WO2002091301A1 - Method for generating image data - Google Patents

Abstract

Image data for reproducing an original image comprising an 8 8 pixel array is generated. Four division levels n=0 to 3 are defined, the original image is divided into 22n blocks and 85 blocks in all and their addresses are defined. A predetermined pixel of the original image included in each block is determined as a representative pixel of that block and the pixel value of each representative pixel is determined as the pixel value of that block. An image data unit representative of the division level, address and pixel value of a block of each of the 85 blocks is determined. Among the 85 image data unit, a plurality of image data units which can reproduce the original image under the condition that when a plurality of items of data represent an identical position, the pixel value of data on more minute division is used are extracted. To judges if a specific image data unit must be extracted, the pixel value of the specific image data unit is compared with the pixel value of image data unit of division rougher by one step including the block of the specific image data unit and the image data units are extracted if they are not identical.

Description

 Description 誊 Image data generation method

 The present invention relates to a method for generating image data for reproducing an original image based on an original image represented by a two-dimensional pixel array in which a large number of pixels are arranged. The present invention relates to a method for generating image data suitable for driving a large display device. Background technology

 BACKGROUND ART Wall-mounted display devices such as electronic bulletin boards and advertisement display boards are widely used as means for presenting information to an unspecified number of people on the street. In these display devices, usually, a large number of display elements for one pixel are arranged in a matrix to form a two-dimensional pixel array, and each display element is driven by electric power to change the display mode, thereby obtaining information. Displaying. For example, an electric bulletin board uses a single light bulb as a display element for one pixel, arranges the light bulbs vertically and horizontally, and emits light at a specific position to display characters and images. Having. Recently, electronic bulletin boards using light emitting diodes instead of light bulbs have become widespread.

In addition, advertising panels and the like that use panel-type display elements as display elements that constitute individual pixels are also used. This panel-type display element has a plurality of display surfaces that do not have the function of emitting light by themselves, and in fact, only one of them is presented. In general, a display surface to be presented can be selected by using a rotating mechanism such as a motor, and if a display surface to be presented is selected for each pixel, characters and images are displayed as a whole. Can be displayed. In this way, a two-dimensional pixel array is formed by arranging a large number of display elements having a function of changing the display mode for one pixel by driving with electric power in a matrix, and information is displayed on the two-dimensional pixel array. Specific display devices for performing the display of, for example, those disclosed in International Publication Nos. WO96 / 10244 and WO97 / 25705 based on the Patent Cooperation Treaty are disclosed. Therefore, an image data format suitable for driving these display devices is disclosed in International Publication No. WO97 / 36279.

 An object of the present invention is to provide a method for generating image data for reproducing the original image by the various display devices described above, based on the original image represented by a two-dimensional pixel array. And Disclosure of the invention

 (1) According to a first aspect of the present invention, based on an original image represented by a two-dimensional pixel array in which a number of pixels having a predetermined pixel value are arranged, image data for reproducing the original image is obtained. In the method of generating,

 A plurality of division modes for dividing the two-dimensional pixel array into a plurality of blocks are defined. At this time, each division mode is set to have a different divisional power, and each division mode indicates a division level. A division level definition stage that can be identified by level;

 An address definition step for defining an address for indicating each block for each individual division mode;

 A pixel value defining step of defining a predetermined pixel on the original image included in each block as a representative pixel of the block, and defining a pixel value of a representative pixel of each block as a pixel value of the block;

As the image data for indicating an image in a specific block determined by a specific division level and a specific address, the specific division level, the specific address An image data defining step of defining unit image data having pixel values of the specific block.

 Is performed.

 (2) According to a second aspect of the present invention, in the method for generating image data according to the first aspect,

 In the division level definition stage, a plurality of division levels in which the details of the division gradually differ are defined, and the k-th division level and the (k + 1) -th division in which the one-step division is finer than this. By dividing one block obtained by the division mode indicated by the k-th division level into a plurality of sub blocks, a plurality of blocks according to the division mode indicated by the (k + 1) -th division level are obtained. In this case, the block obtained by the division mode indicated by the (k + 1) th division level is included in the block obtained by the division mode indicated by the kth division level. A large number of blocks having a hierarchical structure according to a plurality of division levels are defined.

 (3) According to a third aspect of the present invention, in the method for generating image data according to the second aspect described above,

 From the unit image data of all blocks obtained by all division modes defined in the division level definition stage, predetermined unit image data is extracted and extracted according to conditions for reproducing the original image. An image data extraction step of outputting the obtained unit image data group as “image data for reproducing the original image” is further performed.

 (4) In a fourth aspect of the present invention, in the method for generating image data according to the third aspect,

The condition for reproducing the original image is as follows: `` If there is a plurality of unit image data indicating pixel values at the same position on the plane, the pixel value of the unit image data that is finer than the coarser unit image data is "Use" In the image data extraction stage, a plurality of sets of unit image data are extracted from the plurality of unit image data so that the original image can be reproduced under these conditions, and the extracted unit image data group is extracted. Is output as “image data for reproducing the original image”.

 (5) According to a fifth aspect of the present invention, in the method for generating image data according to the fourth aspect described above,

 In the image data extraction stage, the process of determining whether or not to extract a plurality of unit image data as extraction candidates in order from the coarsely divided unit image data to the finely divided unit image data is referred to as “extraction”. If the image in the area indicated by the candidate unit image data is the same as the image reproduced by the already extracted unit image data, it will not be extracted, and if it is not the same, it will be extracted. '' The determination is performed based on the criteria.

 (6) According to a sixth aspect of the present invention, in the method for generating image data according to the fifth aspect,

 When judging a specific unit image data to be an extraction candidate, the “pixel value of this specific unit image data” and “the block of this specific unit image data are included and the division is coarse by one step The pixel value of the unit image data is compared with and the determination is made that if the rain person is the same, it is not extracted, and if it is not the same, it is extracted.

 (7) According to a seventh aspect of the present invention, in the method for generating image data according to the third to sixth aspects described above,

In the image data extraction stage, of a large number of blocks having a hierarchical structure, a first group of blocks including a first block and blocks included therein, and a second group of blocks including the first block and blocks included therein. Of the second group of blocks, each unit image data of the first group of blocks and each unit image data of the second group of blocks correspond to each other. In this case, an identification code is defined for each set of unit image data for the block group of the first group, and the defined identification code is extracted instead of each unit image data for the block group of the second group. It was made.

 (8) According to an eighth aspect of the present invention, in the method for generating image data according to the third to sixth aspects described above,

 When creating the first image data for reproducing the first original image and the second image data for reproducing the second original image, a number of blocks defined for the first original image Of the first group consisting of the first block and the blocks included therein, and the second block of the many blocks defined for the second original image and the second block When the unit image data of the block group of the first group and the unit image data of the block group of the second group match with each other with respect to the block group of the second group An identification code is defined for each set of unit image data for the block group of the group, and the defined identification code is used in place of each unit image data for the block group of the second group. In which it was to be out.

 (9) According to a ninth aspect of the present invention, in the method for generating image data according to the third to eighth aspects described above,

 At the division level definition stage, from the coarsest division mode in which the entire two-dimensional pixel array including the original image is regarded as one block, to the finest division mode in which each pixel of the two-dimensional pixel array is regarded as one block Until then, a plurality of division modes with different division details are defined.

 (10) A tenth aspect of the present invention is the image data generation method according to the ninth aspect,

As divided configuration shown by the division level n, define the dividing manner to obtain 2 ² solid block by respectively 2 ⁿ divided into vertical and horizontal two-dimensional pixel array, n == l, 2, ... , i, ... ^ Define a total of N division modes for, and include the original image As a two-dimensional pixel array, a square pixel array in which 2 ⁿ pixels are arranged vertically and horizontally is used.

 (11) According to a eleventh aspect of the present invention, in the method for generating image data according to the tenth aspect described above,

In the address definition stage, for each of the four blocks obtained in the division mode indicated by the division level n = 1, a 2-bit address of 0, 01, 10 and 11 is defined, and the division is performed. For two blocks obtained in the division mode indicated by the level n = i, 2 ^{2 (i—} "blocks representing the blocks obtained in the division mode indicated by the division level n = (i-1) are indicated. The address is defined by adding one of 00, 01, 10 and 11 to the lower part of the address.

 (12) According to a second aspect of the present invention, in the method for generating image data according to the first aspect described above,

When defining the representative pixel for each block with a division level of _n = 1 or more, add the same two bits as the last two digits of the address to the end of the address of the block as many times as necessary. Calculates the address of a block having the same size as one pixel constituting the original image, and determines the pixel at the position corresponding to the block indicated by the obtained address as the representative pixel. Is

 (13) According to a thirteenth aspect of the present invention, in the method for generating image data according to the tenth aspect described above,

 As a condition for reproducing the original image, set the condition of "display on a display screen whose vertical and horizontal aspect ratio is not 1: 1".

In the image data extraction stage, the aspect ratio of the display screen is set to a part of a square pixel array composed of 22 ⁿ pixels so that the original image can be reproduced under this condition. Defines the effective area with the image and effectively indicates the image information within this effective area A plurality of sets of unit image data necessary for the above are extracted, and the extracted unit image data group is output as “plane image data for reproducing the original image”.

 (14) A fourteenth aspect of the present invention is the method for generating image data according to the first to thirteenth aspects,

 The division level, address, and pixel value are each represented by a bit, and the bit length of the division level is fixed, and the sum of the bit length of the address and the bit length of the pixel value is fixed. The length can be recognized, and the bit length that composes the address can be recognized based on the division level.

 (15) According to a fifteenth aspect of the present invention, in the method for generating image data according to the first to thirteenth aspects,

 A unit image data comprising at least a bit indicating an address and a bit indicating a pixel value, wherein information indicating a division level is expressed as a bit length of the bit indicating the address; Generation method.

 (16) According to a sixteenth aspect of the present invention, in the method for generating image data according to the first to fifteenth aspects,

 In the image data extraction stage, the extracted unit image data is output in order from the coarsely divided unit image data to the finely divided unit image data.

 (17) According to a seventeenth aspect of the present invention, in the method for generating image data according to the first to sixteenth aspects,

 Generating image data for reproducing moving images by preparing multiple still images in a fixed presentation order to compose a moving image and generating each unit image data group using each still image as the original image It is something to do.

 (18) According to an eighteenth aspect of the present invention, in the method for generating image data according to the seventeenth aspect,

When generating a group of unit image data for each still image, The unit image data group is generated using the difference image from the image as the original image.

 (19) A nineteenth aspect of the present invention is the image data generation method according to the first to eighteenth aspects, wherein:

 In the pixel value defining step, for a block obtained by a division mode coarser than a predetermined criterion, a predetermined background pixel value is replaced with a predetermined background pixel value instead of the pixel value of the representative pixel in the block. It is defined as a pixel value.

 (20) A twenty-fifth aspect of the present invention provides a program for executing the image data generation method according to the first to nineteenth aspects using a computer, and stores the program in a computer. It is recorded on a readable recording medium and can be distributed. Brief explanation of drawings

 FIG. 1 is a diagram showing four types of division modes for dividing a two-dimensional pixel array into a plurality of blocks, and addresses defined for each block obtained in each division mode.

 FIG. 2 is a diagram showing division levels and bit representations of addresses for each division mode.

 FIG. 3 is a diagram showing a basic format of a unit image data generated by the image data generating method according to the present invention.

 4A to 4E are diagrams showing an original image and a state in which the original image is represented by a plurality of unit image data.

 FIGS. 5A to 5F are other views showing the original image and the I-dog state expressed by a plurality of unit image data.

Figure 6 shows the individual data when the sum of the address length and the data length is fixed at 32 bits. FIG. 9 is a diagram illustrating bit allocation of address / data for a division mode.

 FIG. 7 is a diagram illustrating an example of a method of expressing an original image by mixing various division modes.

 FIG. 8 is a diagram showing individual pixels constituting the original image.

 9A to 9D are diagrams showing block configurations in various division modes with respect to the two-dimensional pixel array forming the original image shown in FIG.

 Figure 10 shows a total of 85 blocks B (0) to B (1 1 1 1 1 1) defined by the division levels n = 0 to 3 defined for the original image shown in Figure 8. This is a table listing the division levels, addresses, and pixel values.

 FIG. 11 is a diagram showing a first example of an original image to which the method of generating image data according to the present invention is applied.

 FIGS. 12A to 12D are diagrams showing blocks in various division modes defined based on the original image shown in FIG.

 FIG. 13 is a table showing information relating to blocks extracted to reproduce the original image shown in FIG. 11 from among the many blocks shown in FIGS. 12A to 12D. FIG. 14 is a diagram illustrating a second example of an original image to which the image data generation method according to the present invention is applied.

 FIGS. 15A to 15D are diagrams showing blocks in various division modes defined based on the original image shown in FIG.

 FIG. 16 is a table showing information on blocks extracted to reproduce the original image shown in FIG. 14 from the many blocks shown in FIGS. 1558 to 15D. FIG. 17 is a diagram illustrating a third example of an original image to which the method of generating image data according to the present invention is applied.

 FIGS. 18A to 18D are diagrams showing blocks in various division modes defined based on the original image shown in FIG.

Figure 19 shows the original image shown in Figure 17 among the many blocks shown in Figure 18D. It is a figure showing a block which should be extracted in order to reproduce.

 FIG. 20 is a table showing information on blocks extracted from the many blocks shown in FIGS. 18A to 18D in order to reproduce the original image shown in FIG. FIG. 21 is a diagram for explaining the handling when the effective area E of the original image has a shape different from the square pixel array.

 FIG. 22 is a diagram showing address information of a block for defining the effective area E shown in FIG.

 FIG. 23A and FIG. 23B are diagrams showing an example of an original image constituting a moving image. FIG. 24 is a diagram showing a difference image D of the original image shown in FIGS. 23A and 23B. BEST MODE FOR CARRYING OUT THE INVENTION

§ 1. Basic format of image data generated by the present invention

A method for generating image data according to the present invention generates image data for reproducing the original image based on an original image represented by a two-dimensional pixel array in which a large number of pixels having predetermined pixel values are arranged. It is intended to do so. The image data generated by the method according to the present invention is particularly suitable for driving a display device having a large number of display elements constituting pixels arranged in a matrix. In such a display device, each display element has a function of changing a display mode for one pixel by driving with electric power. For example, light bulbs and light emitting diodes are used as display elements in general electric light boards. In addition, advertising display boards and the like that use a no- and flannel-type display element are also used. This panel-type display element does not have a function to emit light by itself, and has a plurality of display surfaces, and in reality, only one of them is presented. Usually, it is possible to select the display surface to be presented using a rotating mechanism such as a motor. The basic format of image data suitable for driving such a display device is disclosed in the above-mentioned International Publication No. WO97 / 36927. The present invention proposes a new method for generating image data having a format disclosed in the above-mentioned publication based on an arbitrary original image. Therefore, first, the substance of this basic format will be described.

 In this basic format, image information on the original image is represented as a set of image data in block units. For this purpose, a plurality of division modes are defined for the two-dimensional pixel array constituting the original image, and an address for indicating each block is defined for each division mode. FIG. 1 is a diagram showing four types of division modes in which a two-dimensional pixel array constituting an original image is divided into a plurality of blocks, and addresses defined for each block obtained in each individual division mode. Each division mode is indicated by a division level n.

 The division mode shown in the first row of FIG. 1 is a division mode indicated by a division level n = 0, and no division is actually performed. That is, only one large block including the entire original image P is defined. Here, for convenience of explanation, this original image P is a two-dimensional image consisting of a total of 6 5 5 3 6 pixels composed by arranging 256 pixels vertically and 256 pixels horizontally. Let's say it consists of a pixel array. In this case, in the division mode represented by the division level n == 0, only one large block including all the pixels of 655336 is defined, and the address for specifying the block is Not required.

On the other hand, the division mode shown in the second row of FIG. 1 is a division mode indicated by a division level n = 1, in which two divisions are performed vertically and horizontally, that is, a total of four divisions are performed. As a result, the two-dimensional pixel array is divided into four blocks a, b, c, and d. Each of the blocks a to d is composed of 1638.sup.4 pixels arranged in a matrix of 128.times.128 in all directions. Further, the division mode shown in the third row of FIG. 1 is a division mode indicated by a division level n = 2. Yes, it is a total of 16 divisions, 4 divisions vertically and horizontally. Each of the 16 blocks is composed of 409 x 6 pixels arranged in a matrix of 64 x 64 in all directions. Further, the division mode shown in the fourth row of FIG. 1 is a division mode indicated by a division level n = 3, which is obtained by performing a total of 64 divisions, eight in each of the vertical and horizontal directions. is there. Each of the 64 blocks is composed of 102 4 pixels arranged in a matrix of 32 × 32 matrix ^ vertically and horizontally.

In FIG. 1, only the division level n = 3 is shown, but if the number of divisions is increased in the same manner, eventually, in the division mode indicated by the division level n = 8, 2 5 Six divisions are performed, so that 6 5 5 3 6 blocking forces ^{s are} formed. In other words, in this division mode with the division level n = 8, one block corresponds to one pixel. In this specification, the number force magnitude les window at the division level n, to be expressed as "Reberuka on" or "high level", conversely, the number mosquito division level ^n? Smaller, the "level We will describe it as "low" or "low level".

In the present invention, as described above, a plurality of division mode forces ^s for dividing the two-dimensional pixel array including the original image into a plurality of blocks are defined, and each division mode is defined by a division level n indicating the fineness of the division. Will be identified. In principle, if there are multiple types of division mode forces ⁵ ′ with different subdivisions, any definition is possible if it can be defined.In practice, as shown in this example in FIG. 1, the division level n as divided embodiment shown in, defining a dividing manner to obtain 2 ^{2 n} blocks by each 2 ⁿ divided into vertical and horizontal two-dimensional pixel array, n = l, 2, ··· , i, · '· It is preferable to define a total of 分割 division modes for Ν.

By defining such a division mode, it is possible to define a plurality of division levels that vary stepwise by s. In addition, regarding the k-th division level and the (k + 1) -th division level in which the one-step division is finer than this, the k-th division level By further dividing one block obtained by the division mode indicated by the eye division level into a plurality of blocks, a plurality of block forces ^s ' can be obtained by the division mode indicated by the (k + 1) th division level Become like That is, the block obtained by the division mode indicated by the (k + 1) th division level is included in the block obtained by the division mode indicated by the kth division level, and as a whole, It will be many blocks force ^s definition having a hierarchical structure corresponding to a plurality of decomposition levels.

 As described above, when a plurality of division modes are defined, an address for indicating each block is defined for each division mode. In the example shown in Fig. 1, no address is defined for the division level n = 0 (there is no need for an address because there is only a single block), but the division state indicated by the division level n = 1 As shown in the figure, the four blocks obtained in the above are defined as addressing s consisting of two bits of 0, 01, 10 and 11 respectively, and are indicated by the division level n = 2. As shown in the figure, for the 16 blocks obtained in the divided mode, addresses each having 4 bits of 0 000, 0001, 00 10, 00 11,... Are defined, and the division level n = As shown in the figure, for each of the 64 blocks obtained in the division mode indicated by 3, an address composed of 6 bits of 000000, 000001,... Is defined.

After all, in the embodiment shown in FIG. 1, the address of 22 ² blocks obtained in the division mode represented by the division level n = i is changed to the next division level n = (i−1) An address definition such as that indicated by an address obtained by adding one of 00, 01, 10 and 11 to the lower part of the address indicating 2 ² — "blocks obtained in the division mode indicated by For example, the addresses for the blocks e, f, g, and h obtained in the division mode indicated by the division level n = 2 are the divisions indicated by the next lower division level n = 1. A block a (block occupying the same position as blocks e, f, g, h) obtained in the embodiment, with addresses 0, 0, 1, 10, and 11 added below the address "0 0" It has become. Here, which two bits should be added to the lower order is determined in the same way as the address definition for the four blocks a, b, c, and d. For example, the mutual position of the four blocks e, f, g, and h is equivalent to the mutual position of the four blocks a, b, c, and d, so the lower two bits of the address of the block e are The address of block a is set to "0 0", the lower 2 bits of the address of block f are set to "0 1", the same as the address of block b, and the lower 2 bits of the address of block g are the address of block c. And the lower 2 bits of the address of block h are the same as the address of block d.

Of course, in practicing the present invention, it is not always necessary to perform the above-described address definition. However, in order to reduce the calculation load and perform an efficient display operation, the above-described address definition is performed. Is preferred. With such an address definition, by removing the lower two bits from the address of a specific block, the address of the block at the same position in the next lower division level can be obtained. The number of bits required for such an address definition is indicated by 2 n bits as shown in FIG. In addition, the number of the display resolution, i.e., resulting et be blocked at each division level n, as shown in FIG. 1, a 2 ^{2 n.}

FIG. 2 is a diagram showing an example of a bit expression of a division level and an address in each of the above division modes. In this example, the division level n is represented by 4 bits, and it is possible to define 16 division modes from n = 0 to 15. On the other hand, the number of address bits required to indicate each block is different for each division level, as described above. Generally, each additional division level requires two extra bits of address. . Therefore, the highest division level n = 15 requires as many as 30 bits of address. Thus, a display resolution as high as 1 G can be obtained.

 As described above, in the present invention, the image information on the original image is represented as a set of image data in blocks, but the image information in one block has a format as shown in FIG. It is indicated by unit image data having. In this format, the division level, address, and pixel value are listed in this order. Here, the bit length of the address portion is determined based on the division level as shown in FIG. 2, and the higher the division level, the longer the bit length of the address. As shown in FIG. 2, if the division level is represented by 4 bits, the first 4 bits of the bit string “00001 0 1” indicate the division level n = 1, and the remaining 2 bits The bit indicates address "01". Therefore, the block b having the division level n = 1 in FIG. 1 is specified by the 6-bit bit string. On the other hand, the pixel value which is the last component of the format of FIG. 3 is a representative pixel value of a pixel constituting the original image in this block, and the corresponding block is displayed in a state corresponding to the representative pixel value. Indicates what to do.

 After all, one unit of image data described in the format of Fig. 3 is data composed of a specific division level, a specific address, and a specific pixel value, and is specified by a specific division level and a specific address. This indicates that pixels having a specific pixel value are spread all over the block.

For example, if 1-bit information of “0” or “1” is assigned as the pixel value, “0” means off and “1” means on, the display corresponding to one pixel It is possible to generate image data suitable for driving a display device in which an element is constituted by one light bulb and a number of light bulbs are arranged vertically and horizontally on a two-dimensional plane. More specifically, the 7-bit unit image data represented by the format shown in FIG. 3 of “00001 01 1” has a division level “0001” and an address “01”. And the data "1", which means that all the display elements (light bulbs) belonging to the block b of the division level n = 1 in Fig. 1 Turn on ".

 Hereinafter, when the unit image data based on the format shown in FIG. 3 is indicated, for convenience of explanation, it is referred to as “a bit sequence indicating a division level / a bit sequence indicating an address Z a bit sequence indicating a pixel value”. As shown, a slash is inserted between each bit string. For example, the 7-bit command described above is represented as "0001/01 1" in this specification. Of course, in practice, there is no slash between bit strings.

As described above, by using the unit image data described in the format shown in FIG. 3, an arbitrary display device in which display elements corresponding to one pixel are arranged in a matrix can be used for an arbitrary display. For example, it is possible to assign one bit information of “0” or “1” as the pixel value, and “0” means off and “1” means on. For example, the unit image data consisting of a 5-bit bit string such as “0000 / (no address) Z1” is a light bulb belonging to the block with the division level n = 0 (all the light bulbs of this display device). ), The unit image data consisting of a 9-bit bit string “00 10/001 1 0” is stored in a block h with a division level n = 2. Turn off all the light bulbs to which it belongs "and display the instruction" 001 1/0 1 0 1 0 1 / The unit image data composed of 11-bit bit strings of 1 "indicates a display instruction of" all the light bulbs belonging to the block i of the division level _n = 3 are lit ". Of course, in order to correctly execute the instruction based on the unit image data described in such a format, a controller that recognizes the address in the display device and performs control for turning on or off a predetermined light bulb is prepared. You need to keep it.

Next, an example will be shown in which information on the original image can be efficiently represented by using the unit image data group described in the format shown in FIG. For example, as shown in Figure 4A, a total of 64 pixels arranged in 8 rows and 8 columns Let's assume that the original image was drawn on the two-dimensional pixel array consisting of white and gray pixels using the pattern shown in the figure. Here, the pixel value of a white pixel is defined as "0", and the pixel value of a gray pixel is defined as "1". As a typical method for indicating such information on the original image as image data, a method using a so-called bitmap format is known. In this method, it is necessary to list the pixel values of all 64 pixels in total as codes, and a total of 64 bits of image data is required.

 On the other hand, if the format according to the present invention is used, it is possible to represent an original image having the pattern shown in FIG. 4A with only nine sets of unit image data. First, as the first unit image data, as shown in the lower section of FIG. 4B, unit image data consisting of a 5-bit bit string of "0000 / (no address) / 0" is prepared. This unit image data is such that all pixels (all 64 pixels in this example) belonging to the block with the division level n = 0 are white (pixel value "0"). And the blocks as shown in the upper section of FIG. 4B. Next, as the second unit image data, as shown in the lower section of FIG. 4C, unit image data relating to a division level n = 1 of “0 001/00/1” is prepared. As shown in the upper section of FIG. 4C, this unit image data has information indicating that all pixels belonging to the upper left block (address “00”) at the division level n = l are gray (pixel value “1”). Is shown. It is not necessary to prepare a unit image data that shows information on other blocks (upper right, lower left, lower right blocks) indicated by broken lines in the upper section of FIG. 4C.

Then, as the third unit image data, as shown in the lower section of FIG. 4D, “00 10/01 10/1 '\” 0010/1001/1 ”,“ 0010/1 101/1 ”,“ 001 Prepare the unit image data for the division level n = 2 which is 0X1 1 10/1 ". As shown in the upper column of Fig. 4D, these four sets of unit image data All pixels belonging to the block are gray (pixel value "1") ". Finally, as the fourth unit image data, as shown in the lower section of FIG. 4E, “00 1 1/01 1 11 0/1”, “00 1 1/101101/1”, “0011” / 1 1 1100/1 "Prepare unit image data for the division level n = 3. As shown in the upper section of FIG. 4E, these three sets of unit image data are “all pixels belonging to three specific blocks at a division level n == 3 are gray (pixel value“ 1 ”).” This shows 1 Seikogura.

Thus, when a total of nine sets of unit image data shown in the lower columns of FIGS. 4B to 4E are prepared, the images represented by these nine sets of unit image data are synthesized. can original image force ^s reproduction shown in FIG. 4 a. However, when synthesizing the image represented by these nine sets of unit image data, if the unit image data indicating the pixel values at the same position on the plane exists, the unit image Use pixel values of unit image data that are more finely divided than data. " For example, the unit image with the division level n = 1 shown in Figure 4C is better than the pixel value "0" of the unit image data "0000 / (no address) Z0" with the division level n = 0 shown in Figure 4B. Since the pixel value "1" of the data "0001 / 00Z1" is used, the upper left 1Z4 part of the image shown in the upper part of FIG. It will be replaced by the upper left 1Z4 image shown in the upper row of 4C (the gray area surrounded by the solid line). Similarly, part of the other white part of the image shown in the upper part of FIG. 4B is replaced by the gray part of the upper part of FIG. 4D and the upper part of FIG. 4E. In other words, the images are sequentially superimposed on each other in the order of FIG. 4B, FIG. 4C, FIG. 4D, and FIG. 4E.

Thus, the finally obtained composite image is the same as the original image shown in FIG. 4A. As a result, the original image shown in FIG. 4A was efficiently reproduced by the nine sets of unit image data shown in FIGS. 4B to 4E. However, in the above example, comparing only the number of bits required to represent the original image shown in FIG. By using the expression in the conventional general bitmap format, the total

For 64 to live in the image data of the bit, by utilizing the representable by Fomatsu bets according to the present invention, a total of 8 1-bit image data force ^s required of, utilizing a full Omatsu bets according to the present invention The result is that more bits are required. However, this is because the original image shown in FIG. 4A is a simple image composed of a total of 64 pixels arranged in 8 rows and 8 columns. In general, the larger the two-dimensional pixel array that includes the original image, the larger the number of bits required for the conventional bitmap format representation, and thus the format according to the present invention. Using expressions is more efficient.

 Further, the format according to the present invention is also effective when generating image data for giving an instruction to rewrite from one display mode to another display mode. For example, if the original image shown in Fig. 4A changes to the original image shown in Fig. 5A, this change can be expressed by only two unit image data. First, as shown in FIG. 5B, the pixel values of 16 pixels belonging to the upper left block of the four-dimensional two-dimensional pixel array are divided according to the unit image data regarding the division level n = 1, such as "0001 / 00Z0". Are all "0". Next, as shown in FIG. 5C, a new pixel value is defined for one pixel once defined as the pixel value “0” by the unit image data regarding the division level n = 2 of “0011/00111 1 Z1”. Redefined as "1". In other words, the synthesis is performed with priority given to the pixel values of the unit image data that is finely divided (Fig. 5B and Fig. 5C show the synthesis based on each unit image data in the image shown in Fig. 5A). Went to show 1) dog state).

If the original image shown in FIG. 5A further changes to the original image shown in FIG. 5D, first, as shown in FIG. 5E, the division "0000 / (no address) / 1" After the unit image data for level n = 0 shows that the pixel values of the 64 pixels are all “1”, the division “001 0/1 1 00/0” is performed as shown in FIG. 5F. Depending on the unit image data for level n = 2, four images The element may be newly defined as a pixel value "0".

 As described above, by using the format according to the present invention, the information of the original image can be efficiently expressed as image data by appropriately combining the unit image data having different division levels. As shown in FIG. 2, in the format according to the present invention, the larger the division level n, the longer the bit length required for the address. For this reason, in the examples described so far, the total bit length of each unit image data differs depending on the division level. However, in practice, it is often more convenient to handle the unit image data with a fixed bit length. As described above, when the unit image data has a fixed length, in the format shown in FIG. 3, the bit length of the division level is fixed, and the bit length of the address and the bit length of the pixel value are not changed. The sum should be fixed length.

Figure 6 shows the address / pixel for each division mode when the bit length of the division level is fixed at 4 bits and the sum of the address length and the data length for the pixel value is fixed at 32 bits. FIG. 6 is a diagram illustrating bit assignment of values. The division level represented by 4 bits is 16 levels with n = 0 to 15, and the display resolution at n = 15 reaches 1 G, which is sufficient for practical use. On the other hand, since the address Z pixel value is 32 bits in both cases, when the division level n is low, the power ⁵ ′ that can secure a sufficient pixel value data length ⁵ ′ and the division level n is high If this happens, it will not be possible to secure a sufficient pixel value data length. For example, in the example of Fig. 6, at the division level n = 0, a sufficient bit length of 32 bits is secured as the pixel value data bits, while at the division level n = 15, the pixel Only two data bits are reserved for values.

However, even if the bit allocation as shown in FIG. 6 is performed, there is no practical problem. In other words, the bit allocation method of “fixing the sum of the address length and the pixel value data length” is the pattern recognition characteristic of human vision. It fits. First, the pixel value data length “32 bits” secured at the division level 0 is sufficient for displaying a color image. For example, suppose that the pixel value of one pixel was represented as 256 gradation values for each of the three primary colors of RGB. In this case, the power ^s that makes it possible to express so-called full color (1,670,000 colors) using the three primary colors of RGB ^s , and the gradation expression of one color only needs to have 8-bit data. To represent a pixel value, it is enough to have 24 bits of data. The “32-bit” pixel value data length described above is sufficient for such a full-color display. On the other hand, the data length “2 bits” secured at the division level 15 can represent only four types of pixel values, but the area displayed using this division level 15 is the entire display area. It is a very small area compared to the screen, and even if you can select only four types of pixel value expression, there is no discomfort when performing pattern recognition by human vision (§ 3 (2) described later) See).

 FIG. 7 is a diagram showing an example of an image expressed by mixing unit image data in various division modes. In this image, the upper left area is division level 1 (data length: 30 bits), the upper right area is division level 2 (data length: 28 bits), and the lower left area is division level 3 (data length: 30 bits). The lower right area is displayed at division level 4 (data length: 24 bits). Here, the larger the block, the longer the allocated data length is, and the more precise color expression is possible. Conversely, the smaller the block, the shorter the allocated data length, It can be seen that the color expression is coarse. This characteristic is consistent with human visual pattern recognition. In other words, the human eye is sensitive to the color expression in a large area colored with the same color, but becomes insensitive to a small area with a small area. Expressions with short data lengths do not cause discomfort.

Note that the unit image data of each division level shown in FIG. It consists of a bit sequence of bits. First, the division level n is recognized by the first four bits, and the next 2 n bits from the beginning of the following 32 bits are the address. The remaining bits are recognized as bits indicating the pixel value. Also, in this example, only 4 bits are used to indicate the division level n, but since the information of this division level is important information that will affect the subsequent address and pixel value if misrecognized. However, in practice, it is preferable to add an error code or repeat the same information twice to express the word with redundancy. In particular, unit image data with a division level of n = 0 and other unit image data with a low division level become unit image data that affects the entire screen or large area of the reproduced image. It is preferred that the method be redundant. '

 § 2. Basic method of generating image data according to the present invention

 In the above §1, the basic format of image data generated by the method according to the present invention and its advantages have been described. A method of driving a display device using image data described in such a format has been described. Is a known method already disclosed in the above-mentioned International Publication No. WO97 / 36927. The intention of the present invention is to generate image data (image data expressed in the format described in §1) for reproducing this original image based on an arbitrary original image.

 For example, based on an original image P 0 defined on a two-dimensional pixel array composed of a total of 64 pixels arranged in 8 rows and 8 columns as shown in FIG. 8, the original image P 0 is reproduced. Consider the case of generating image data for the purpose. Here, for convenience of explanation, the pixel located at the i-th row and j-th column in this two-dimensional pixel array is referred to as a pixel.

Let us call it P (i, j) and let the pixel value of the pixel be Q (i, j). For example, in FIG. 8, the pixel P (1, 1) at the upper left corner is a pixel having a pixel value Q (1, 1), and the pixel P (8, 8) at the lower right corner is a pixel value Q (8, 8) To Pixel.

Now, for the two-dimensional pixel array shown in FIG. 8, if four types of division levels n = 0 to 3 as shown in FIG. 1 are defined, a large number of blocks as shown in FIGS. 9A to 9D can be obtained. Can be defined. That is, at the division level n == 0 shown in FIG. 9A, one block B (0) is defined, and at the division level n = 1 shown in FIG. 9B, four blocks B (00) to B ( 1 1) is defined, the division level n = 2 shown in FIG. 9 C, 1 6 blocks ^{B (0000) ~:? B} (1 1 1 1) force is defined, the division level n shown in FIG. 9 D At = 3, 64 blocks B (000000) ~: B (1 1 1 1 1 1) are defined, and a total of 85 blocks can be defined. Note that, here, blocks having a division level n = 1 or more are denoted by “B (address of the block)”. The image data according to the present invention is composed of unit image data having the format shown in FIG. 3, and this one unit image data corresponds to one block. That is, one of the 85 blocks described above is specified by the division level address in the format shown in FIG. 3, and one unit of image data is There is no other data indicating the pixel value of.

In the present invention, from the coarsest division mode in which the entire two-dimensional pixel array including the original image is regarded as one block to the finest division mode in which each pixel of the two-dimensional pixel array is regarded as one block. The fineness of division s A plurality of different division modes may be defined. For example, if the two-dimensional pixel array including the original image is an eight-row, eight-column pixel array as shown in FIG. 8, the division mode represented by the division level n == 0 (two-dimensional pixel array From the coarsest division mode where the whole is one block) to the division mode indicated by the division level n = 3 (the finest division mode where each pixel of the two-dimensional pixel array is one block), 4 What is necessary is just to define the division mode over the types of division levels. Now, in order to efficiently represent the information of the original image P0 as shown in FIG. 8 by a plurality of unit image data described in the format shown in FIG. There are issues. The first issue is how to determine the pixel value of each block. For example, when generating the unit image data for the block B (0) shown in FIG. 9A, the division level and the address can be uniquely determined (the division level is "00001", and the address is "none". "). In order to determine the pixel value, some algorithm must be defined. The second problem is whether to use unit image data of any of the 85 blocks described above as image data for reproducing the original image P0. As exemplified in §1, image data described in the format according to the present invention generally has higher resolution of the original image than image data described in the bitmap format. Image information can be expressed efficiently. However, in order to be able to efficiently represent the information of the original image, it is necessary to reproduce the original image with as few blocks as possible (in other words, with as few unit image data as possible). is there. For example, as an example of expressing the information of the original image shown in Fig. 4A, Figs. 4B to 4E show examples using a total of 9 blocks (unit image data). In the case of an original image consisting of a two-dimensional pixel array of 8 rows and 8 columns as in the example shown in Fig. 4A, it is possible to define a total of 85 blocks (unit image data) as described above. In the examples shown in FIGS. 4B to 4E, nine sets are extracted from the total of 85 pieces, and are used as image data for reproducing the original image shown in FIG. 4A. The combination of unit image data to reproduce the same original image is not necessarily one, but the final combination depends on which unit image data is to be extracted from a total of 85 unit image data. The content of the image data obtained at the beginning differs.

In the present invention, the first problem described above, that is, how to determine the pixel value of each block As for the point of determination, one pixel on the original image included in each block is determined as a representative pixel for the block, and the pixel value of the representative pixel for each block is determined as the pixel of the block. Value. For example, in the case of block B (0) shown in FIG. 9A, the pixels in the original image included in this block are all 64 pixels constituting the original image P0 shown in FIG. One of the pixels is defined as a representative pixel, and the pixel value of this representative pixel is defined as the pixel value of block B (0). The representative pixel may be any one of the 64 pixels. For example, if the pixel at the upper left corner is defined as the representative pixel, the pixel P (1, 1) shown in FIG. 8 is the representative pixel of the block B (0), and this representative pixel P (1, 1) ) Is defined as the pixel value of block B (0).

On the other hand, in the case of blocks B (00) to B (11) shown in FIG. 9B, the pixels in the original image included in each block constitute a region obtained by dividing the original image P0 shown in FIG. Since there are 16 pixels, one of the 16 pixels is defined as a representative pixel for each block, and the pixel value of this representative pixel is defined as each of the blocks B (0 0) to B ( 1) It is defined as the pixel value of 1). For example, in the case of block B (0 0), 16 pixels P (1, 1) to P (1, 4), P (2, 1) to P (2, 4) shown in FIG. , P (3,1) to P (3,4), P (4,1) to P (4,4) are determined as representative pixels. The representative pixel may be any one of the 16 pixels. Here, the representative pixel is selected based on the following criteria. That is, if the block address is "00", the pixel at the upper left corner is the representative pixel, and if the block address is "01", the pixel at the upper right corner is the representative pixel. If the block address is "10", the pixel at the lower left corner is the representative pixel, and if the block address is "11", the pixel at the lower right corner is the representative pixel. And As a result, pixel P (1, 1) becomes the representative pixel of block B (00), and its pixel value Q (1, 1) becomes the pixel value of block B (00). Pixel P (1, 8) becomes the representative pixel of block B (01), and its pixel value Q (1, 8) becomes the pixel value of block B (01). Similarly, pixel P (8, 1) becomes the representative pixel of block B (10), and its pixel value Q (8, 1) becomes the pixel value of block B (10). In addition, pixels? (8, 8) becomes the representative pixel of block B (1 1), and its pixel value Q (8, 8) becomes the pixel value of block B (1 1).

Similarly, for a block having a division level n = 2 or more, a predetermined pixel on the original image included in each block is determined as a representative pixel for the block, and the pixel value of the representative pixel for each block is It is defined as the pixel value of the block. In this case as well, it does not matter which pixel is selected as the representative pixel. In this case, if the last two digits of the block address are "00", the pixel at the upper left corner is set as the representative pixel. If it is "0 1", the pixel at the upper right corner is the representative pixel, and if it is "10", the pixel at the lower left corner is the representative pixel, and it is "1 1". Then, the representative pixel is selected based on the criterion that the pixel at the lower right corner is the representative pixel. As a result, for example, for block B (0000) shown in Fig. 9C, the last two digits of the address are "00", and the pixel P (1, 1) at the upper left corner in this block is the representative pixel. , And the pixel value Q (1, 1) becomes the pixel value of block B (0000). In addition, for the block B (01 0 1) shown in FIG. 9C, the last two digits of the address are "01", so that the pixel P (1, 8) at the upper right corner in this block becomes the representative pixel, The pixel value Q (1, 8) becomes the pixel value of block B (01 01). Similarly, for block B (10 10) shown in Fig. 9C, the last two digits of the address are "10", so that the pixel P (8, 1) at the lower left corner in this block is the representative pixel. The pixel value Q (8, 1) becomes the pixel value of block B (1010). Furthermore, for the block B (1 1 1 1) shown in FIG. 9C, since the last two digits of the address are "1 1", the right The pixel P (8, 8) in the lower corner becomes the representative pixel, and its pixel value Q (8, 8) becomes the pixel value of block B (1 1 1 1). The blocks B (000 000) to B (1 1 1 1 1 1) shown in FIG. 9D are the same as the individual pixels of the two-dimensional pixel array shown in FIG. The value itself becomes the pixel value of each block.

 Figure 10 shows a total of 85 blocks B (0) to B (1 1 1 1 1 1) defined by the division levels n = 0 to 3 defined for the original image shown in Fig. 8, respectively. This is a table listing the division levels, addresses, and pixel values. The unit plane image data corresponding to each block is data in which the division levels, addresses, and pixel values in this table are arranged as a bit string, as shown in the format of FIG. As described above, in the present invention, in order to solve the first problem described above, that is, the problem of how to determine the pixel value of each block, the original image included in each block is A method is adopted in which one predetermined pixel is defined as a representative pixel for the block, and the pixel value of the representative pixel for each block is defined as the surface element value of the block. The method of determining the representative pixel is effective in simplifying the calculation for obtaining the pixel value. For example, as a method of determining the pixel value of the block B (0) shown in FIG. 9A, the average value of the pixel values of the 64 pixels constituting the original image P0 in FIG. Although it is possible to perform a calculation that obtains, the burden of such a calculation increases as the number of target pixels increases. In particular, when the pixel values are represented by three primary colors such as RGB, the computational burden becomes enormous. If a method of determining any one representative pixel is adopted as in the present invention, no complicated calculation is required at all.

However, the representative pixel selected by the method according to the present invention is not always a pixel representing a large number of pixels belonging to the block. For example, in the original image shown in Fig. 8, only pixel P (1, 1) in the upper left corner shows a red color. If the pixel value of the block B (0) has the pixel value Q (red) and all the other 63 pixels have the pixel value Q (blue) indicating a blue color, the pixel value of the block B (0) has the majority of the pixels. It is more appropriate to adopt the pixel value Q (blue) from the original meaning. If a method of determining the most frequently occurring pixel value as the pixel value of the block is adopted, in such a case, the pixel value of the majority of pixels is used as the pixel value of block B (0). become Q (blue) force ^s adopted is that. However, according to the above-described method, the pixel P (1, 1) at the upper left corner is selected as the representative pixel of the block B (0), so that the pixel value of the block B (0) is the pixel value Q (red). It turns out that. This is because the method of generating image data according to the present invention focuses on reducing the computational load.

In short, in the above-described embodiment of the present invention, when deciding the pixel value of each block, the specific pixel value of each pixel included in the block is determined in order to emphasize the reduction of the calculation load. Instead, the pixel at a specific position in the block is uniquely selected as the representative pixel. Pixel Therefore, as in the example described above, there majority of pixels force in the block B ^(0)? Despite to a blue, a predetermined specific position, i.e., the upper left corner position Since the color was P (1, 1) red, Q (red) indicating the red color is selected as the pixel value of block B (0). Thus, as the pixel values of one block, as the pixel value of the minority was elected, the image data of to reproduce the original image - in performing a process of generating data, any problem force ^s 'The points that do not occur will be described later with specific examples.

In practicing the present invention, for each block, any pixel located in the block may be selected as the representative pixel. However, in the above embodiment, when the last two digits of the block address are “00”, the pixel at the upper left corner is set as the representative pixel, and when the last two digits of the block address are “01”, the pixel at the upper right corner is set. If the pixel is the representative pixel and it is "1 0", the pixel at the lower left corner is the representative pixel and it is "1 1" In this case, the representative pixel is selected based on the criterion that the pixel at the lower right corner is the representative pixel. The reason why the representative pixel is selected based on such criteria is nothing but to simplify the logical operation for selecting the representative pixel. In other words, if the representative pixels are selected based on the above-described criteria, when determining the representative pixels for each block of the division level n = 1 or more, `` the end of the address of the block, the end of the address 2 By adding the same two bits as the digits repeatedly as many times as necessary, an operation is performed to obtain the address of a block having the same size as one pixel that constitutes the original image, and the obtained address is used. The pixel at the position corresponding to the block shown is determined as the representative pixel. "Let's show this in a concrete example.

 For example, in the case of the original image P0 shown in Fig. 8, a block having the same size as one pixel is a block with a division level n = 3 (each block shown in Fig. 9D), and has 6 bits. This is a block indicated by the address. Therefore, for example, in the case of block B (00) shown in Fig. 9B, the same two bits "00" as the last two digits "00" are repeated as many times as necessary at the end of the block address "00". To obtain a 6-bit address by adding, "00" may be repeated twice and added. In this case, the obtained address is "000000", and the pixel P (1, 1) at the position corresponding to the block B (000 000) indicated by this address is selected as the representative pixel.

Similarly, in the case of the block B (10) shown in FIG. 9B, the same two bits "10" as the last two digits "10" are added to the end of the address "10" of the block as many times as necessary. In order to obtain a 6-bit address by repeated addition, "10" may be repeated twice and added. In this case, the obtained address is “10101 0”, and the pixel P (8, 1) at the position corresponding to the block B (1 01 0 10) indicated by this address is selected as the representative pixel. become.

In addition, in the case of the block B (0 1 1 0) shown in FIG. 9C, two bits “1 0” which are the same as the last two digits “1 0” are added to the end of the address “0 1 1 0” of the block. To obtain a 6-bit address by repeatedly adding "" as many times as necessary, "10" need only be added once. In this case, the obtained address is "0 1 1 0 1 0", and the pixel P (4, 5) at the position corresponding to the block B (0 1 1 01 0) indicated by this address ^. It will be selected as the representative pixel.

 In general, the operation of adding a bit to the end is extremely light in calculation load. Therefore, selecting a representative pixel for each block on the basis of the above-described criterion is an efficient operation in the image data generation method according to the present invention. This is extremely effective in converting

 Subsequently, a solution of the present invention for the second problem, that is, the problem of which one is extracted from a large number of definable unit image data will be described. For example, given the original image shown in FIG. 8, a total of 85 blocks as shown in FIG. 10 can be defined, and a total of 85 sets of unit image data can be defined. However, if all the 85 sets of unit image data are output directly as image data and the original image is reproduced, the redundancy becomes extremely high. In order to generate efficient image data, some of the necessary unit image data are extracted from the 85 sets of unit image data, and only the extracted unit image data is used as image data for reproducing the original image. Need to output.

Therefore, in the present invention, a condition that “when a plurality of unit image data indicating pixel values at the same position on a plane exists, the pixel value of the unit image data that is finely divided than the unit image data that is coarsely divided” is used In order to be able to reproduce the original image below, multiple sets of unit image data are extracted from the multiple unit image data, and the extracted unit image data group is referred to as `` an image for reproducing the original image. It is output as "data". Here, specifically, a method of generating a group of unit image data for reproducing the original image P1 based on the original image P1 as shown in FIG. 11 will be considered. The original image P1 shown in FIG. 11 is a pattern painted in green and purple, and is an image composed of an 8 × 8 two-dimensional pixel array, like the original image P0 shown in FIG. is there. As described above, a total of 85 blocks can be defined for this two-dimensional pixel array, and a total of 85 unit image data can be defined. FIGS. 12A to 12D are diagrams showing a total of 85 blocks filled with a color corresponding to each pixel value. The pixel value of each block is determined by selecting a representative pixel based on the above-described criteria.

 For example, for block B (0) shown in FIG. 12A, a green pixel value is given by selecting one green pixel at the upper left corner of FIG. 11 as a representative pixel. For the block B (00) shown in FIG. 12B, a green pixel value is given by selecting one green pixel at the upper left corner of FIG. 11 as a representative pixel. On the other hand, for block B (01) shown in Fig. 12B, the purple pixel value is given by selecting one purple pixel in the upper right corner of Fig. 11 as a representative pixel. For B (10), a green pixel value is given by selecting one green pixel in the lower left corner of Fig. 11 as a representative pixel, and for block B (11), The purple pixel value is given by selecting one purple pixel in the lower right corner of 1 as the representative pixel. For the 16 blocks shown in FIG. 12C and the 64 blocks shown in FIG. 12D, representative pixels are selected on the same basis, and the pixel value of each representative pixel is given. If pixel values for each block are determined in this way, pixel values can be defined by a very simple operation, as described above.

Now, as shown in the table of Fig. 13, seven blocks are extracted from the 85 blocks. By extracting the images shown by these seven blocks, It can be seen that the original image P 1 in Fig. 11 can be reproduced (where Q (Green) and Q (purple) indicate green and purple pixel values, respectively). That is, the block B (0) shown in FIG. 12A and the blocks B (0 1) and B shown in FIG.

(1 1), blocks B (0 100) and B (0 1 01) shown in FIG. 12C, and blocks B (0 1 001 1) and B (0 1 0 100) shown in FIG. 12D. By superimposing a total of seven blocks (all indicated by bold lines), the original image P 1 shown in Fig. 11 is reproduced. However, when stacking seven blocks,

The condition that “when there are a plurality of unit image data forces ⁵ ′ indicating pixel values at the same position on the plane, the pixel values of unit image data that is finely divided is used rather than unit image data that is coarsely divided” is necessary. is there. In other words, if each of these seven blocks is a square paper painted in a predetermined color, the blocks are placed in a predetermined position in order from the large area paper (coarse unit image data). It is sufficient to perform the following processing. For example, if purple paper corresponding to blocks B (01) and B (1 1) shown in Fig. 12B is placed on green paper corresponding to block B (0) shown in Fig. 12A, The right half of the green paper corresponding to block B (0) is hidden. In addition, when a layer of red paper corresponding to blocks B (01 00) and B (01 01) shown in Fig. 12C is placed on top of it, the purple paper corresponding to block B (01) is The upper half is hidden. Eventually, only the paper (finely divided unit image data) placed on top of the top layer will be in a visible state, and "a plurality of unit image data indicating pixel values at the same position on the plane exists." In this case, the pixel values of the finely divided unit image data are used rather than the coarsely divided unit image data. "The original image P1 shown in FIG. 11 is reproduced. However, the process of “laying another paper on one paper” in the above example is executed as a bit data replacement process on an actual display device, so it is usually nsec or This is performed by high-speed processing in units of sec. Therefore, actually, the original image P1 is displayed on the screen of the display device at a speed that cannot be grasped by the naked eye. By the way, in order to extract “image data for reproducing an original image” from a plurality of unit image data as extraction candidates, the following image data extraction processing may be performed. That is, for a plurality of unit image data to be extracted, the process of determining whether to extract unit image data from coarsely divided unit image data to finely divided unit image data is sequentially described as “unit image to be extracted candidate”. If the image in the area indicated by the data is the same as the image reproduced by the already extracted unit image data, it will not be extracted, and if it is not the same, it will be extracted. '' What should be done is.

 Specifically, such a determination is made based on “pixel values of unit image data that are extraction candidates” and “pixel values of unit image data that include the block of the unit image data and are coarsely divided by one stage”. It is sufficient to compare with and judge that if the rain people are the same, they are not extracted, and if they are not the same, they are extracted.

This image data extraction processing is performed from the total of 85 blocks (unit image data) shown in FIGS. 12A to 12D, from the seven blocks shown in the table of FIG. An example of extracting the criterion (unit image data) will be specifically described. As described above, the process of determining whether or not to extract should be performed in order from the coarsely divided unit image data to the finely divided unit image data. In this example, the division shown in FIG. 1 unit image data with level n = 0, 4 unit image data with division level n = = 1 shown in Fig. 12B, 16 unit images with division level n = 2 shown in Fig. 12C This determination processing is executed in the order of the data and the 64 unit image data of the division level n = 3 shown in FIG. 12D. First, the unit image data power ^s corresponding to the block B (0) indicated by the thick line in Fig. 12A is extracted unconditionally as data to be extracted (the division level is coarser). Since there is no unit image data, there is no image that has already been extracted, and of course, it is not the same as the image that has already been extracted.)

Next, it corresponds to the four blocks B (0 0) to B (1 1) shown in FIG. The unit image data is determined, but blocks B (00) and B (10) are reproduced by the already extracted unit image data (unit image data corresponding to block B (0)). Since they are the same as the image (the image in the left half area of block B (0)) (all of them are images filled with green pixels), they will not be extracted. On the other hand, blocks B (01) and B (1 1) indicated by bold lines in the figure are reproduced by the already extracted unit image data (unit image data corresponding to block B (0)). Since it is not the same as the image (the image in the right half area of block B (0)) (one is an image filled with green pixels, while the other is an image filled with purple pixels). These are extracted.

 Next, a determination is made on the unit image data corresponding to the 16 blocks B (0000) to B (1111) shown in FIG. 12C. In this judgment,

,, Only two blocks B (01 00) and B (01 0 1) are determined to be blocks to be extracted. The other 14 blocks are the images reproduced by the three unit image data already extracted (the unit image data corresponding to blocks B (0), B (01), and B (1 1)). Are not extracted because they are the same as Similarly, the judgment is performed on the unit image data corresponding to the 64 blocks B (000000) to B (1 1 1 1 1 1) shown in FIG. 12D. In this determination, only two blocks B (0 100 1 1) and B (01 0100) indicated by thick lines in the figure are determined to be blocks to be extracted. The other 62 blocks are the same as the images reproduced by the three extracted unit image data (unit image data corresponding to blocks B (0), B (01), and B (1 1)). Therefore, it will not be extracted. Thus, as shown in the table of FIG. 13, only seven unit image data are extracted as “image data for reproducing the original image”. In general, the operation for determining whether or not the images in a given area are the same can be performed as a comparison operation between bits, so that a general logic circuit or CPU is used. Can be done at high speed.

When outputting the group of unit image data extracted in this way, it is preferable that the unit image data is output in order from the coarsely divided unit image data to the finely divided unit image data. This is transmitted to the display unit of the unit image data group ^s , and when the reproducibility ^s ' of the original image is performed on the display unit, the synthesis process of the original image is gradually performed at such a speed that the naked eye can confirm it. When presented (for example, when the original image is reproduced by a display device with a slow response speed such as a liquid crystal display device), the unit image data is divided from coarsely divided unit image data to finely divided unit image data. This is because it is better to show them in order. In other words, by presenting in order from a coarse image to a fine image, a blurry and rough image first appears, and gradually the finer partial power ^s ' gradually becomes clearer, and finally the original image is completely reproduced , And the presentation form can be taken. However, on the contrary, it is also possible to present the images in order from a finer image to a coarser image for a special visual effect. In this case, the small part is expressed first, and the presentation form is such that the overall strength ^s ' can be confirmed gradually.

Subsequently, a method of generating a unit image data group for reproducing the original image P2 based on the original image P2 as shown in FIG. 14 will be described. Similar to the example shown in FIG. 11, the original image P 2 shown in FIG. 14 is an image composed of an eight-row, eight-column two-dimensional pixel array ⁵ , only one pixel at the upper left corner is green, and other pixels are green. 63 This is an extreme pattern in which three pixels are purple. Also for such an original image P2, a total of 85 blocks can be defined, and a total of 85 unit image data can be defined. FIGS. 15A to 15D are diagrams in which the 85 blocks in total are filled with a color corresponding to each pixel value. The pixel value of each block is determined by selecting a representative pixel based on the aforementioned criteria.

As described above, this criterion uniquely defines a pixel located at a specific position in each block as a representative pixel, and the pixel value of the representative pixel remains unchanged. Also, it is defined as the pixel value of the block. Therefore, the pixel value defined for each block is not always the pixel value of many pixels belonging to the block. In fact, regarding block B (0) shown in Fig. 15A, the pixel in the upper left corner became the representative pixel despite the fact that the majority of the pixels belonging to this block were purple, so block B (0) The pixel value of is green (Q), which indicates green. As described above, the reason why the representative pixel is uniquely selected based on the position in the block is to obtain the advantage of reducing the calculation load.

 Now, to extract some unit image data necessary to reproduce the original image P2 shown in Fig. 14 from the 85 blocks defined in this way, that is, 85 unit image data, As described above, the processing for determining whether or not to extract in order from the coarsely divided unit image data to the finely divided unit image data is referred to as “the region indicated by the unit image data as the extraction candidate”. If the image in is not the same as the image reproduced by the already extracted unit image data, it will not be extracted, and if it is not the same, it will be extracted. '' Good. First, the unit image data corresponding to the block B (0) indicated by the thick line in FIG. 15A is unconditionally determined as the data to be extracted and extracted.

Subsequently, the judgment power s is determined for the unit image data corresponding to the four blocks B (00) to B (11) shown in FIG. 15B, but the block B (00) has already been extracted. It is the same as the image reproduced by the unit image data (unit image data corresponding to block B (0)) (the image in the upper left quadrant of block B (0)). The image is full), and will not be extracted. On the other hand, blocks B (01), B (1 0), and B (1 1) indicated by bold lines in the figure are not the same as the images reproduced by the already extracted unit image data. These are images filled with green pixels, while the other is filled with purple pixels.) Will be.

 Next, a determination is made on the unit image data corresponding to the 16 blocks B (0000) to B (1 1 1 1) shown in FIG. 15C. In this determination, three blocks B (0001), B (00 10), and B (001 1) indicated by thick lines in the figure are determined to be blocks to be extracted. The other 14 blocks are based on the four unit image data already extracted (unit image data corresponding to blocks B (0), B (0 1), B (1 0), and B (1 1)). Since it is the same as the image to be reproduced, it will not be extracted. Similarly, the judgment power s is determined for the unit image data corresponding to the 64 blocks B (000 000) to B (1 1 1 1 1 1) shown in FIG. 15D. In this determination, three blocks B (00000 1), B (0000 10), and B (00001 1), which are indicated by thick lines in the figure, are determined as blocks to be extracted. The other 61 blocks are not extracted because they are the same as the images reproduced by the seven unit image data already extracted. Thus, as shown in the table of FIG. 16, ten unit image data are extracted as “image data for reproducing the original image”.

 Eventually, the first extracted block B (0) shown in Fig. 15A was an image far apart from the original image P2 shown in Fig. 14, but the 10 units shown in the table in Fig. 16 By sequentially synthesizing the image data, the original image P2 shown in FIG. 14 is reproduced without any trouble. Therefore, there is no problem if the representative pixel for each block is uniquely selected based on the position without considering the pixel value of each pixel.

Finally, another method of generating a unit image data group for reproducing the original image P3 based on the original image P3 as shown in FIG. 17 will be described. The original image P 3 shown in Fig. 17 is an image consisting of a two-dimensional pixel array of 8 rows and 8 columns, as in the previous examples, but a green English letter “I” is added to the white background. The pattern is as shown. Even for such an original image P3, a total of 85 blocks Can be defined, and a total of 85 unit image data can be defined. FIG. 18A to FIG. 18D are diagrams showing a total of 85 blocks filled with a color corresponding to each pixel value. The pixel value of each block is determined by selecting a representative pixel based on the above-described criteria.

As described above, the pixel values defined for each block do not necessarily reflect the pattern of the original image. In fact, in block B (0) shown in FIG. 18A and blocks B (00) to B (11) shown in FIG. 18B, only a white background pattern appears. As in this example, even if the image indicated by the block defined by one division level i and the image indicated by the block defined by the next division level (i + 1) are completely the same. It should be noted that the images indicated by the blocks defined after the upper division level (i + 2) are not necessarily the same. In other words, because the image represented by the block defined at one division level i and the image represented by the block defined at the next division level (i + 1) are completely identical. Therefore, it is not possible to omit the block definition for a higher division level. Now, from the 85 blocks thus defined, that is, 85 unit image data, some unit image data necessary to reproduce the original image P 3 shown in FIG. 17 are extracted. For this purpose, first, the unit image data power ^s corresponding to the block B (0) indicated by a thick line in FIG. 18A is unconditionally determined and extracted as data to be extracted. Next, the unit image data corresponding to the four blocks B (00) to B (11) shown in FIG. 18B is determined, and all of these four blocks have already been extracted. Since it is the same as the image reproduced by the unit image data (unit image data corresponding to block B (0)), it will not be extracted. Next, a determination is made on the unit image data corresponding to the 16 blocks B (0000) to B (1111) shown in FIG. 18C. In this determination, two blocks B (01 1 0) shown by a thick line in FIG, B (1 1 00) ^force?, Extraction should do plot Is determined. Similarly, determination force of the unit image data corresponding to 64 blocks B shown in FIG. 1 8 D (000000) ~B ( 1 1 1 1 1 1)? Is performed. In this judgment, the eight blocks indicated by thick lines in the figure, that is, the blocks B (0001 1 1), B (010010), B (01001 1), and B (0 10 1 10) drawn in FIG. , B (1 0 1 10 1), B (1 1 1 000), B (1 1 1 0 01>, B (1 1 1 100) f, are determined to be blocks to be extracted. As shown in Table 20, twenty-one unit image data is extracted as “image data for reproducing the original image”.

 When the original image is an image composed of an object such as a character, as in the original image P3 shown in FIG. 17, the pattern usually has the object placed on the background. As described above, in the case of the original image indicating “an object arranged on the background”, information on the image of the background portion can be omitted from the image data. For example, the information of the block B (0) shown in the uppermost column of the table in FIG. 20 is information indicating a white background, so that the unit image data corresponding to the block B (0) may be omitted. I don't care. That is, the unit image data corresponding to block B (0) is deleted from the table of FIG. 20, and the remaining 10 unit image data are extracted as “image data for reproducing the original image”, Even if this is transmitted to the display device, the display device handles it on the assumption that “the transmitted image data is the original image to be displayed on a specific background” (for example, 17), the original image P3 shown in FIG. 17 can be reproduced.

§ 3. Examples of ears

 Finally, some modified examples of the image data generation method according to the present invention other than the above-described embodiment will be described.

 (1) Selection method of representative pixel

In the above-described embodiment, the substitution for the block B (0) with the division level n = 0 is performed. As a front pixel, one pixel in the upper left corner is selected ⁵ ′, and any pixel included in the block may be selected as the representative pixel. For example, one pixel at the upper right corner, lower left corner, or lower right corner may be selected as the representative pixel, or a pixel near the center of the block may be selected as the representative pixel. This is the same for blocks having a division level n = 1 or more. However, as described above, for each block with a division level of n = l or more, one of the upper-left corner, upper-right corner, lower-left corner, or lower-right corner depends on the last two digits of the block address. If is used as the representative pixel, the arithmetic processing is greatly simplified.

 () A method of giving a background pixel value to a block with a coarse division mode

 In the embodiments described above, one pixel on the original image included in the block is defined as a representative pixel as the pixel value of each block, and the pixel value of the representative pixel is defined as the pixel value of the block as it is. However, it is not always necessary to give the pixel value of the representative pixel included in the block, but rather a different pixel value that is not the pixel value of the representative pixel. Sometimes it is preferable.

For example, in the original image shown in FIG. 8, if the pixel value of the upper left pixel P (1, 1) is red, the method according to the above-described embodiment uses the block B having the division level n = 0 shown in FIG. The pixel value for (0) is red. Therefore, at the moment when the display device displays the unit image data for the block B (0) shown in FIG. 9A, the entire display screen is displayed in red. Of course, the unit image data for blocks B (00), B (01), B (10), and B (1 1) shown in Fig. 9B is displayed. As the unit image data for the block shown in Fig. 9D and the block shown in Fig. 9D is displayed, the original original image will be displayed, but at least the unit image data for block B (0) will be displayed. The moment the force is displayed, the whole screen will turn red for a moment. In this way, the whole screen is covered with a vivid color like red. This is not desirable depending on the display environment. In particular, in the case of a display device that takes a long time to switch the display, the entire screen ^s remains red, and depending on the display environment, it may be desirable to avoid such a situation as much as possible. In such a situation, a background color such as white or black, which is generally displayed on the entire screen and does not cause any discomfort, should be determined in advance, and the background pixel value indicating this background color should be determined. I do. Then, in the pixel value definition stage, for a block obtained by a division mode that is coarser than a predetermined reference, a predetermined background pixel value is used instead of the pixel value of the representative pixel in the block. What is necessary is just to perform the process defined as the pixel value of the said block.

Specifically, for example, based on the division level n = 1, the blocks obtained by the coarse division mode in which the division level is 1 or less, that is, blocks B (0) and 9B shown in FIG. 9A For the blocks B (00), B (01), B (10), and B (1 1) shown in (1), the background pixel value (for example, the pixel value indicating black) is forcibly given. Then, when the display device displays the unit image data of block B (0), the entire display screen has a black background color, and then the blocks B (00), B (0 1), and B (0) When displaying the unit image data of (10) and B (11), this black state will continue. Eventually, the original image power s is gradually displayed as the unit image data power ^s of the blocks shown in FIG. 9C and the blocks shown in FIG. 9D are displayed more finely. In this way, the unit image data for blocks B (0), B (00), B (01), B (10), and B (1 1) is based on the original information transmission of the original image. Although it cannot play a role, it can play a presentation effect that the original image gradually appears from a black background color. Of course, such a background pixel value is forcibly given only to blocks obtained by the coarse division mode, so that the block shown in FIG. 9C and the block shown in FIG. The professionals obtained by the fine division mode It is possible to faithfully reproduce the original image by using the unit image data of the block.

 (3) Extraction judgment of unit image data

 In the above-described embodiment, when extracting some unit image data to be “image data for reproducing the original image” from all the unit image data theoretically definable for the original image, , The pixel value of the unit image data that is a candidate for extraction and the pixel value of the unit image data that includes the block of the unit image data and whose division is coarse by one stage are compared. It is determined that it is not extracted in some cases, and extracted if it is not the same. For example, when determining whether to extract unit image data corresponding to block B (01 1 0) in FIG. 18C, one of the four blocks shown in FIG. Block containing block B (01 1 0): B (01) is selected as the comparison target, and it is determined that the pixel is not extracted if the pixel values of the rain block are the same, and it is extracted if the pixel values are not the same. . As described above, in general, it is sufficient to make a determination as a comparison target for a block whose division is only one step coarse, but in some cases, a block whose division is two or more steps coarse may be used as a comparison target. For example, the unit image data for the four blocks B (01) to B (11) shown in FIG. 18B is data that is determined to be “no extraction”, and the division level n = All the unit image data of 1 will not be extracted. In such a case, the determination on the block B (0110) with the division level n = 2 may be performed by comparing the block B (0) with the division level n = 0, which is roughly divided by two stages.

As described above, the basic principle of the unit image data extraction determination is as follows: `` Is the image in the area indicated by the unit image data that is an extraction candidate the same as the image reproduced by the already extracted unit image data? However, the term “images are the same” in the present invention does not necessarily mean that the pixel values are completely the same, but that “the image can be regarded as the same by human senses”. One "Correspondence" or "matching within a range that cannot be distinguished in the display device to which the image data is to be transmitted". For example, when a pixel value is represented by 8-bit data, a pixel value in a range from 0 to 254 is defined. In this case, for example, if the difference between the pixel values is within ± 5, it may be determined that they are the same. It is also possible to change such an allowable range for each division level. Specifically, the higher the division level, the smaller the area of the block, so that even if the permissible range is increased, the human eyes are less likely to feel uncomfortable. In the present invention, “the same image” does not necessarily need to be determined based on “matching colors”, and may be determined based on “matching brightness” in some cases. For example, if the display device to which image data is to be transmitted is a device that can only display monochrome gradations, it is better to make decisions based on “brightness match” instead of “color match” on the original image. It is preferable from the viewpoint of reproducing the sound.

 () Coding of unit image data group

In the embodiments described so far, a large number of blocks having a hierarchical structure are defined. For example, the block B (010 1 01) at the upper right corner obtained by the division mode of the division level n = 3 shown in FIG. 9D is changed from the block B (0) as the root to the block B (0) —B (01) — B (0 101) → B (010 1 01), which can be reached by drawing the branches on this hierarchical structure. Therefore, a total of 85 blocks B (0) to B (1 1 1 1 1 1) shown as a list in FIG. 10 form a single tree structure with the block B (0) as the root. Will be. That is, one trunk B (0) extending from the root is divided into four branches B (00) to B (1 1), and each branch has four middle branches B (0000) to B (1 1). The tree structure ^{5 is} formed, which is divided into four branches B (0 00000) to B (1 1 1 1 1 1). Here, if these blocks with a tree structure Of the first group located on the branch branched from the first group (in other words, the block group of the first group consisting of the first block and the blocks included in the first block among many blocks having a hierarchical structure) ) And the second group of blocks located on the branch branched from the second location (in other words, from the block in the second block and the blocks contained in it, among many blocks having a hierarchical structure). If the unit image data of the block group of the first group matches the unit image data of the block group of the second group, the first group An identification code is defined for each set of unit image data for the block group, and this identification code is used instead of each unit image data for the block group of the second group. If to appear, it is efficient eliminates the need to extract overlapping the same data.

For example, the same object is drawn with the same positional relationship in the upper right quadrant and the lower right quadrant of the original image shown in Fig. 8, and the block B (01) shown in Fig. 9B And the pattern in B (1 1) are exactly the same. In this case, the pixel value given for block B (0 1) and the pixel value given for block B (1 1) are, of course, the same, and the finer division levels contained in these blocks For each block, the same pixel value is given to both blocks at the corresponding positions. For example, with respect to the block of the division level n = 2 shown in FIG. 9C, the same pixel value is given to the block B (0 100) and the block B (1 100), and the block B (01 01) is The same pixel value is given to block B (1 1 01), and the same pixel value is given to block B (01 1 0) and block B (1 1 10). 11) and block B (111) have the same pixel value. The same applies to the block of the division level n = 3 shown in FIG. 9D. In other words, the block indicated by the address whose first two bits are s “01” and the block whose first two bits are indicated by the address “1 1” are the addresses 3 If the bits below are the same, the same pixel value will be given.

Therefore, the block indicated by the address of the first two bits “0 1” (block B (01) and the block obtained in a finer division manner included therein) is defined as the first group of blocks, and the first two bits are assigned. If the block indicated by the address “1 1” (block B (1 1) and the block obtained by the finer division mode included therein) is the block group of the second group, the first group The difference between the set of unit image data for the block group of the group and the set of unit image data for the block group of the second group is only the first two bits of the address. Therefore, for example, an identification code such as “group (0 1)” is defined in the set of unit image data for the block group of the first group, and each unit for the block group of the second group is defined. If this identification code “group (01)” is extracted instead of image data, it is not necessary to extract the same data redundantly. In other words, of the 85 unit image data shown in FIG. 10, the unit image data indicated by the address having the leading 2-bit power ^s “11” does not need to be extracted. It suffices to extract the code that is “group (0 1)”. However, in order to perform replacement using such an identification code, information that clarifies which part has been replaced with which part is added to the finally extracted data (data transmitted to the display device). It is necessary to keep. That is, in the case of the above example, the unit image data indicated by the address of “the leading two bits s“ 1 1 ”is replaced with a set of unit image data indicated by the identification code“ group (0 1) ”. The information should be included in the unit image data that is ultimately extracted.

However, in order to perform such replacement using the identification code, the display device that performs display processing using the unit image data thus extracted can also handle such identification code. It is necessary to prepare such a function. Above In the case of the above example, the lower right quadrant area to be displayed using “unit image data whose first two bits are indicated by“ 1 1 ”” is identified by the identification code “group (01)”. The display processing power that displays the same image as the image displayed in the upper right quadrant by the unit image data indicated (that is, the unit image data indicated by the address whose first two bits are “10”) You should be able to do it.

After all, the above-mentioned method encodes the image displayed in the upper right quadrant with an identification code of “group (01)”, and places the “group ( It can be said that this method reduces the data capacity required for reproducing the original image by describing that the image strength indicated by the identification code “0 1)” is displayed. If a block definition with a tree structure is performed ⁵

The image indicated by the identification code “group (0 1)” is an image indicated by a block group located on a branch branched from the block B (01). In the above example, the image coded in this way is used only in one place (block B (11)). Of course, if the same image exists in three or more places on the original image, It is possible to use coded images in multiple places.

In short, in the method described here, if a single original image contains multiple parts with the same pattern, an identification code is given to the image data indicating the one pattern part and the code is coded. This is a method based on the basic idea that the identification code is used instead of the image data for the picture part in other places. Such a method can be applied to two or more original images. That is, when generating the first image data for reproducing the first original image and the second image data for reproducing the second original image, a portion of the first original image is generated. If a part of the same pattern is included in a part of the second original image, an identification code is also given to the image data indicating the pattern part of the first original image, and the image data is coded. If this is done, it becomes possible to substitute this identification code for the same picture portion on the second original image instead of the image data.

 In other words, the first group of blocks including the first block of the many blocks defined for the first original image and the blocks included therein, and the second group of the second original image Regarding a second block of a large number of defined blocks and a second group of blocks including blocks included therein, each unit image data of the first group of blocks, If the unit image data of the block group of the group matches the unit image data, an identification code is defined for the set of unit image data of the block group of the first group, and the second original image In the image data extraction stage, the identification code may be extracted instead of each unit image data of the block group of the second group.

 Of course, this method can also be applied to three or more original images, and the coded part on the first original image can be used for all subsequent original images. It is possible. When presenting a moving image as described later, a plurality of original images are sequentially displayed along the time axis. At this time, an identification code is assigned to a part of the original image displayed earlier in time. If the identification code is used to represent the same part on the original image to be displayed later using this identification code, it is very effective in reducing the total data amount for presenting a moving image. is there. (5) Omission of image data extraction stage

In the embodiments described so far, a predetermined unit image is selected from unit image data of all blocks obtained by all division modes defined in the division level definition stage according to the conditions for reproducing the original image. Although the image data extraction step of extracting data and outputting the extracted unit image data group as “image data for reproducing the original image” was performed, the image data extraction step may be omitted. Absent. For example, when the division levels n = 0 to 3 as shown in Fig. 9A to Fig. 9D are defined for the original image shown in Fig. 8, as shown in Fig. 10, all 85 blocks are defined. Thus, a total of 85 sets of unit image data can be prepared. In the above-described embodiment, an image data extracting step of extracting some unit image data necessary for reproducing the original image from the 85 sets of unit image data is performed based on a predetermined algorithm. However, from the viewpoint of creating image data, it is not always necessary to perform the image data extraction step. Of course, these 85 sets of unit image data are redundant data for reproducing the original image, and not all of these data are required for display. However, from the viewpoint of selecting necessary unit image data at the time of displaying an image, it is not necessary to select at the time of generating image data.

Specifically, for example, consider the case where some image data is distributed as a web page. In such a case, it is preferable to prepare the above-mentioned 85 sets of unit image data in the Sano device for distributing the image data. At present, client devices that receive web pages and display them on the screen are not limited to personal computers, but also cover a wide variety of game devices and mobile phones. Therefore, the display screen resolution of these client devices also varies, and some devices can only display low-resolution images even if high-resolution image information is transmitted. Alternatively, high-resolution image information may not be needed depending on the application, such as when recognizing the presence or absence of an object. In such a case, it is sufficient to select and transmit only the unit image data up to a certain division level. Conversely, there may be times when you want to receive as much redundant image information as possible because of the poor communication environment. By preparing the above-mentioned 85 sets of unit image data in the server device, the necessary unit image data can be selected each time according to the performance, application, communication environment, etc. of the client device. Distribution. (6) Support for original images of any size

Embodiments described so far, which had been subjected to the creation of image data based on the original image constituted by a two-dimensional pixel array of 8 rows and 8 columns, generally vertical contact and laterally the 2 ⁿ respectively If the original image consists of a square pixel array with pixels arranged (for example, a pixel array of 16 rows and 16 columns, 32 rows and 32 columns, and 64 rows and 64 columns), the description has been given so far. mosquitoes that carry out the definition of exactly the same split level you Yopi § de-less as the ^embodiment? can.

On the other hand, the size of the original image does not always match the square element array in which 2 "pixels are arranged vertically and horizontally. When handling such an original image of an arbitrary size, As a two-dimensional pixel array including the original image, a square pixel array in which 2 ⁿ pixels are arranged vertically and horizontally is defined, and a part of this square pixel array (displaying the original image Image data to be displayed on a display screen with a vertical / horizontal aspect ratio of 1: 1. For example, in the example shown in Fig. 21, in the square pixel array of 16 rows and 16 columns, only the area of 9 rows and 16 columns surrounded by the thick line is the effective area for displaying the original image. It is defined as E. For high-definition television, etc., Asupeku Ratio of lateral 9: having the desired § scan Bae transfected ratio in a part of the 1 6 and going on the force thus square pixel arrangement arranged 2 ⁿ pixels in the vertical and horizontal By defining the effective area E, the method for generating image data according to the present invention can be applied to an original image having an arbitrary aspect ratio. For example, if the block address information for each division level as shown in Fig. 22 is prepared, the effective area E shown in Fig. 21 can be defined by using this information. For the pixels in the sky area EX that constitute the outer part of the effective area E, some pixel value Q (φ) indicating that the pixel is a pixel in the sky area EX is defined. In short, in effect, this is effectively A plurality of sets of unit image data necessary for indicating the image information in the area E may be extracted, and the extracted unit image data group may be output as “image data for reproducing the original image”. .

 In the example shown in FIG. 21, the pixel values of the blocks B (1 0) and B (1 1) at the division level n = 1 are both Q (φ) if defined by the method according to the above-described embodiment. (The representative pixel of block Β (1 0) is the lower left corner, and the representative pixel of Β (1 1) is the lower right corner). In this case, the unit image data corresponding to blocks Β (10) and Β (11) is data indicating that no original image exists in this area. On the other hand, if one pixel at the upper left corner is selected as the representative pixel of block Β (10) and one pixel at the upper right corner is selected as the representative pixel of block Β (11), the pixel values of these blocks are This is the actual pixel value of one pixel on the image. If the screen of the display device that reproduces the original image has the same size as the effective area Ε, the original image can be normally reproduced even if the representative pixel is determined by any of the methods described above. . However, if a display device with a screen that is wider in the vertical direction than the effective area Ε is used, the empty area Ε X will be displayed as a margin, so the representative pixel must be determined by the former method described above. For example, if the representative pixel is determined by the latter method described above, the margin part is the power to be displayed as a screen corresponding to the pixel value Q (φ). The image will be presented.

To avoid such a phenomenon, it is considered several ways force ^s. The first method is to extract only unit image data corresponding to blocks that fall within the effective area E (blocks that do not enter the empty area Ex) when extracting unit image data. However, this method has a problem that the number of unit image data to be extracted increases, and the total data amount increases. For example, in order to represent the original image in which all of the effective area E is painted green, it is usually sufficient to prepare only one set of unit image data for the block with the division level n = 0, but When extracting only the unit image data corresponding to the blocks that fall within the effective area E, it is necessary to prepare the unit image data for 18 sets of blocks shown in Fig. 22 ^. Occurs.

 The second method is a method in which information as shown in FIG. 22 is transmitted to the display device in advance, and the size of the effective area E is defined. In this case, at the stage of preparing the unit image data, it is not necessary to impose the restrictions as in the first method described above. The display device performs a process of generating a reproduced image of the original image based on the received unit image data. At this time, a process may be performed such that a portion other than the effective area E is not displayed. Alternatively, among the display elements constituting individual pixels on the screen of the display device, display elements existing in a portion other than the effective area E may be set to an OFF state in a hardware manner.

 In the case of a display device with a function that allows the viewer to perform a zoom operation, the original image is enlarged or reduced, so the effective area E is also enlarged or reduced. It is necessary to perform processing that treats it as having been lost. Further, FIG. 22 shows the information indicating the effective area E. Conversely, information indicating the empty area EX may be prepared and transmitted to the display device side. In this case, the display device may perform a process of removing a portion of the empty region EX from the image reproduced by the received unit image data and displaying the image.

 (7) Image data generation for moving image display

The examples described so far are all examples in which a still image is used as an original image and image data for reproducing the original image is generated, but the present invention provides an image data for reproducing a moving image. It can also be used for generation. More specifically, a plurality of still images having a fixed presentation order may be prepared to compose a moving image, and a unit image data group may be generated using each still image as an original image. If the generated unit image data group is played back in still image units in order, a moving image can be played. When generating a unit image data group for each still image, if the unit image data group is generated using the difference image from the preceding still image as the original image, the overall data amount is reduced. can do. For example, by preparing a moving image original image Ml as shown in FIG. 23A and a moving image original image M2 as shown in FIG. 23B, the original images Ml and M2 are displayed in this order. Consider the case where a video is presented. This video is a video in which the green letter “I” displayed on a purple background moves to the left. In this case, first, image data is generated using the moving image original image Ml as the first original image. Subsequently, a difference image D of the original moving images Ml and M2 as shown in Fig. 24 is created. This difference image D is an image in which the pixels at the same position in the original moving image Ml and M2 are compared, and only the pixels having a difference are defined as the pixel values of the original moving image M2. There is no difference and the pixel value indicating the background is defined. Then, using the difference image D as an original image, image data is created by the method of the present invention. However, for the background image part, the creation of image data may be omitted. Specifically, in the case of the difference image D shown in FIG. 24, it is sufficient that a unit image data group capable of indicating green pixels and purple pixels can be prepared.

 Thus, if the image data indicating the difference image D is transmitted after the image data indicating the moving image original image Ml, the display device first displays the moving image original image Ml, and then displays the difference image D If the processing of rewriting only the image portion other than the background indicated by is performed, the presentation of a moving image becomes possible.

As described above, first, image data representing the original image in the initial state is created and transmitted to the display device, and thereafter, a difference image D representing only the changed portion is created and transmitted, whereby the power of change can be obtained. Since image data needs to be transmitted only when s' occurs, the total amount of data to be transmitted can be reduced. In addition, since the calculation for obtaining the difference image D can be performed by a simple comparison operation between bits, the calculation burden is reduced. (8) Application to comparison operation between images

 The present invention is directed to a method for generating image data for reproducing the original image based on a predetermined original image. The image data created by the method according to the present invention is a comparison operation between images. It can also be used for As described above, since the comparison operation between the unit image data can be performed by a simple bit comparison operation, high-speed operation is possible. This makes it possible to compare the size and color of an arbitrary object included in the image, and for example, it is possible to perform processing such as comparing the shapes of two products at high speed. If applied to image recognition, high-speed recognition processing becomes possible.

 Also, if the product is photographed with the product placed in an environment with a background having a predetermined color, and only the unit image data for expressing pixels having a color other than the background color is extracted, the background is removed. Image data indicating only products can be generated. Also, it is possible to perform processing to combine with a different background. Furthermore, if a code is added to the unit image data group indicating a specific object, the unit image data group expressing the specific object can be searched by this code, and the background image can be included. It is possible to construct an image database for the object images that do not exist.

(9) Unit image data format

In Fig. 6, the division level, address, and pixel value are each represented by a bit, and the bit length of the division level is fixed, and the bit length of the address and the bit length of the pixel value are represented. The format of the unit image data with fixed length is the sum of However, in implementing the present invention, the format of the unit image data is arbitrary, and does not necessarily need to be the format shown in FIG. Each bit length can be arbitrarily set, and the bit itself indicating the division level can be omitted. For example, if it is possible to distinguish between the bit portion indicating the address and the bit portion indicating the pixel value data in some way (for example, by sandwiching them with a separator consisting of a unique bit string), Ad Even less length Ya data length of any length, it is possible ^force? Be distinctively recognize on the bit portion indicating a bit portion and a pixel value data indicating the address of the specific bit sequence. If the bit part indicating the address can be recognized, the division level can be recognized based on the bit length. As described above, many methods are known for expressing addresses and data as a bit string by distinguishing them from each other according to various communication methods. In implementing the present invention, the format of unit image data is as follows. It may be something like this. INDUSTRIAL APPLICABILITY The method for generating image data according to the present invention can be widely used for generating image data for driving various display devices. It can be used for generating image data suitable for driving a large-sized display device such as the like.

Claims

Scope of request

1. A method of generating image data for reproducing an original image based on an original image represented by a two-dimensional pixel array in which a large number of pixels having predetermined pixel values are arranged,

A plurality of division modes for dividing the two-dimensional pixel array into a plurality of blocks are defined. At this time, each division mode is made to have a different fineness force ^s , and each division mode indicates a division fineness. A division level definition stage that can be identified by level;

 An address definition stage for defining an address for indicating each block for each individual division mode;

 A pixel value defining step of defining a predetermined pixel on the original image included in each block as a representative pixel of the block, and defining a pixel value of a representative pixel of each block as a pixel value of the block;

 As image data for indicating an image in a specific block determined by a specific division level and a specific address, unit image data having the specific division level, the specific address, and the pixel value of the specific block, Defining the image data to define,

 A method for generating image data, comprising:

2. In the method for generating image data according to claim 1,

In the division level defining step, so as to define a plurality of split levels Yuku different fineness force stepwise division, the k-th division level and than this one step dividing ^month? Finer the (k + 1) -th With respect to the division level, one block obtained by the division mode indicated by the k-th division level is further divided into a plurality of blocks, whereby a plurality of blocks according to the division mode indicated by the (k + 1) -th division level are obtained. B The lock is obtained, so that it is included in the block obtained by the partitioning mode indicated by the (k + 1) th partitioning level, the partitioning mode indicated by the kth partitioning level. A method for generating image data, wherein a large number of blocks having a hierarchical structure corresponding to a plurality of division levels are defined.

3. In the method for generating image data according to claim 2,

 From the unit image data of all blocks obtained by all division modes defined in the division level definition stage, predetermined unit image data is extracted according to the conditions when reproducing the original image, and the extracted unit An image data generating method, further comprising an image data extracting step of outputting an image data group as “image data for reproducing the original image”.

4. The method for generating image data according to claim 3,

 The condition for reproducing the original image is as follows: `` If there is a plurality of unit image data indicating pixel values at the same position on the plane, the pixel value of the unit image data that is finer than the coarser unit image data is "Use"

 In the image data extracting step, a plurality of sets of unit image data are extracted from the plurality of unit image data so that the original image can be reproduced under the above-mentioned conditions, and the extracted unit image data group is referred to as “ And outputting the image data as image data for reproducing the original image.

5. In the method for generating image data according to claim 4,

In the image data extraction stage, for a plurality of unit image data as extraction candidates, a process of determining whether to extract the unit image data in order from coarsely divided unit image data to finely divided unit image data is referred to as “extraction candidate”. According to the unit image data If the image in the area indicated by is not the same as the image reproduced by the already extracted unit image data, it is not extracted, and if it is not the same, it is extracted. '' A method of generating image data.

6. The method for generating image data according to claim 5,

 When judging a specific unit image data to be an extraction candidate, the “pixel value of the specific unit image data” and “the block of the specific unit image data are included and the division is coarse by one stage A pixel value of the unit image data ”and are compared, and if both are not the same, the extraction is not performed, and if they are not the same, the extraction is performed.

7. The method for generating image data according to any one of claims 3 to 6, wherein, in the image data extracting step, the first block and the blocks included in the first block among a plurality of blocks having a hierarchical structure. Of the first group of blocks, and the second group of blocks consisting of the second block and the blocks included therein. And when the unit image data of the second group of block groups matches each other, an identification code is defined in a set of each unit image data of the first group of block groups, and A method for generating image data, wherein the identification code is extracted instead of each unit image data for two groups of blocks.

8. The method for generating image data according to any one of claims 3 to 6, wherein the first image data for reproducing the first original image and the second image for reproducing the second original image. When generating data, the first block of a number of blocks defined for the first original image and the blocks contained therein A first group of blocks consisting of a second group of blocks defined for the second original image and a second group of blocks consisting of blocks contained therein, When each unit image data of one block group and each unit image data of the block group of the second group match, the unit image data of each block image of the first group is A method for generating image data, wherein an identification code is defined for a set, and the identification code is extracted instead of each unit image data for the blocks of the second group. 9. The method for generating image data according to any one of claims 3 to 8, wherein, in the division level defining step, the coarsest division mode in which the entire two-dimensional pixel array including the original image is regarded as one block, A method for generating image data, comprising defining a plurality of division modes having different division details up to the finest division mode in which individual pixels of a two-dimensional pixel array are regarded as one block.

10. The method for generating image data according to claim 9, wherein

As a division mode indicated by a division level n, a division mode in which 2 ^{2 n} blocks are obtained by dividing a two-dimensional pixel array vertically and horizontally by 2 "is defined, and n = 1, 2,, ..., i, A total of Ν division modes are defined for 態 様, 正, and a square pixel array in which 2 ^η pixels are arranged vertically and horizontally is used as a two-dimensional pixel array that encompasses the original image. How to generate image data.

11. The method for generating image data according to claim 10, wherein

In the address definition stage, for each of the four blocks obtained in the division mode represented by the division level η = 1, an address consisting of 2 bits of 0 0, 0 1, 10 and 11 is defined. In the division mode shown by n = i Of the 2 ^{2 1} blocks obtained in this manner, 0 0 is added below the address indicating 2 ^{2 (i-} "blocks" obtained in the division mode represented by the division level n = (i-1). A method for generating image data, characterized by defining an address to which any one of the following is added.

12. In the method for generating image data according to claim 11,

 When determining the representative pixel for each block of division level 1 or higher, the same two bits as the last two digits of the address are repeatedly added to the end of the address of the block as many times as necessary. An operation for obtaining an address of a block having a size equivalent to one pixel constituting the original image is performed, and a pixel at a position corresponding to the block indicated by the obtained address is determined as a representative pixel. How to generate image data.

13. In the method for generating image data according to claim 10,

 As a condition for reproducing the original image, set the condition of "display on a display screen whose vertical and horizontal aspect ratio is not 1: 1".

In the image data extraction step, the aspect ratio of the display screen is partially included in a square pixel array including 22 ⁿ pixels so that the original image can be reproduced under the above conditions. Defines an effective area that has an image, extracts a plurality of sets of unit image data that are necessary to indicate image information in the effective area, and describes the extracted unit image data group as `` reproduce the original image A method for generating image data, characterized in that the image data is output as “image data to be processed”.

14. The image data generation method according to any one of claims 1 to 13, wherein the division level, the address, and the pixel value are each represented by a bit, and the bit length of the division level is fixed. Together with the bits that make up the address. Image data, characterized in that the sum of the data length and the bit length of the pixel value is fixed, and the bit length of the address can be recognized based on the division level. Method.

15. The method for generating image data according to any one of claims 1 to 13, wherein the unit image data includes at least a bit indicating an address and a bit indicating a pixel value, and indicates an information level indicating a division level. A method for generating image data, wherein the image data is represented as a bit length of a bit indicating the address.

16. The image data generation method according to any one of claims 1 to 15, wherein in the image data extraction step, the extracted unit image data is converted from coarsely divided unit image data to finely divided unit image data. A method for generating image data, characterized in that image data is output in order.

17. The image data generation method according to any one of claims 1 to 16, wherein a plurality of still images in a predetermined presentation order are prepared to form a moving image, and each still image is converted to an original image. A method for generating image data, characterized by generating image data for reproducing a moving image by generating a group of unit image data.

18. The method for generating image data according to claim 17, wherein

 A method for generating image data, characterized in that when generating a unit image data group for each still image, a unit image data group is generated using a difference image from a preceding still image as an original image.

19. The method for generating image data according to any one of claims 1 to 18, wherein In the pixel value defining step, for a block obtained by a division mode coarser than a predetermined criterion, a predetermined background pixel value is replaced with a predetermined background pixel value instead of the pixel value of the representative pixel in the block. A method for generating image data characterized by being defined as pixel values.

20. A program for executing the method of generating image data according to any one of claims 1 to 19 using a computer, or a computer-readable recording medium storing the program.