the density change is sharper, without being conscious of the difference in the gradient of the density distribution of the original image. Preservation and high fidelity reproduction are possible.

Embodiment 2 FIG. 6 is a drawing showing a second embodiment of the present invention. The number of pixels of the original image in the present embodiment is the same as that of the first embodiment, but is applied to a moving image of light and shade.

In this embodiment, first, the image plane is defined as x, y,
When time is t and luminance is D (x, y, t), x,
Δx × Δ in four-dimensional Euclidean space of y, z, D
The volume under the luminance distribution curved surface D ( x , y, t) in the four-dimensional rectangular parallelepiped space having a volume of y × Δt × ΔD is recorded or transmitted.

As in the first embodiment, Δx = 8 pixels, Δy =
It is assumed that 8 pixels and Δt = 6 frames (one frame group). Further, the luminance data is represented by 8 bits from 0 to 255, and Δ
D = 16. Therefore, there are 16 times data in the luminance direction.

The transmission process is basically the same as that of the first embodiment, but the differences are as follows.

First, the number of adders is 81 × 16 instead of 81 counters. Each set includes a counter corresponding to 16 levels of luminance from the first to the 16th. Then, when counting, assuming that the upper 4 bits of the luminance data of the pixel are a and the lower 4 bits are b, the first to a-th adders add 16 respectively, and the value b is added to the a + 1-th adder. (However, when a = 0,
Only the value b is added to the first counter).

The processing by the last adder 40 is also performed for each density level. Therefore, the processing requires 16 times.

After the processing in the adder 40, the R _V values 0 to 61
Work of compressing the data of No. 44 to 0 to 127 (divider 41
To reduce the data to 7 bits / cuboid.

Immediately before output to the transmission path, redundancy removal is performed. Specifically, in a cell i, j and a frame group f,

[0073]

R _V (i, j, k, f) = 0 for all k such that k> k ₂ R _V (i, j, k, f) = 127 for all k such that k <k ₁ ( 9) k ₁ , k ₂ and R _V (i, j, k, f) (k ₁ ≦ k ≦
k ₂ ) only. Further, well-known run-length compression is also combined. By performing the redundancy removal, the amount of data to the transmission path is reduced from 2.77 Mbit / sec to about 0.3 Mbit / sec to about 1.2 Mbit / sec.

Further, it goes without saying that the present method can be combined with a well-known resolution exchange control.

Next, reproduction of image data will be described with reference to FIG. In the reproduction of image data, first, the data from which the redundancy has been removed is restored (51) and stored in the buffer memory. Next, R _V data is called from the buffer memory, and the R _V ^* value at each vertex of the four-dimensional rectangular parallelepiped is determined as an average value of R _V of 16 rectangular parallelepipeds connected to the vertex. Specifically, it can be obtained by addition and bit right shift.

Using these 16 R _V ^* , for example, the R _V value in the rectangular parallelepiped can be interpolated as follows.

[0077] That is, rectangular arbitrary pixel x, y and luminance level D, the R _V value in the frame f, and cuboid region {(x, y, D, f) | x i ≦ x ≦ x i + Δx∧y _{j
≦ y ≦ y j + Δy∧D k} ≦ D ≦ D k + Δk∧f m ≦ f ≦
f _m + Δf｝,

[0078]

(Equation 10)

Can be calculated. here,

[0080]

[Equation 11]

Is defined as This calculation can also be performed by addition / subtraction and bit shift (53 in FIG. 7).

Therefore, pixels x and y, density level D,
Given a frame f, an R _V value is obtained. If this value is equal to or greater than a predetermined value, an operation of adding 1 to the image data φ (x, y, f) may be performed on the image memory. In addition, about 64 is usually used as a predetermined value.

In this embodiment, as in the previous embodiment, any spatial resolution, luminance resolution, and time resolution are defined for the cell sizes Δx and Δy, the luminance level width ΔD, and the size of the frame group (6 frames). is not.

In this embodiment, processing using a microprocessor is also possible. In this case, it is also possible to use an Overhauser cubic polynomial, a Lagrangian polynomial, a spline function, or the like in addition to the above as the interpolation function.

Embodiment 3 FIG. 8 is a drawing showing a third embodiment of the present invention. INDUSTRIAL APPLICABILITY The present invention can be applied to not only transmission and reproduction of TV signals but also storage and output of computer graphics and calculation results of various CAEs.

For example, the result of numerical calculation of the movement of a celestial body performed by a large computer is first sent to a workstation through a network (Ethernet cable). At the workstation, the moving image data is visualized by a so-called post processor (software: not shown), and the moving image data is accumulated on the disk by executing a program having the same function as in the second embodiment.

At a later date, this data is stored in a computer program (FIGS. 9 and 10) having the same function as the image reproduction of the second embodiment.
), Sent to the frame memory, and recorded on the VTR as needed. It should be noted that the visualization can be performed on a graphic display simultaneously with the development of the image.

The recording of an image according to the present invention has an excellent compression ratio, and thus can be stored in an external storage device of a computer.

Embodiment 4 The present invention is applicable to a color image in addition to a monochrome image. In this case, processing may be performed independently for R, G, and B, or another separation method, for example, processing using signals of Y (luminance) I and Q (color difference) may be used.

[0090]

As described above, according to the present invention, the present invention can be applied to a wide range of media, for example, TV signals, HDTVs, personal computers, etc., and is simple for recording, transmitting, and reproducing moving images. A processing method can be realized. In particular, the present invention has an excellent compression ratio and fidelity, and can be similarly applied to an image with a violent motion, or to a binary image, a gray image, a monochrome image or a color image. .

[Brief description of the drawings]

FIG. 1 is a diagram showing a typical example of a binary moving image.

FIG. 2 is a diagram conceptually illustrating an image encoding method.

FIG. 3 is an enlarged view around a point P shown in FIG. 2;

FIG. 4 is a block diagram illustrating an apparatus that performs an encoding process according to the first embodiment of the present invention.

FIG. 5 is a block diagram showing an apparatus for reproducing an image according to the first embodiment of the present invention.

FIG. 6 is a block diagram illustrating an apparatus that performs an encoding process according to a second embodiment of the present invention.

FIG. 7 is a block diagram showing an apparatus for reproducing an image in a second embodiment of the present invention.

FIG. 8 is a block diagram showing a third embodiment of the present invention.

FIG. 9 is a flowchart illustrating an image recording program according to a third embodiment.

FIG. 10 is a flowchart illustrating an image reproduction program according to a third embodiment.

[Explanation of symbols]

1 In a xyt space, a three-dimensional surface connecting the shapes of the original 2 One of the rectangular parallelepipeds

Continuation of the front page (56) References JP-A-62-266974 (JP, A) JP-A-6-343130 (JP, A) IEEE 8th International Conference on Video, Audio and Data Recording, 1990, p. 144 -150 IEEE Journal on Selected Areas in Communication, Vol. 1, 1992, p. 97-121 (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68 H04N 5/91-5/956 H04N 1/41-1/419 G06T 9/00

Claims (13)

(57) [Claims] At the time of recording or transmitting a moving image, a set of points where the image data of the moving image represented by binary data is 1 is assigned to each of xy orthogonal coordinates representing an image plane and each of the coordinate axes. Expressed as one or a plurality of solids in a three-dimensional space having an orthogonal time axis, dividing an area ranging from the image start time to the end time height into a plurality of rectangular parallelepipeds with the image plane as a bottom surface, The ratio of the volume inside the solid in each of the divided rectangular solids is recorded or transmitted as an Rv value. When a moving image is reproduced, the Rv value of the rectangular solid of interest and the Rv value of the rectangular solid adjacent to the rectangular solid of interest are recorded. From the above, the normal vector of the three-dimensional curved surface in the cuboid of interest is calculated, and the position of the curved surface is calculated from the calculated normal vector and the Rv value of the cuboid of interest, thereby obtaining a curved surface that traverses the cuboid of interest. Calculating and determining whether each pixel at each time in the rectangular parallelepiped is inside or outside the calculated curved surface, and determining the value of pixel data of each pixel based on the determination result, A moving image processing method characterized by reproducing a moving image represented by binary data by repeating a rectangular parallelepiped. 2. When recording or transmitting a moving image, a four-dimensional Euclidean space having xy orthogonal coordinates representing an image plane composed of a plurality of pixels, a time axis and an axis representing a gray level is divided into a plurality of four-dimensional rectangular parallelepipeds. Then, the moving image represented by the multi-value data is divided into four divided images.
The volume ratio on the lower side of the distribution surface representing the gray level in the three-dimensional rectangular parallelepiped is recorded or transmitted as an Rv value. When a moving image is reproduced, the image data is initialized, and the Rv value of the four-dimensional rectangular parallelepiped and the attention Using the Rv values of the four-dimensional rectangular parallelepiped adjacent to the four-dimensional rectangular parallelepiped, calculate the Rv value at each pixel, each shading level, and each time in the rectangular parallelepiped, and update the image data according to the calculated Rv values. A moving image represented by multi-valued data by repeating the process for each rectangular parallelepiped. 3. The method according to claim 1, wherein, when recording or transmitting an image, the Rv value is compressed and then recorded or transmitted, and before the image is reproduced, the compressed Rv value is expanded. A moving image processing method. 4. The moving image processing method according to claim 1, wherein a part of the rectangular parallelepiped is subdivided. 5. The moving image processing method according to claim 1, wherein a diffusion process is performed on the Rv value when the image is reproduced. 6. The moving image processing method according to claim 1, wherein the gray level of the image is represented by luminance information of the image. 7. A moving image processing method according to claim 1, wherein the gray level of the image is represented by image density information. 8. A moving image processing method according to claim 1, wherein the gray level of the image is represented by luminance or density information of each of the three primary colors of the color image. 9. The moving image processing method according to claim 1, wherein the gray level of the image is represented by luminance and color difference information of the color image. 10. A set of points where the image data of a moving image represented by binary data is 1 is defined as an xy orthogonal coordinate representing an image plane and a three-dimensional space having a time axis orthogonal to each of the coordinate axes. Expressed as one or more solids, with the image plane as the bottom surface, divide the area ranging from the image start time to the end time height into a plurality of rectangular parallelepipeds, and A moving image processing method characterized by recording or transmitting a ratio of a volume inside a solid as an Rv value. 11. A set of points where the image data of a moving image represented by binary data is 1 is defined as an xy rectangular coordinate representing an image plane and a three-dimensional space having a time axis orthogonal to each of the coordinate axes. Expressed as one or more solids, with the image plane as the bottom surface, divide the area ranging from the time of the image start to the height of the end time into a plurality of rectangular parallelepipeds, and A storage unit for storing a ratio of the volume inside the solid as an Rv value; and a method of calculating the curved surface of the solid in the target cuboid from the Rv value of the cuboid of interest and the Rv value of the cuboid adjacent to the cuboid stored in the storage unit. Calculate the line vector, calculate the position of the curved surface from the calculated normal vector and the Rv value of the cuboid of interest to calculate the curved surface that crosses the cuboid of interest, and at each time in the cuboid of interest By determining whether a pixel is inside or outside the calculated curved surface and determining the value of pixel data of each pixel based on the determination result, it is represented by binary data by repeating for each rectangular parallelepiped. And a reproducing unit for reproducing a moving image to be processed. 12. An xy representing an image plane composed of a plurality of pixels.
A four-dimensional Euclidean space having rectangular coordinates, a time axis, and an axis representing a gray level is divided into a plurality of four-dimensional rectangular parallelepipeds, and a moving image represented by multi-value data is converted into a gray level in each of the divided four-dimensional rectangular solids. Recording or transmitting the ratio of the volume under the distribution surface representing Rv as an Rv value. 13. An xy representing an image plane composed of a plurality of pixels.
A four-dimensional Euclidean space having rectangular coordinates, a time axis, and an axis representing a gray level is divided into a plurality of four-dimensional rectangular parallelepipeds, and a moving image represented by multi-value data is converted into a gray level in each of the divided four-dimensional rectangular solids. Storage means for storing as a Rv value the ratio of the lower volume of the distribution surface representing Rv, and Rv values of the target four-dimensional cuboid stored in the storage means and four adjacent to the target four-dimensional cuboid. Using the Rv value of the three-dimensional cuboid, calculating the Rv value at each pixel, each gray level, and each time in the cuboid of interest, and updating the image data according to the calculated Rv value is repeated for each cuboid. And a reproducing means for reproducing a moving image represented by multi-valued data.