JPH1141625A - Color moving image data compression method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To relieve a processing load for color compression processing of color moving image data and to improve the data compression efficiency. SOLUTION: A row of frames 2 configuring color moving image data are divided into a frame group 20 consisting of a prescribed frame number. Color compression processing is applied to head frame data 22 of the frame group 20 to generate color palette information 24. Each frame in the frame group 20 is subject to palette processing based on the color palette information 24 generated from the head frame. The color palette information 24 is stored only in the head frame data 22. In the case of reproducing a moving image, the color palette information 24 extracted from the head frame data 22 is set in a memory and decoding processing to RGB values is conducted based on a palette index of succeeding frames in the frame group while referring to the color palette information 24.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression method for color moving image data, and more particularly to an improvement in the efficiency of data compression processing by color compression (palletizing processing) for limiting the number of colors included in an image.

[0002]

2. Description of the Related Art As a method of expressing color moving image data,
An expression format (referred to as a YUV format) using a luminance value (Y) and two color difference signals (U, V), an RGB (Red / Green / Blue) format using three primary colors, and a CMY (Cyan / Magen
ta / Yellow) format.

Conventionally, YU has been used for transmitting color moving image data.
The V format is often used. For example, moving image data is compressed with respect to the YUV format data by a technique such as MPEG compression, and data transmission and storage are performed by the compressed YUV format data.

On the other hand, an input unit of a system that handles color moving images, for example, an output signal of an image sensor, and an output unit of the system, for example, an output device such as a display device for displaying and printing, use an RGB format or a CMY format. Often. Therefore, data of different formats such as RGB format / CMY format and YUV format are often used in one moving image system.

[0005] Regarding the number of colors constituting an image, for example, in the RGB format, if each of R, G, and B is given an 8-bit gradation, about 16.7 million colors can be expressed. However, the data amount increases according to the number of colors that can be expressed, and the load of processing such as storage, display, and transmission increases.

[0006] Some of the constituent devices of the system have a limited number of colors that can be handled. For example, the number of colors that a color display device can handle simultaneously is often based on 256 colors.

For this reason, color compression may be performed on RGB and CMY format data by performing color reduction processing for limiting the number of colors. According to this color compression, a table is prepared that represents the correspondence between the RGB values of the color selected by limiting the number of colors and information such as codes for identifying the RGB values, and an image is configured using this table. The pixel value of each pixel is R
It is converted from the GB value to an identification code. For example, if the number of colors is 2
When the number of colors is limited to 56, the identification code can be represented by 8-bit binary, and the data amount is compressed. The table used for this conversion is called a color palette, and the color reduction processing is also called palletizing.

According to the conventional color compression, for example, the distribution area of the pixels constituting the frame in the three-dimensional RGB space defined by the three axes of R, G, and B is equal to the number of limited colors. , For example, into 256 blocks. Then, the identification number of the same block is given to the pixels included in the same block. For example, a color corresponding to the center point of the corresponding block is associated with each identification number by a color palette. That is, the pixels belonging to the same block are all displayed in the same color. Although accurate colors are not reproduced by this process, if the number of colors is set to a certain number, for example, about 256 colors, naturalness of a color image is not so much lost due to human visual characteristics.

By this processing, the amount of data is increased by the amount of the information of the color palette added to each frame data. However, since the data amount is reduced as described above by replacing the color data of each pixel with the block identification number, the data amount of each frame is generally compressed as a whole.

[0010]

Since the conventional color compression processing is performed on a frame basis, there is a problem that the processing time becomes long. That is, there is a problem in that the processing load of the processing device that performs color compression increases and the processing speed decreases.

There is another problem that color pallet information is added to each frame data and the data compression ratio is reduced.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a color moving image data compression method by color compression with a reduced processing load and an improved data compression ratio.

[0013]

In a first color moving image data compression method according to the present invention, a sequence of frames constituting a moving image is divided into a plurality of frame groups, and based on a predetermined color compression procedure. 1 for each frame group
One of the color pallet information is generated, and each of the frames included in each of the frame groups is indexed based on the color pallet information corresponding to the frame group.

A second color moving picture data compression method according to the present invention is applied to accumulated moving picture data, wherein the color compression is performed on the moving picture data included in each frame group. The color pallet information corresponding to the frame group is generated.

[0015] In a preferred aspect of the present invention, the color compression is performed on one virtual image obtained by two-dimensionally arranging a plurality of frames included in the frame group, and Is generated.

A third color moving image data compression method according to the present invention is applied to moving image data obtained in real time, and the color compression is performed on the first frame included in each frame group. This is performed as an object, and the color pallet information corresponding to the frame group is generated.

A fourth color moving image data compression method according to the present invention is applied to moving image data obtained in real time, and corresponds to each index based on the color pallet information. Determining an assigned color data range that includes the selected color data, and selecting the assigned color data range that includes the color data of each of the pixels; By the corresponding index,
Replacing the color data of the pixel.

In a preferred aspect of the present invention, the assigned color data range is a numerical value having the same number of predetermined upper digits as the selected color data associated with the assigned color data range via the index. The upper and lower limits are defined based on the maximum value and the minimum value.

In a fifth color moving picture data compression method according to the present invention, in the first and second methods, the color pallet information of each of the frame groups corresponds to one of the frames included in the frame group. That is, it is built in only the frame data obtained.

[0020] In a sixth color moving image data compression method according to the present invention, in the first to fourth methods, the color pallet information of each frame group is stored in the first frame included in the frame group. It is built into the corresponding frame data.

[0021]

Next, embodiments of the present invention will be described with reference to the drawings.

[First Embodiment] FIG. 1 is a schematic diagram for explaining a color compression method according to a first embodiment of the present invention applied to accumulated color moving image data.

The stored color moving image data processed here is moving image data already stored in a storage device such as a magnetic disk. That is, a frame sequence 2 which is image data at each time to be subjected to color compression processing
Is basically obtained at the start of the process. The color compression processing is basically performed by a central processing unit (CPU: Ce
ntral Processing Unit). In brief, for example, a small general-purpose computer or the like can be used, and a processing device dedicated to color compression can be configured to achieve high-speed processing.

In the processing of this embodiment, such a processing device reads out m and n pieces of frame data in order from the beginning of a frame sequence 2 which is a sequence of frame data stored in the storage device in time order, for example. I do. That is, the frame sequence 2 is divided into a plurality of frame groups each including m and n frames, and is processed for each frame group. m and n may be default values set in advance in the processing device, or may be parameters that can be set and changed in the processing device by an operator.

Alternatively, a plurality of frame groups 3 each including mn frames continuous in the time axis direction are formed. The m · n frames included in each frame group 3 are treated as one virtual image 4 in which m frames are arranged in the horizontal direction and n frames are arranged in the vertical direction. The figure shows an image 4 when m = 3 and n = 3.

Generally, an algorithm for color compression is designed to handle frame data as two-dimensional array data. In other words, when the pixels in one frame are a two-dimensional array of j rows and k columns, a j × k two-dimensional array area is defined in the processing device, specifically, on the memory.

In this embodiment, the color compression algorithm for one frame can be used as it is. That is, when the normal frame has a two-dimensional array of pixels of horizontal j pixels × vertical k pixels as described above, in the present embodiment, the one frame group is divided into a two-dimensional array of horizontal mj pixels × vertical nk pixels. Is regarded as one image. Then, by replacing j and k in the conventionally used color compression algorithm with mj and nk, respectively,
A color compression process is performed on one virtual frame, which is a two-dimensional array of pixels × nk pixels, to generate a palletized virtual enlarged image 6.

It is assumed that the color compression process limits the number of colors to N _C. The processing device generates a color palette (color palette information 8) based on a conventional algorithm.
That the color palette is stored palette index is RGB values and distinguishing identification codes them respectively as N _C is the same as conventional. However, in the present embodiment, the color compression processing is different from the conventional color compression processing in that mn frames that are temporally continuous are processed instead of one frame. In other words, the colors stored in the color palette are selected based on the color distribution in the entirety of the continuous mn frames. The color distribution of each successive frame usually has a correlation with each other, except for the exception that a moving image displayed by those frames changes abruptly due to a scene change or the like in the middle. Therefore, the expansion of the range of the color distribution is small compared to the increase in the continuous number (m · n) of moving images to be processed. Therefore, as long as the exceptions described above are not included, even if the number of moving images to be processed collectively is increased, deterioration of the naturalness of color expression is suppressed.

On the other hand, by improving the processing efficiency by performing a plurality of processes collectively, the time processing required for the color compression processing is mn times longer than the time required for processing mn sheets individually. Reduced. Specifically, the color reduction processing which was required m and n times in the conventional method is performed by one common processing with respect to mn images, so that (mn-1) color reduction processing is omitted. can do.

The data to be processed is array data. In general, this array is often a two-dimensional array corresponding to an image, but can be handled in a form expanded to a one-dimensional array. In order to process these array data at high speed, it is effective to perform parallel processing in a pipeline system using a vector processor. At this time, generally, the larger the array size is, the higher the processing efficiency is improved by parallelization. In other words, in the case of performing the vector calculation processing, the processing of the present method in which continuous image data in a time series is treated as virtually enlarged image data can be further speeded up.

The pixel value of each pixel of the palletized virtual enlarged image 6 is calculated by RG according to a conventional method.
It has been converted into a pallet index corresponding to the N _C different colors of the palette from the B value. This enlarged image 6
Corresponds to an image 4 in which a temporally continuous frame sequence 2 is spatially enlarged, and is converted again into a frame group 10 of mn temporally continuous frames. A series of frame groups 10 that are sequentially generated and a frame train after color compression processing are configured.

According to the above-described color compression method, the generated color pallet information 8 is common to each frame included in the frame group 10 which is a unit of the color compression processing. In this method, the color pallet information 8 generated for each frame group of the frame sequence 2 is added to the data of the first frame of the corresponding frame group 10. The first frame contains the color pallet information 8 and can reproduce a color image only with its own data. In this respect, the first frame is similar to an I picture in the MPEG system, so this frame will be referred to as an I picture here. On the other hand, other frames in the frame group 10 that do not include the color pallet information 8 cannot reproduce a color image without referring to the color pallet information 8 of the first frame. Since frames other than the head frame are similar to MPEG P-pictures in that an image cannot be reproduced only by own data, each frame other than the head frame in the frame group 10 is referred to as a P-picture here. Shall be referred to.

Conventionally, each frame is an I picture in which the color palette information 8 and the palette index are stored. However, in this method, only the first frame of the frame group 10 is an I picture, and the other frames that occupy most of the frames are P pictures. That is, most frames include the palette index, but do not hold the color palette information 8. Thus, the data amount of the entire color-compressed frame sequence is reduced.

When the frames of the frame sequence 2 are grouped, for example, assuming at the end of the frame sequence 2, m ·
It is conceivable that only n frames are not collected. In this case, the same frame of less than mn frames collected, for example, the last frame in a time series is repeatedly used to form the mxn enlarged image. Alternatively, when a specific color that will be included in less than mn collected frames is estimated, the specific color may be used to fill in a portion corresponding to the insufficient frame of the enlarged image. Good.

When reproducing a moving image from the sequence of frame data generated in this way, the processing device acquires the color pallet information 8 from the first frame of the frame group 10 and stores it in the memory. For the frames belonging to the same frame group 10, the color palette information 8 is searched using the color palette information 8 in the memory and the palette index as a key, and the R
Reproduce the GB values.

Here, the first frame of the frame group 10 is an I picture. This is because there is an advantage that a frame sequence can be processed in a time series at the time of reproducing a moving image as described above. However, with respect to the accumulated moving image data as described above, it is possible to refer to subsequent frame data in a certain frame reproduction process. Therefore, it is possible to set the I picture to a frame other than the head of the frame group 10.

[Second Embodiment] FIG. 2 is a schematic diagram for explaining a color compression method applied to color moving image data obtained in real time according to a second embodiment of the present invention. is there.

The color moving image data obtained in real time, which is processed here, is moving image data obtained momentarily from an imaging device such as a video camera. That is,
When performing color compression processing for a certain frame, basically no subsequent frame is obtained.

Note that the color compression processing of this system is also realized as a program processing by a computer including a CPU, as in the above embodiment.

Also in this case, the frame sequence 2 is divided into a plurality of frame groups 20, and the processing is performed in units of the frame groups. However, as described above, a certain frame group 2
At the start of the process of 0, all the frame data of the frame group 20 has not been obtained. Frame number N _F that constitute the frame group is set from the outside, depending on the nature of the moving image to be processed. For example, for moving image data in which the change in image between frames is large or in which the movement is intense, a frame group is configured with a small number of frames, and conversely, the change in image between frames is small or the movement is small. For moving image data, a frame group can be composed of many frames.

The processor obtains a first frame data 22 of the frame group 20 performs the conventional similar color compression processing for the first frame data 22, the color palette information 24 of the color number N _C, of the frame A pallet index is generated. By this color compression processing, the top frame data 22 is converted into frame data including color palette information 24 and a palette index. The converted frame data corresponding to the first frame data 22 is the I picture 26 described in the above embodiment.

The other frame data belonging to the same frame group 20, that is, (N _F -1) frame data obtained subsequent to the first frame data 22, are subjected to the first frame data without being subjected to a separate color compression process. 2
Palletizing is performed based on the color pallet information 24 generated from Step 2.

In this palette, color pallet information 2
The one closest to the RGB value of the pixel to be processed among the RGB values registered in No. 4 is searched, and the corresponding palette index is assigned to the pixel. This assignment is referred to herein as color matching.

Here, an example of this color matching will be described. The R, G, and B values of the pixel to be processed are represented by I _R ,
I _G, and I _B, which is registered in the color palette information 24 R,
_Let G and B be P _R (i), P _G (i), and P _B (i). Here, i is a pallet index and has N _C types. In the color matching, the palette index i satisfying the conditions represented by the following equations (1) to (3) is stored in the color palette information 24.
And is assigned to the pixel.

| I _R −P _R (i) | ≦ ε _R (1) | I _G −P _G (i) | ≦ ε _G (2) | I _B −P _B (i) | ≦ ε _B (3) Here, ε _R , ε _G , and ε _B are parameters set from the outside, and correspond to the allowable error amounts of the R, G, and B values in palletizing. Ε _R , ε _G , ε _B are
RGB values I _R assigned to each palette index, I _G, a distribution width of I _B.

If a pallet index that satisfies all of these conditions is not found, for example, the sum of squares of the error amount ｛(I _R -P _R (i) ) ² + (I _G −P _G (i)) ² + (I _B −
A palette index that minimizes P _B (i)) ²探索 is searched for and assigned.

Since the frame group 20 is determined according to the nature of the moving image as described above, the correlation of the color distribution between the frames belonging to the same frame group 20 is strong. _{That, I R, I G, ε} R is the distribution width of the I _B, ε _G,
Even if ε _B is set small, there is a high probability that a pallet index that satisfies the conditions of equations (1) to (3) exists. This is because the color compression processing is performed on the frame group 20 as described above.
This means that fine color gradations are maintained even when represented by the processing for the first frame of, and that the deterioration of the naturalness of color expression is suppressed.

By this color matching, (N _F -1) frame data subsequent to the I picture 26 generated corresponding to the frame group 20 is generated. These are the P pictures 28 described in the above embodiment. As described above, the frame sequence 2 is converted into a frame sequence in which a plurality of frame groups 30 each including one I picture and the following (N _F -1) P pictures are repeatedly arranged.

[0049] According to the color compression method described above, it is possible to reduce the conventional color reduction processing has been performed for each frame to once per N _F frames, it is possible to reduce the processing load of the computer, processing Time is shortened, and high-speed color compression processing is realized. The color pallet information 24 includes
It is stored only in the I picture 26 generated for every N _F images. As a result, the data compression method using the color compression of the present method reduces the data amount of the entire frame group 30 generated by the processing as compared with the conventional data compression method using the color compression in which color pallet information is held in each frame. Higher data compression efficiency is realized.

Incidentally, the generated frame group 3
The method of reproducing a moving image from 0 is the same as that in the above embodiment, and the description is omitted.

Third Embodiment FIG. 3 shows a third embodiment of the present invention, which describes another color compression method applied to color moving image data obtained in real time. It is a schematic diagram.

In the second embodiment, the number of frames included in a frame group is determined by a parameter which can be set from the outside but is kept constant during the color compression processing. On the other hand, a third embodiment described below is
This is a method having, as one feature, changing the number of frames included in a frame group according to a moving image to be processed. In the present embodiment, another form of color matching is employed.

First, in FIG. 3, several frames are put together in order from the beginning of the frame sequence 2 to form a frame group. It is undecided at the time of data input of the head frame how many frames from the head frame of a certain frame group are included in the frame group.
How the first frame is determined, that is, where the frame sequence 2 is divided and a frame group is defined will be described later.

In this method, first, color compression processing is performed on the data of the first frame 40. Color reduction processing when this is carried out by the same method as conventional, thereby selecting color data is RGB values of N _C colors selected (palette data) in the form in which the palette index identifies it were associated with each other Is stored in the color pallet information 42.
In this method, basically, similar to the second embodiment, the palletizing of the frame subsequent to the first frame 40 is performed based on the color pallet information 42 obtained from the data of the first frame 40. This is made possible because, as described in the above embodiment, their color distributions have a strong correlation with each other between successive moving image frames.

However, even if there is a correlation, some fluctuation in color distribution may occur between frames. In order to cope with this fluctuation, an assigned color data range that includes pallet data and represents an ambiguous range is defined. Then, a palette dictionary 44 storing the correspondence between the allocated color data range and the palette index is set. FIG. 4 is a schematic diagram showing an example of the format of the palette dictionary. In the illustrated example, the assigned color data range is stored as a table integrated with the color pallet information 42. In other words, for each pallet index,
The palette data and the assigned color data range are stored in association with each other.

Note that the color pallet information 42 and the assigned color data range can be stored as separate tables. In this case, two tables are provided: a table for storing the palette index and the palette data, and a table for storing the palette index and the assigned color data range.

The assigned color data range shown in this embodiment is
For example, the top 5 in 8-bit binary representation of palette data
This is a range in which the upper limit and the lower limit are defined by the maximum value and the minimum value of the numerical value having the same bit. The range width is the size of the numerical range “8” that can be represented by the lower 3 bits.
It is.

It is sufficient for the palette dictionary to store the information of the upper 5 bits which are the common digits. However, in the present processing apparatus, in consideration of the easiness of data processing in the computer, an area of lower 3 bits corresponding to the fluctuation width of the allocated color data range is also secured in the memory. That is, the assigned color data range is held here as 8-bit data instead of 5-bit data for each of R, G, and B. However, since the lower three bits are ignored in the color mapping process, any bit pattern may be stored. Therefore, simply, the lower three bits may be all “0” or “1”, or the pallet data itself may be stored in eight bits of data in the allocated color data range. Note that the number of lower bits ignored is not required to be 3 bits, but is increased or decreased according to the allowable fluctuation width.

As described above, the definition of the assigned color data range by ignoring the lower several bits does not necessarily mean that the value of the palette data is located at the center of the range. That is, the distance from the pallet data value generally differs between the lower limit and the upper limit of the allocated color data range. Therefore, this allocated color data range does not absorb fluctuations between pallet data frames from top to bottom with respect to the pallet data of the first frame. However, the accuracy is not particularly required here, and such a method is easy to be realized by operating the computer at the data processing instruction level, and the processing speed is significantly increased. There is.

The color matching is performed based on which assigned color data range the pixel value of the target pixel belongs to. Here, the upper and lower limits of the assigned color data range corresponding to the palette index i are represented by symbols _rmH (i) and r, respectively.
_{Expressed in mL} (i). m is one of the color types R, G, and B. If these symbols are used, color matching basically corresponds to selecting a palette index i that simultaneously satisfies the following equations (4) to (6).

R _RL (i) ≦ I _R ≦ r _RH (i) (4) r _GL (i) ≦ I _G ≦ r _GH (i)... (5) r _BL (i) ≦ I _B ≦ r _BH (i) (6) If a palette index that satisfies all of these conditions is not found, for example, color error is minimized, for example, to minimize the error amount (color deviation). sum of squares of the deviations _{{(I R -P R (i} )) 2 + (I G -P G (i)) 2 + (I B -
A palette index that minimizes P _B (i)) ²探索 is searched for and assigned.

Next, a method of determining the first frame will be described. Color matching is performed as described above based on a palette dictionary generated from data of a certain first frame. At this time, based on the color deviation of each pixel, intra-frame color deviation is represented by the following formula Ipushironshiguma, the maximum color deviation epsilon _max is determined. Note that the sum of Expression (7) is the sum of each pixel in the frame, and “ma” on the right side of Expression (8).
x "means that the maximum value of the sum of the color deviations within the frame is taken.

[0063] _{εσ = Σ (| I R -P} R (i) | + | I G -P G (i) | + | I B -P B (i) |) ......... (7) ε max = max _{(| I R -P R (i} ) | + | I G -P G (i) | + | I B -P B (i) |) ......... (8) In addition, (7), (8) sum of squares of the color deviation in parentheses in each right side of _{{(I R -P R (i} )) 2 + (I G -P G (i)) 2 +
Εσ as _{(I B -P B (i)} ) 2}, may be determined epsilon _max.

The εσ thus obtained indicates the magnitude of the overall pixel value shift between the frame to be processed and the first frame. If this value is large, it is highly likely that a scene change of a moving image has occurred in the frame. on the other hand,
ε _max represents a local shift within the frame. This may be a large value in, for example, a frame to which a spot light is applied. For example, in such a case, it is considered that there are many pixels for which the palette index is determined not based on (4) to (6) but by an exceptional process of minimizing the sum of squares of the color deviation.

[0065] For frame being processed as belonging to a frame group of the first frame with those Ipushironshiguma, epsilon it _max is large, the color matching accuracy of using color palette information 42 generated based on the data of the first frame Is inadequate for that frame. Therefore, in this method, ε
sigma, sets an upper limit value epsilon _max of each acceptable parameters, and Ipushironshiguma, frames both epsilon _max exceeds the respective upper limit is redefined as a new first frame. That is, the frame sequence 2 is a preceding frame group up to the frame before this frame, and the subsequent frame becomes a new frame group. In this way, the number of frames included in the frame group is dynamically set.

For the frame newly set as the first frame, the process of generating the color palette information 42 and the palette dictionary 44 is performed using the data as described above.

[0067] Here, to determine a new first frame, Ipushironshiguma, epsilon although both _max is with the proviso that more than each upper limit value, instead Ipushironshiguma, either an upper limit of epsilon _max A condition that the value is exceeded may be adopted.

The head frame of the frame group after the color compression processing is the above-described I picture 26, which stores both the color palette information 42 and the palette index information of the frame. Another frame belonging to the frame group is the P picture 28, and only the palette index information of the frame is stored.

The I and P picture data have a data structure and information for distinguishing each other. At the time of reproduction, the P-picture frame is restored from the pallet index to an RGB value with reference to the color pallet information 42 stored in the I-picture located at the head of the frame group to which the frame belongs. That is, when the frame is identified as an I picture, the color pallet information 42 is extracted from the frame data and stored in the memory of the processing device. The I picture is sequentially restored with RGB values of each pixel based on its own color pallet information 42, and is stored in the frame buffer. When the processing of one frame is completed, data is output from the frame buffer to the display device and an image is displayed. On the other hand, when the reproduced frame is recognized as a P picture, the RGB values of each pixel are determined using the color pallet information 42 stored in the memory, and are stored in the frame buffer. Hereinafter, image display is performed in the same manner as the I picture.

According to the above-described color compression method, the color reduction processing conventionally performed for each frame can be reduced to once for each frame group, so that the processing load on the computer can be reduced and the processing time can be reduced. Shortened and faster color compression processing is realized. Further, the color pallet information 42 is held only in the I picture 26 generated one for each frame group. As a result, in the data compression method using color compression of the present method, the data amount of the entire frame sequence generated by the processing is reduced as compared with the conventional data compression method using color compression in which color pallet information is held in each frame. High data compression efficiency is realized.

Incidentally, the method described here is applied directly to a moving picture signal outputted in total of 24 bits, that is, 8 bits each of R, G and B outputted from the video camera. The number can be arbitrarily changed as needed.

The current computer has 32 bits (= 4 bits).
Bytes) are often handled as a basic unit of one word. Therefore, if the above-described 24-bit, that is, 3-byte moving image data is added with, for example, a dummy 1 byte, and the entire RGB value is set to one word, the above-described color matching processing for searching the palette dictionary is performed. The number of execution instructions of a program for performing processing can be reduced, and the processing speed can be further increased. For example, VLIW (Very Long Instruction)
The speed of the processing can be further increased by performing a color mapping process using one word obtained by adding a dummy one byte to 24 bits (3 bytes) of RGB using a CPU of a Word) type.

[0073]

According to the color moving picture data compression method of the present invention, a frame sequence is divided into frame groups, and one piece of color pallet information is generated for each frame group. Then, all the frames belonging to the frame group are indexed based on the one color pallet information. Therefore, the color pallet information is generated only for each frame group in the color moving image data, so that the processing load is reduced and the processing speed is increased.

According to the color moving image data compression method of the present invention applied to the stored moving image data, the color compression processing is performed on moving image data included in a frame group as a unit. Thereby, one common color pallet information is generated based on the data of all the frames included in the frame group, so that an effect that accurate palletizing is performed for each frame in the frame group is obtained. .

According to the color moving image data compression method of the present invention applied to the stored moving image data, one virtual image obtained by two-dimensionally arranging a plurality of frames included in a frame group Is subjected to the color compression. As a result, the effect that the color compression processing conventionally applied to one frame can be used only by changing the image size to be processed and the processing can be easily realized is obtained. By increasing the size of the array data, which is image data, it is possible to obtain an effect that processing efficiency and speed can be improved.

According to the color moving image data compression method of the present invention applied to moving image data obtained in real time, color compression is performed on the first frame included in each frame group. Accordingly, the color compression processing can be sequentially performed from the frame that has just been input without waiting for the input of a frame that has not yet been obtained among the moving image data obtained in real time. That is,
Since processing can be performed according to the time-series flow of data,
The effect that processing is simple is obtained, and further, since there is no need to accumulate frames to perform processing, the required amount of resources such as the memory and storage device of the computer executing the processing can be suppressed. Is also obtained.

According to the color moving image data compression method of the present invention applied to moving image data obtained in real time, the assigned color data including the selected color data corresponding to each index for each index based on the color pallet information. Determining a range and indexing includes selecting an assigned color data range that includes the color data of each pixel, and replacing the pixel color data with an index corresponding to the selected assigned color data range. Will be As a result, there is an effect that the moving image data can be compressed by the color compression processing which is strong against the fluctuation of the frame data which occurs even under the strong correlation between the frames of the moving image data.

According to the color moving image data compression method of the present invention applied to moving image data obtained in real time, the assigned color data range is a numerical value whose predetermined upper digit number is the same as the selected color data associated therewith. The upper and lower limits are defined based on the maximum and minimum values of. As a result, processing using a computer is facilitated, and in particular, the effect of suppressing the number of executed instructions and improving the processing speed is obtained.

According to the color moving picture data compression method of the present invention, the color pallet information of each frame group is included only in the frame data corresponding to one frame included in the frame group. Thus, since the compressed color moving image data has only one color pallet information for each frame group, the effect of improving the data compression ratio is obtained.

According to the color moving picture data compression method of the present invention, the color pallet information of each frame group is incorporated in the frame data corresponding to the first frame included in the frame group. As a result, the generated compressed moving image data can be reproduced in real time from the beginning. In other words, since the processing can be performed in accordance with the time-series flow of the data, the effect that the processing is simple is obtained. Further, since there is no need to accumulate the data, the memory or the storage device of the computer that executes the processing can be obtained. Also, the effect that the required amount of resources can be suppressed can be obtained.

[Brief description of the drawings]

FIG. 1 is a schematic diagram for explaining a color compression method according to a first embodiment of the present invention applied to accumulated color moving image data.

FIG. 2 is a schematic diagram for explaining a color compression method applied to color moving image data obtained in real time according to a second embodiment of the present invention.

FIG. 3 is a schematic diagram for explaining another color compression method applied to color moving image data obtained in real time according to a third embodiment of the present invention.

FIG. 4 is a schematic diagram showing an example of a format of a palette dictionary.

[Explanation of symbols]

2 frame sequence, 4,6 images, 8, 24, 42 color palette information, 3, 10, 20, 30 frame groups, 22 first frame data, 26 I pictures, 2
8P picture, 40 first frame, 44 palette dictionary.

Claims (8)

[Claims] 1. A method of generating color-compressed moving image data by performing color compression for limiting the number of colors on color moving image data, wherein the compressed moving image data is selected by color compression. A color moving image that includes color pallet information that defines the correspondence between selected color data representing a color and an index that distinguishes the selected color data, and that compresses the data amount by indexing that replaces the color data of each pixel with the index. In the data compression method, a column of frames constituting a moving image is divided into a plurality of frame groups, and one color palette information is generated for each frame group based on a predetermined color compression procedure. Each of the frames included in the frame group,
The method according to claim 1, wherein the indexing is performed based on the color pallet information corresponding to the frame group. 2. The color moving image data compression method according to claim 1, wherein the moving image data compression method is applied to the stored moving image data. Wherein the color pallet information corresponding to the frame group is generated. 3. A virtual one obtained by two-dimensionally arranging a plurality of frames included in the frame group.
3. The color moving image data compression method according to claim 2, wherein the color compression is performed on one image, and the color pallet information corresponding to the frame group is generated. 4. The color moving image data compression method according to claim 1, wherein the color compression is performed on the first frame included in each of the frame groups. A color moving image data compression method, wherein the color pallet information corresponding to the frame group is generated. 5. An assigned color data range including the selected color data corresponding to each index is determined based on the color palette information, and the indexing includes color data of each pixel. 5. The method according to claim 4, further comprising: selecting the assigned color data range, and replacing the color data of the pixel with the index corresponding to the selected assigned color data range. The color moving image data compression method described in the above. 6. The allocation color data range is based on a maximum value and a minimum value of numerical values having the same predetermined number of upper digits as the selected color data associated with the allocation color data range via the index. 6. The color moving image data compression method according to claim 5, wherein the upper and lower limits are within a defined range. 7. The apparatus according to claim 1, wherein the color pallet information of each of the frame groups is included only in frame data corresponding to one of the frames included in the frame group. The color moving image data compression method according to any one of the above. 8. The apparatus according to claim 1, wherein the color pallet information of each frame group is included in the frame data corresponding to the first frame included in the frame group. The color moving image data compression method according to any one of the above.