JP2000222420A - Image processor, its method, and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to transform a certain color label string data into another color label string data by removing a color label indicating a background color from color label string data of an original image without requiring the original image. SOLUTION: A color label string related to an image to be transformed and its attribute information are inputted from a color label string input part 2010. A transformation attribute specification part 2020 specifies attribute information related to labeling in a transforming color space. A color space matching part 2030 matches the color substitution of the inputted color label string based on the inputted attribute information and the specified attribute information. A background color judging part 2035 judges a control color of an input image as a background color based on the inputted color label string. A block division matching part 2040 matches a block removing the background color with a transformed block and a color label string output part 2050 generates and outputs the transformed color label string.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus and method which divides an image into a plurality of blocks, and expresses a color of each block with a label according to a feature amount, and a storage medium.

[0002]

2. Description of the Related Art In recent years, for example, various techniques for searching for images have been proposed. Among them, there is a technique of inputting an image (hereinafter, referred to as a query image) to be used for a search and searching for an image closest to the image from a group of registered images. In this technique, in a technique which has been put to practical use to some extent, an image is divided into a plurality of blocks, and a histogram of each color is taken to determine a most frequent color for each block. on the other hand,
The color, for example, RGB color space is divided, and a label is given to each cell (small color space piece) as shown in FIG. Then, a color label to which the most frequent color of each block corresponds is determined for all the blocks, and this is set as a color label string.

[0003] The color label string is registered in a database, and the color label string is similarly detected from the query image.
By performing pattern matching between them, an image closest to the query image is searched. Such techniques include, for example, JP-A-8-249349 and JP-A-10-260983.

[0004]

However, the smaller the number of block divisions of an image (the coarser the division), the more the dominant color (in many cases, the background color) in the entire image becomes the majority of the color label string. It occupies and loses the color of objects in the background. In this case, in particular, in the above-described image search device, only the search using the background color can be performed, and a search result more than necessary is output, and it is difficult to reach the target image.

The present invention has been made in view of such a problem, and does not require an original image, and removes a background color which is dominantly present in the original image, and separates another color from the color label string data. It is an object of the present invention to provide an image processing apparatus, method, and storage medium that enable conversion into label string data.

[0006]

To solve this problem, for example, an image processing apparatus according to the present invention has the following arrangement. That is, an image processing apparatus that divides an image into a plurality of blocks, and attaches a color label according to a feature amount to each block to express the image with color label string data, wherein the conversion-source color label string data And input means for inputting the attribute information thereof, determining means for determining the background color of the original image based on the color label string input from the input means, and attribute information about the converted color label string data. A designation means for designating, and a color label data of the conversion source, based on the attribute information inputted by the input means and the attribute information designated by the designation means, and further based on the judgment result of the judgment means, Conversion means for converting to column data.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A preferred embodiment according to the present invention will be described below with reference to the accompanying drawings.

FIG. 1 is a block diagram showing a control configuration of the color label string conversion device of the present embodiment. In the figure, 10
Reference numeral 1 denotes a CPU which executes various controls in the image search device of the present embodiment. Reference numeral 102 denotes a ROM, which stores a boot program executed when the apparatus is started and various data. A RAM 103 stores a control program for processing by the CPU 101 and has a CP
It provides a work area when the U101 executes various controls. A keyboard 104 and a mouse 105 provide various input operation environments for the user.

Reference numeral 106 denotes an external storage device, which comprises a hard disk, a floppy disk, a CD-ROM, or the like. The external storage device 106 stores an OS and various image processing applications.
By being loaded into the M103 and executed, it functions as a color label string conversion device. Reference numeral 107 denotes a display, which is configured by a display or the like, and displays a result or the like to a user. A network interface 108 enables communication with each device on the network. It is assumed that a plurality of devices such as an image search device and an image recognition device exist on the network. A bus 111 connects the above components.

In the above configuration, the external storage device 1
Reference numeral 06 may be replaced by one located on the network.

FIG. 2 is a block diagram showing a functional configuration of the color label string conversion device of the present embodiment. Each component described below is mainly based on the CPU 101 when shown in correspondence with FIG. That is, most of the components are realized by software executed by the CPU 101 (the program is stored in an external storage device and is loaded into the RAM 103 and executed).

Referring to FIG. 1, reference numeral 2010 denotes a color label string input unit, which receives a color label written in a format shown in FIG. 6 from the image recognition apparatus of the conversion source via the external storage device 106 or the network interface 108. This is for inputting column information. Reference numeral 2020 denotes a conversion attribute specifying unit for specifying an attribute of a conversion destination of the color label string. Reference numeral 2030 denotes a color space matching unit that matches a color space between a conversion source and a conversion destination. A background color determination unit 2035 determines a color label to which a dominant color (background color) in an image belongs. Reference numeral 2040 denotes a block division matching unit that matches the block divisions of the conversion source and the conversion destination. Reference numeral 2050 denotes a color label string output unit, which outputs the converted color label string to the image search apparatus via the external storage device 106 or the network interface 108 or directly.

An operation example of the color label string conversion apparatus of the present embodiment having the above-described configuration will be described below.

The position of the color label converting apparatus according to the embodiment is as follows. For example, an image (query image) to be searched is read from an image feature extracting apparatus to generate a color label sequence, and the image feature extracting apparatus reads the image. When supplying the supplied color label string as a search condition to the image search device,
It is easy to understand if it is considered as a device that converts the image into a device that matches the label of the color space used in the image search device.

FIG. 6 shows an example of a color label string format according to the present embodiment. In this format, the color space (color space) is the name of the color space used in the feature extraction stage of the image search device, for example, RGB space, L
^* u ^* v ^* space and the like are stored. The color space division number is the number of color space divisions for labeling. The block division number stores how much the image is divided in the horizontal direction (x direction) and the vertical direction (y direction). The label string stores the actual color label string itself. FIG. 7 shows the number of divisions 27 (= 3) in the RGB space.
This is an example of dividing into 3 × 3) and labeling.

FIGS. 3 to 5 show the procedure of the color label string conversion processing according to the present embodiment, which is executed by the CPU 101. FIG. Hereinafter, a process of converting the color label string according to the above will be described. In the embodiment, the attribute information on the color labels of the conversion destination and the conversion source is assumed to be stored in advance. For example, the attribute information can be obtained by requesting a device to be converted via a network. Is also good.

First, in step S3000, the color label sequence input unit 2010 inputs a color label sequence to be converted. Next, in step S3005, the attribute of the conversion destination is specified by the conversion attribute specifying unit 2020. The attributes to be specified are attributes (color space, number of color space divisions, number of block divisions) relating to the color labels shown in the format of FIG. Next, in step S3010, the color space matching unit 2030 compares the conversion source color space division number with the conversion destination color space division number, and compares the conversion source color space with the conversion destination color space to determine whether they are different. judge. If they are different, in step S3020, the color space matching unit 2030 corrects the color label mismatch caused by the difference in color space. This processing will be described in detail with reference to FIG.

First, in step S5010, the intermediate color of the cell (the color at the center of the cell) corresponding to each label of the conversion source is determined. In the example of FIG. 7, since the cells corresponding to the respective labels are cubes, the colors of the eight vertices (R ₀ , G ₀ , B ₀ ), (R ₁ , G ₁ , B ₁ ) are provided for each cell. ,
(R ₂ , G ₂ , B ₂ ), (R ₃ , G ₃ , B ₃ ), (R ₄ , G ₄ ,
B ₄ ), (R ₅ , G ₅ , B ₅ ), (R ₆ , G ₆ , B ₆ ), (R ₇ ,
G ₇ , B ₇ ) may be averaged.

R _mid = 1/8 (R ₀ + R ₁ + R ₂ + R ₃ + R ₄ + R ₅ + R ₆ + R ₇ ) Equation (1) G _mid = 1/8 (G ₀ + G ₁ + G ₂ + G ₃ + G ₄ + G ₅₎ + G ₆ + C ₇ ) Formula (2) B _mid = 1/8 (B ₀ + B ₁ + B ₂ + B ₃ + B ₄ + B ₅ + B ₆ + B ₇ ) Formula (3) In the above formula, mid indicates an intermediate color. ing.

Next, in step S5020, it is determined whether or not the conversion source and destination color spaces are different. If they are different, the intermediate color corresponding to each color label obtained by the above equation is converted into a value in the color space of the conversion destination. For example, the conversion source is RGB space and the conversion destination is L ^* u ^* v ^*
If it is a space, it can be converted by the following conversion formulas (4) to (11). X = 0.62R + 0.17G + 0.18B Equation (4) Y = 0.31R + 0.59G + 0.11B Equation (5) Z = 0.066G + 1.02B Equation (6) u ′ = 4X / (X + 15Y + 3Z) Equation (7) V ′ = 9Y / (X + 15Y + 3Z) Equation (8) L ^* = 25 (100Y) ^1/3 −16 Equation (9) u ^* = 13L ^* (u′−0.201) Equation (10) v ^* = 13L ^* (V′−0.461) Expression (11) Here, R, G, and B are R _mid , G for each cell obtained earlier.
_mid and B _mid .

Next, in step S5040, the coordinates of the intermediate color corresponding to the color label of the conversion source are mapped on the color space of the conversion destination, and it is determined which cell in the color space of the conversion destination contains this. By determining the color label, the color label of the conversion source is corrected. At this time, for example, the color space is the RGB space for both the conversion source and the conversion destination, the conversion source color space division number is 3 (R direction) × 3 (G direction) × 3 (B direction) = 27, and the conversion destination color When the number of divisions of the space is 6 (R direction) × 6 (G direction) × 6 (B direction) = 216, which is twice as large in each direction, the intermediate color of the conversion source is the cubic cell of the conversion destination. At the top,
Unable to determine color label. Step S5050
In step S5060, a cell in a specific direction is set as a corresponding cell, and the color label of the cell is adopted. For example, a cell close to black where R = 0, G = 0, and B = 0 among cells near the position of the intermediate color may always be adopted. The above is the details of the processing for correcting the label mismatch caused by the difference in the color space.

Next, the processing of the background color judgment unit 2035 will be described.

In step S3021 of FIG. 3, the histograms of the entire color label sequence of the conversion source are totaled. next,
In step S3022, it is determined whether the first (most frequent) color label occupies the majority. If the first color label indicates a majority, it is determined that the color indicated by the label is the dominant color for the entire image, and in step S3023, the first color label is set to the background. Register as a color label. Registration destination is, for example, R
A predetermined address position of the AM 103 is set. If the first-ranked label does not occupy the majority, it is not set as the background color label. However, label 0 is registered at the same address position as an initial value. That is, in a later process, if the value stored at this address position is "0", it is determined that the background color is not registered, and if a value other than "0" is stored, Determines that the label corresponding to that value is the background color.

Next, the processing in the block division matching unit 2040 will be described. Here, for simplicity of explanation, as an example, as shown in FIG. 9, the number of block divisions of the conversion source is 16 (x direction) × 16 (y direction) (the block division shown by the thin line), and The case where the number of block divisions is 4 (x direction) × 4 (y direction) and there are (block divisions shown by thick lines) will be described. That is, in this case, at the conversion destination, the blocks of the conversion source are grouped by 4 (x direction) × 4 (y direction).

In step S3030, histograms of 4 × 4 color labels of the source block corresponding to each of the destination blocks are totaled. For example, in the conversion destination block of interest, if the block of the color label of the conversion source is as shown in FIG. 10, if the histogram is taken, 11 color labels 1, 3 color labels 20, and 3 color labels 8 is two. Next, in step S3040, the top color label in the histogram of this color label is
It is determined whether it is the majority. In the example described above, the first place is the color label 1, which is eleven, which is the majority. At this time,
In step S3091, it is determined whether the number one color label is a background color label. Whether the color is the background color may be compared with the value registered and registered as described above.

If the first color label has a majority and is not the background color, the flow advances to step S3100 to use the first color label as it is for the destination color label. For example, in the state of FIG. 10, the first place is the color label 1, which occupies the majority, and if the background color registered is other than the color label 1, the first place is the color label. Is converted into a color label according to the attribute of the color space of the conversion destination.

If the first color label is the background color (if the first color label matches the color label set and registered as the background color), the process proceeds to step S3092, and proceeds to the second step S3092. It is determined whether or not the second color label can be determined as the destination label. For example,
In the state of FIG. 10, if the color label 1 is registered as the background color, the second place is the color label 20 (the second frequency), and the same number of other color labels do not exist. In such a case, the process proceeds to step S3093, and the second-rank color label is used for the color label according to the attribute of the color space of the conversion destination.

If there are color labels with the same number of frequencies in the second place or lower, a specific color label among them is adopted in step S3094. As a specific color label, for example, there is a method of selecting a color label having a lower color label from the second or lower color labels. If there are three or more second-level color labels, those having the shortest distance among the colors may be put together and selected based on the number.

In any case, if one color label has been determined, the flow advances to step S3110 to determine whether or not the processing has been performed for all the blocks.
It will return to 3030.

On the other hand, if the color label at the top of the target block does not occupy the majority, the process proceeds from step S3040 to step S3050.

For example, when the target block is in the state shown in FIG. 11, the color label histogram shows six color labels 1, five color labels 2, and three color labels 20.
Since there are two color labels 8, the color label 1 in the first place does not occupy the majority.

In this case, in step S3050, a process is performed to combine labels that can be approximated. In this method, for example, an original image having a different color label due to the presence of light and shade is regarded as approximately the same color. As a method of determining which color label and a color label are considered to be approximately the same color, a method of grouping adjacent cells in a color space, a combination of color labels that can be approximated manually, and a method of approximating A method of creating a table of possible color labels may be considered. Here, an example of the former method is shown.
In FIG. 11, the color labels 1 and 2 are cells adjacent to each other in the color space field as shown in FIG. At this time, the color label is combined with the color label with higher frequency. That is, there are 11 color labels 1 (here, if the frequency is the same, for example, a color label with a small number,
It only needs to match a specific color label). In this way, it is determined whether or not all the color labels can be approximated, and when the color labels can be approximated, a process of combining color labels with higher frequency is performed.

By this processing, in the example of FIG. 11, the color label histogram shows that 11 color labels 1 and 3 color labels 20
And two color labels 8. Next, step S3060
, It is determined whether the first-ranked color label can be determined. In the example of FIG. 11, eleven color labels 1 rank first. Then, the aforementioned steps S3091, S3092,
By performing the processing of S3093, S3094, and S3100,
The destination color label is determined in consideration of the background color label.

As shown in FIG. 12A, (the block of interest is the center block of the bold line.
The corresponding original image is shown in FIG. 12B. If the color label histogram is eight and the color label 20 is eight, the process proceeds to step S3060.
, It is not possible to determine the number one color label.

Therefore, in step S3070, a block around the block of interest is added to the histogram of the color label in consideration of the block around the block of interest. This is to determine which color label is more dominant from the surrounding color labels. In the example of FIG. 12A, as can be seen from FIG. 12B of the corresponding original image, the color label 20 is more dominant. Here, considering, for example, two blocks around the transformation source of the target transformation destination block (the area shown by the thick dotted line in FIG. 12A), the color label histogram has 52 color labels 20, 52 There are 12 labels 1. In the next step S3080, it is determined whether or not the top color label can be determined.
In the example of FIG. 12A, the number of the color labels 20 is 52 and the color label 20 is the first place. Then, steps S3091, S309 described above are performed.
2. The processing of S3093, S3094, and S3100 is performed to determine the destination color label in consideration of the background color label.

If the first place cannot be determined in step S3080, that is, if the processing cannot be determined even in the processing from step S3030 to step S3080, the following is performed.

For example, as shown in FIG.
If it is the block at the center of the bold line and the corresponding original image is (b) in FIG.
At 90, the specific block of the conversion source is set as the color block of the conversion destination. As the specific block, there is a method of adopting a block at a specific position such as the upper left of the block of interest. If the conversion destination block of interest in step S3070 is a block located at the end of the image, there is no portion where the corresponding block around the conversion source protrudes from the image area. At this time, an unused color label, for example, 0 is assigned to a non-existing peripheral block, and if it is counted, it has no effect.

The above processing is repeated until it is determined in step S3110 that all the conversion destination blocks have been determined. When the processing for all the blocks is completed, in step S3120, the color label string output unit 2050 outputs the color label string to the conversion destination.

In the above-described embodiment, as shown in FIG. 6, the format of the color label string includes the attributes of the color space, the number of color space divisions, and the number of block divisions. Is not only a number, but in RGB space, R direction, G direction, B direction
Of course, more detailed attributes such as designation by the number of divisions in each direction may be given, or other attributes such as an attribute relating to a color palette may be added.

Further, in the above embodiment, in step S3050, when color labels that can be approximated are grouped together, color labels with higher frequency are adopted. However, weighting can be performed with the frequency of each color label to approximate. Determine the average color between the intermediate colors of the color label (for example,
If color label 1 has three frequencies and color label 2 has two frequencies, and these can be approximated, then R _mid = 1/5 (3R ₁
+ 2R ₂ ), G _mid = 1/5 (3G ₁ + 2G ₂ ), B _mid =
It is obtained from 1/5 (3B ₁ + 2B ₂ ). There is also a method of adopting the color label of the cell to which it belongs.

Further, in the above embodiment, step S
At 3021, histograms are taken for all the color label strings of the conversion source. That is, a dominant color label was obtained for the entire image. However, it is also possible to obtain a dominant color label from the conversion source color label sequence for a sufficiently large area around the block of interest at the conversion destination. At this time, it is possible to determine a dominant color label for each block of interest at the conversion destination.

Further, in the above-described embodiment, an application example particularly in an image retrieval apparatus has been described. However, an image processing apparatus which divides an image into a plurality of blocks and expresses a color in each block with a label in accordance with a feature amount. If there is, it goes without saying that the utilization method is not particularly limited.

In the above-described embodiment, a format in which a color label string extracted in another image processing apparatus is used has been described. However, in one image processing apparatus, for example, a division smaller than a predetermined block division may be used. It is also conceivable that a color label string is obtained by using the above formula, and this is converted into a predetermined block division.

Further, even if the present invention is applied to a system including a plurality of devices such as a host computer, an interface device, a reader, and a printer, a device including one device (eg, a copying machine, a facsimile machine, etc.)
May be applied.

An object of the present invention is to supply a storage medium storing software program codes for realizing the functions of the above-described embodiment (processing corresponding to FIGS. 3 to 5 and FIG. 8) to a system or an apparatus. Needless to say, the present invention is also achieved when the computer (or CPU or MPU) of the system or the apparatus reads out and executes the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the function of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, hard disk, optical disk, magneto-optical disk, CD-ROM, C
DR, magnetic tape, nonvolatile memory card, ROM
Etc. can be used.

When the computer executes the readout program, not only the functions of the above-described embodiment are realized, but also an OS (Operating System) running on the computer based on instructions of the program code. It goes without saying that a case where the functions of the above-described embodiments are implemented by performing some or all of the actual processing, and the processing performs the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instructions of the program code, It goes without saying that the CPU included in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

As described above, according to the present embodiment,
For example, when searching for an image similar to the image read by the image reader, attribute information (color space name, number of divisions of color space, number of divisions of image) when labeling the conversion source image, and conversion Even if the attribute information of the destination (the search device having the image database) is different, the color label of the conversion destination is generated from the original again. The conversion can be directly performed from the color label of the conversion source, and the image can be searched. In addition, although the background color of an image is an important color for the image, the background color can be removed when searching for an image, so that it is possible to search for a characteristic portion of the image, and it is also possible to narrow down efficiently. .

[0051]

As described above, according to the present invention, it is possible to convert color label string data to another color label string data without requiring an original image.

[Brief description of the drawings]

FIG. 1 is a block diagram of a color label string conversion device according to an embodiment.

FIG. 2 is a functional block diagram of the color label string conversion device of the embodiment.

FIG. 3 is a flowchart illustrating a procedure of a color label conversion process according to the embodiment.

FIG. 4 is a flowchart illustrating a procedure of a color label conversion process according to the embodiment.

FIG. 5 is a flowchart illustrating a procedure of a color label conversion process according to the embodiment.

FIG. 6 is a diagram illustrating an example of a color label string format in the embodiment.

FIG. 7 is a diagram illustrating an example in which an RGB color space is divided and labeled.

FIG. 8 is a flowchart illustrating a detailed procedure of a color space matching process.

FIG. 9 is a diagram illustrating an example of image block division according to the embodiment.

FIG. 10 is a diagram illustrating an example of a conversion source block corresponding to a conversion destination block of interest in the embodiment.

FIG. 11 is a diagram illustrating an example of a conversion source block corresponding to a conversion destination block of interest in the embodiment;

FIG. 12 is a diagram illustrating an example of a case where a destination color label is determined with reference to surrounding blocks according to the embodiment.

FIG. 13 is a diagram illustrating an example in which a destination color label cannot be determined by referring to surrounding blocks in the embodiment;

Claims (15)

[Claims] 1. An image processing apparatus which divides an image into a plurality of blocks and assigns a color label according to a feature amount to each block to express the image by color label string data, wherein the conversion source color Input means for inputting label string data and attribute information thereof; determining means for determining a background color of an original image based on the color label string input from the input means; Specifying means for specifying attribute information; converting the color label data of the conversion source on the basis of the attribute information input by the input means and the attribute information specified by the specifying means; An image processing apparatus comprising: a conversion unit configured to convert the data into label color string data. 2. The attribute information includes a type of a color space, a number of labels corresponding to the number of small color space pieces constituting the color space, and a number of divisions of an image. 2. The image processing device according to claim 1. 3. The conversion means: means for obtaining an intermediate color of each color label in the color space of the conversion source excluding the color label of the background color determined by the determination means; and converting the obtained intermediate color into the color space of the conversion destination. 2. The image processing apparatus according to claim 1, further comprising: means for obtaining a color label of a conversion destination by mapping the color label. 4. A first block size according to the number of divisions of the source image is a second block size according to a number of divisions of the destination image.
If the size of the block is an integer fraction of the block size, a histogram of the labels of the blocks of the first block size included in the block of the second block size is taken, and the most frequent label is set as a representative label. 2. The image processing apparatus according to claim 1, wherein a destination label based on the label is assigned. 5. The representative label has a condition that it has a majority of the number of blocks of the first block size included in the block of the second block size and excludes a background color. The image processing device according to claim 4. 6. The method according to claim 4, wherein when the most frequent label cannot be specified, the color labels assigned to positions that are similar in color space are identified with each other. Image processing device. 7. The image processing apparatus according to claim 6, wherein the position approximated in a color space indicates a shortest distance in a color space. 8. If the color labels are less than the majority even if they are identified with each other, a representative label is determined by labels attached to blocks in a predetermined range exceeding a block of a second block size of interest. The image processing apparatus according to claim 6, wherein: 9. When a representative label cannot be determined by a label attached to a block in a predetermined range exceeding the block of the second block size, a label of a block at a predetermined position is determined as a representative label. An image processing apparatus according to claim 8. 10. When the type of the color space of the conversion source and the type of the color space of the conversion destination are different, the conversion unit determines the color in the conversion destination color space from the color information represented by the label of the conversion source color space. The image processing apparatus according to claim 1, further comprising a calculating unit that calculates information. 11. An image processing method for dividing an image into a plurality of blocks and assigning a color label according to a feature amount to each block to express the image by color label string data, the image processing method comprising: An input step of inputting label string data and its attribute information; a determining step of determining a background color of an original image based on the color label string input from the input step; A specifying step of specifying attribute information, and converting the color label data of the conversion source based on the attribute information input in the input step and the attribute information specified in the specifying step, and further based on the determination result of the determining step. A conversion step of converting the data into label color string data. 12. An image processing apparatus that reads and executes a computer to divide an image into a plurality of blocks, and attaches a color label to each block according to a feature amount to express the image by color label string data. A storage medium storing a functioning program code, comprising: input means for inputting the color label string data of the conversion source and its attribute information; and Determining means for determining the background color; specifying means for specifying attribute information on the converted color label string data; and specifying the color label data to be converted by the attribute information input by the input means and the specifying means. Based on the attribute information and the determination result of the determination step, a program code that functions as a conversion unit that converts the data into destination color label string data is stored. Storage medium. 13. An image is divided into N1 blocks, and each block is assigned one of labels defined to each of M1 small space groups constituting a first color space representing the image. Is divided into N2 blocks, and a second label having a label defined for each of the M2 small space groups forming the second color space for each block. An image processing apparatus for converting into a column data, a determining unit that determines a background color of an image based on a label of the N1 blocks, and a first unit included in one block indicated by the second label column data. If there are multiple blocks of label column data,
Means for determining a representative label based on a group of labels in a first color space assigned to the plurality of blocks, excluding a color label of a background color determined by the determining means; and Converting means for converting the color information in one color space into color information in the second color space; determining which of the labels assigned to the second color space the converted color information belongs to; Means for labeling a subsequent block. 14. An image is divided into N1 blocks, and each block is assigned one of labels defined to each of M1 small space groups constituting a first color space representing the image. Is divided into N2 blocks, and a second label having a label defined for each of the M2 small space groups forming the second color space for each block. An image processing method for converting into column data, a determining step of determining a background color of an image based on a label of the N1 blocks, and a first step included in one block indicated by the second label column data. If there are multiple blocks of label column data,
Excluding the color label of the background color determined in the determination step, determining a representative label based on a label group in the first color space allocated to the plurality of blocks; and determining the representative label represented by the determined label. A conversion step of converting color information in one color space into color information in the second color space; and determining which of the labels assigned to the second color space the converted color information belongs to. A step of labeling a subsequent block. 15. An image is divided into N1 blocks, and each block is assigned one of labels defined to each of M1 small space groups constituting a first color space representing the image. Is divided into N2 blocks, and a second label having a label defined for each of the M2 small space groups forming the second color space for each block. A storage medium for storing a program code functioning as an image processing device for converting into column data, wherein a program code of a determination step of determining a background color of an image based on a label of the N1 blocks; When there are a plurality of blocks of the first label column data included in one block indicated by the column data,
A program code for determining a representative label based on a group of labels in the first color space assigned to the plurality of blocks, excluding a color label of a background color determined in the determination step; A program code of a conversion step of converting the color information in the first color space into color information in the second color space, and the converted color information belongs to a label assigned to the second color space. A storage medium for storing a program code of a step of determining whether or not the converted block is a label of the converted block.