JP2003338936A - Image data compression equipment and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide image data compression equipment which can prevent deterioration of quality of printed letters which is to be caused by non-reciprocal compression, by avoiding a matter that image data exceed capacity of a storage device in an apparatus wherein margin of capacity of the storage device is small. <P>SOLUTION: This equipment is provided with a partial image extracting means 5 wherein original image data are divided into a plurality of partial regions and image data of each of the partial regions are extracted; a luminance value calculating means 6 for calculating a luminance value regarding the respective picture element data of the partial region; a section calculating means 7 wherein a maximum value and a minimum value of luminance value of the respective picture element data of the partial region are obtained, and a quantized section is output in which a section between the maximum value and the minimum value are divided into a plurality of sections; a central value deciding means 8 for deciding a central value of picture element data belonging to each of the quantized sections; and a quantizing means 11 for quantizing the respective picture element data by making all picture element data belonging to the quantized section a central value in the quantized section, in each of the quantized sections. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data compression apparatus and an image data compression method, and in particular,
The present invention relates to an image data compression apparatus and an image data compression method suitable for use in a device such as a printer or a copier having a small memory capacity.

[0002]

2. Description of the Related Art Conventionally, as a data compression method for compressing image data, a reversible data compression method using entropy coding such as Huffman coding and arithmetic coding, and JP
A lossy data compression method using band compression such as the EG method is known.

[0003]

However, in the conventional lossless data compression method such as entropy coding,
The compression rate depends on the redundancy of the original image.
Therefore, if the image to be compressed is not specified, the code amount of the compressed image data cannot be predicted. Therefore, in a device such as a printer or a copying machine having a small memory capacity margin, the data amount of the compressed image data exceeds the memory capacity of the device, and there is a drawback that unprintable state may occur. Was.

Further, in the conventional irreversible data compression method such as the JPEG method, high-frequency components in the frequency space of the image are omitted, so that in the image data in which the outline of the figure such as a character or line drawing is important, There is a drawback that the print quality is degraded. That is, by removing the high-frequency component, there is a drawback that mosquito noise is generated in a portion where the pixel value changes abruptly, such as the contour portion of the figure in the image, and it becomes difficult to read characters and line drawings.

The present invention solves the above-mentioned conventional problems. Even in a device having a small memory capacity margin, it is possible to avoid the situation where the image data exceeds the memory capacity, and at the same time, the outline information of the image is displayed. Leave exactly,
An object of the present invention is to provide an image data compression device and an image data compression method capable of preventing deterioration of print quality due to lossy compression.

[0006]

In order to solve the above problems, an image data compression apparatus according to the present invention divides original image data into a plurality of partial areas and extracts partial image data of each partial area. A brightness value calculating means for calculating a brightness value for each pixel data of the partial area extracted by the partial image extracting means; and a maximum value of the brightness value of each pixel data of the partial area output from the brightness value calculating means, Section calculating means for obtaining a minimum value and outputting a quantization section obtained by dividing the section between the maximum value and the minimum value into a plurality of sections, and a representative for determining a representative value of pixel data belonging to each of the quantization sections A value determining means, and a quantizing means for quantizing each pixel data by setting all pixel data belonging to the quantization interval as the representative value in the quantization interval in each of the quantization intervals. The constitution with.

With this configuration, even in a device having a small memory capacity margin, it is possible to avoid the situation where the image data exceeds the memory capacity, and at the same time, the information of the contour of the image is accurately left and the lossy compression is performed. It is possible to provide an image data compression device capable of preventing deterioration of print quality.

In the image data compression method of the present invention, the original image data is divided into a plurality of partial areas and the pixel data of each partial area is extracted, and the partial image extraction step is performed. A brightness value calculation step of calculating a brightness value for each pixel data of the region,
A section calculation step of obtaining a maximum value and a minimum value of the brightness value of each pixel data of the partial area, and outputting a quantization section obtained by dividing a section between the maximum value and the minimum value into a plurality of sections, A representative value determining step of determining a representative value of pixel data belonging to a quantization section; and, in each of the quantization sections, all pixel data belonging to the quantization section is set to the representative value in the quantization section. And a quantizing step for quantizing each pixel data.

With this configuration, even in a device having a small memory capacity, it is possible to avoid the situation where the image data exceeds the memory capacity, and to accurately leave the contour information of the image and to perform the lossy compression. It is possible to provide an image data compression method capable of preventing deterioration of print quality.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An image data compression apparatus according to claim 1 of the present invention divides original image data into a plurality of partial areas, and extracts partial pixel data of each partial area. Brightness value calculating means for calculating the brightness value for each pixel data of the partial area extracted by the image extracting means, and the maximum and minimum values of the brightness value of each pixel data of the partial area output from the brightness value calculating means, Interval calculation means for outputting a quantization interval obtained by dividing an interval between the maximum value and the minimum value into a plurality of intervals, a representative value determination means for determining a representative value of pixel data belonging to each quantization interval, and each quantization In the section
All pixel data belonging to the quantization section is configured to include a quantizing unit that quantizes each pixel data by setting the representative value in the quantization section as a representative value. The following action can be obtained.

First, the partial image extracting means divides the original image data into partial areas and extracts pixel data for each partial area. Next, the brightness value calculation means calculates a brightness value for each pixel data in the partial area. Next, the section calculation means obtains the maximum value and the minimum value among the brightness values of each pixel data in the partial area. Then, the section between the maximum value and the minimum value is divided into a plurality of sections, and each section is set as a quantization section. Note that the reason why the threshold value is decided by paying particular attention to the "brightness value" is because it has the greatest effect on human vision. Next, the representative value determining means determines the representative value of the pixel data in the partial region whose luminance value is in each quantization section. Next, the quantizing means quantizes each pixel data by using as a representative value all of the pixel data of each partial area whose luminance value is within each quantization interval.

By the above operation, the information amount of the pixel data of each partial area of the original image data is reduced, so that the image data can be compressed. Specifically, for example, when the quantization interval is 2 intervals, the partial area of the RGB color pixel data is a square area of 4 × 4 pixels, and the pixel value of each color of each pixel data is 8 bits, The information amount of data is 48 bytes, but by binarizing the square area, the information amount is reduced to 8 bytes and becomes 1/6. Therefore, the code amount after compression can be theoretically derived from the code amount of the original image for any image, and the size of the image data after compression can be predicted, so that the margin of the memory capacity is small. Even in a device such as a printer or a copying machine, it is possible to avoid the situation where the image data exceeds the memory capacity. Further, by adopting a means of quantizing the interval between the maximum value and the minimum value of the brightness value in the partial area, instead of deleting the high frequency component, for example, the outline portion of the line of a line drawing or a character image. As described above, the contour information can be accurately left even in a portion where the brightness changes abruptly. Therefore, it is possible to avoid deterioration of print quality due to reduction of the information amount of the original image (that is, lossy compression).

Here, "original image data" means image data before data compression processing. The type of “original image” is not particularly specified, and various images such as a monochrome image (gray scale), RGB color image, CMYK color image, etc. are used. The “partial area” means an area included in the entire image, and the shape and size of the area are not particularly specified. Therefore, it is possible to consider an area divided by a plane filling figure such as a rectangular area or a hexagonal area, or an area divided corresponding to the contour of the image. "Pixel data" refers to the data of each pixel that makes up the image data,
For example, if it is an RGB color image, R (red), G
A set of intensity data for each color of (green) and B (blue). “Luminance value” refers to the brightness value of each pixel, for example,
In the case of an RGB color image, the brightness value Y is Y = 0.2
It is calculated by a predetermined calculation formula such as 99R + 0.587G + 0.114B. The number of divisions by which the section calculation unit divides the section is not particularly specified, and it is determined according to the characteristics of the image to be compressed, such as 2 sections, 3 sections, 4 sections, or the purpose of the device using the image data compression apparatus. . Further, the method for the section calculation means to divide the maximum value and the minimum value of the brightness value is not particularly specified, and may be equal division or unequal division. The “representative value” refers to a value of one pixel data representative of all the pixel data in one quantization section. For example, in the case of an RGB color image, the representative value is (R, G, B) intensity data set. The method by which the representative value determination means determines these representative values is not particularly specified. For example, when the number of quantization intervals is two, when determining the representative value of each quantization interval, For example, the method of using the maximum value or the minimum value of the brightness values of each pixel data in the section as a representative value, the median value or the average value or the mode value of the brightness values of the pixel data in the section, For example, a method of using pixel data having a certain brightness value as a representative value is adopted. The term "quantization" used here means that the luminance value is divided into several sections, and the pixel data having the luminance value in each section are grouped with the same numerical value (for example, in the case of an RGB color image, (R, G,
This is represented by the set of numerical values in B)).

According to a second aspect of the present invention, in the image data compression apparatus according to the first aspect, the section calculating means equally divides the section between the maximum value and the minimum value into two quantization sections. The quantization means binarizes each pixel data of the partial region into 1-bit data into pixel data of a quantization section having a large luminance value and pixel data of a quantization section having a small luminance value. The binarized image data is converted into a representative value of pixel data in a quantization section having a large luminance value and a representative value of pixel data in a quantization section having a small luminance value. Is obtained.

After the representative value deciding means decides, for each pixel data in the partial region, a representative value belonging to a quantization section having a large luminance value and a representative value belonging to a quantization section having a small luminance value, The quantizing means binarizes each pixel data of the partial area into 1-bit data for the pixel data of the quantization section having a large luminance value and the pixel data of the quantization section having a small luminance value. The data is converted into a representative value of pixel data in a quantization section having a large luminance value and a representative value of pixel data in a quantization section having a small luminance value. As a result, the code amount after the conversion by the quantizing means is (the number of pixels in the partial area) × (1 bit) + (the number of colors) ×
(Bit number of one pixel of one color). Therefore, the code amount after the data compression processing can be efficiently reduced.

Here, "binarization" means replacing with two values. Specifically, pixel data belonging to a quantization interval having a large luminance value is a set of one value (for example,
(A set of (R, G, B)), which means replacing pixel data belonging to a quantization interval having a small luminance value with another set of values.

According to a third aspect of the invention, there is provided the image data compression apparatus according to the second aspect, wherein the representative value determining means is a pixel data for a quantization section having a large luminance value among the pixel data of the partial area. The representative value of is the value of the pixel data having the maximum brightness value in the partial area, and the representative value of the pixel data for the quantization section having the smaller brightness value in the pixel data of the partial area is the minimum brightness value in the partial area. With this configuration, the pixel data in the quantization section having a large luminance value among the pixel data in the partial area is all pixel data having the maximum luminance value in the partial area. And the pixel data in the quantization section having a small luminance value is the value of the pixel data having the smallest luminance value in the partial area. Therefore, especially for an image composed of characters and line drawings, the contour of the line is emphasized, a large amount of contour information can be left, and the deterioration of the image quality of the image data after data compression can be prevented.

The invention according to claim 4 is the invention according to claim 1 or 2.
In the image data compression device described in (1), the representative value determining means determines a value of the representative value in each quantization section based on the histogram generating means for generating a histogram of the luminance value of the pixel data of the partial area and the histogram. The representative value extracting means and the representative value extracting means are provided, and with this configuration, the histogram generating means generates a histogram of the brightness values of each pixel of the partial region, and according to the distribution of the brightness values in the histogram. , Determine the typical value. Therefore, the optimum representative value can be determined according to the color change state of the image in the partial area of the original image, and the deterioration of the image quality due to the data compression can be minimized.

Here, the histogram generating means appropriately determines the section width of the histogram in accordance with the number of pixels in the partial area. For example, when the partial area has a size of 4 × 4 = 16 pixels, the number of sections is 1/2 to 1 times the number of pixels (8 to 16 sections).
The degree. The method for the representative value extracting means to determine the representative value from the histogram is not particularly limited, but an example of using two quantization intervals is, for example, as follows.

When the histogram has a shape having two maximum values (peak) of large and small, the image of the area is likely to include a contour portion, and therefore the brightness value is set to leave a large amount of contour information. The representative value of the pixel data in the large quantization section is the pixel data with the largest luminance value among all the pixel data belonging to the quantization section with the large luminance value, and the representative value of the pixel data in the quantization section with the small luminance value. The value is the pixel data having the smallest luminance value among all the pixel data belonging to the quantization section having the smaller luminance value.

Further, when the shape of the histogram is a gentle distribution having one maximum value (a distribution in which the distribution value of the histogram is a predetermined value or more), the color of the image of the partial region continuously changes. Since the image is likely to be blurred, the representative value of the pixel data in the quantization section with a large luminance value is all the pixels belonging to the quantization section with a large luminance value in order to smooth the color change. The pixel data having the brightness of the median of the brightness values of the data, and the representative value of the pixel data of the quantization interval having the low brightness value is the median value of the brightness values of all the pixel data belonging to the quantization intervals having the low brightness value. Pixel data having a luminance of.

When the shape of the histogram is a steep distribution having one maximum value (a distribution in which the variance value of the histogram is less than or equal to a predetermined value), the image of the partial area is an image with almost no color change. Therefore, in order to minimize the color change, the representative value of the pixel data in the quantization section with a large luminance value is the luminance value of all the pixel data belonging to the quantization section with a large luminance value. The pixel data having the minimum value of luminance is represented, and the representative value of the pixel data in the quantization section having the small luminance value has the luminance of the maximum value of the luminance values of all the pixel data belonging to the quantization section having the small luminance value. This is pixel data.

The invention described in claim 5 is the invention according to claims 1 to 4.
The image data compression apparatus according to any one of the items, wherein the partial image extraction means has a partial area size setting means for setting a size of a partial area for dividing the original image data, With this configuration, it is possible to freely set the resolution at the time of compression by adjusting the area size according to the fineness of the image to be compressed, and the data is compressed at the optimum resolution in conformity with the original image to be compressed. It becomes possible.

According to a sixth aspect of the present invention, there is provided a partial image extracting step of dividing original image data into a plurality of partial areas and extracting pixel data of each partial area, and a partial image extracting step. A brightness value calculation step of calculating a brightness value for each pixel data of the extracted partial area, a maximum value and a minimum value of the brightness value of each pixel data of the partial area are obtained, and a section between the maximum value and the minimum value is divided into a plurality of sections. A section calculation step of outputting a quantized section divided into sections,
By the representative value determination step of determining the representative value of the pixel data belonging to each quantization interval, and by setting all the pixel data belonging to the quantization interval in each quantization interval as the representative value in the quantization interval. Quantization step for quantizing each pixel data, and with this configuration, since it is possible to theoretically derive the code amount after compression from the code amount of the original image, the image after compression The size of data can be predicted. Therefore, even in a device such as a printer or a copying machine having a small memory capacity margin, it is possible to avoid the situation where the image data exceeds the memory capacity. Further, by adopting a means of quantizing the interval between the maximum value and the minimum value of the brightness value in the partial area, instead of deleting the high frequency component, for example, the outline part of the line of a line drawing or a character image. As described above, the contour information can be accurately left even in a portion where the brightness changes abruptly. Therefore, it is possible to avoid deterioration of print quality due to reduction of the information amount of the original image (that is, lossy compression).

The invention described in claim 7 is the image data compression method according to claim 6, wherein in the interval calculation step, the interval between the maximum value and the minimum value is equally divided into two quantization intervals. Then, in the quantization step, each pixel data of the partial region is binarized by 1-bit data into pixel data in a quantization section having a large luminance value and pixel data in a quantization section having a small luminance value. The image data is converted into a representative value of pixel data in a quantization section having a large luminance value and a representative value of pixel data in a quantization section having a small luminance value. Since the code amount after conversion by the conversion means can be set to (the number of pixels in the partial area) × (1 bit) + (the number of colors) × (the number of bits of one pixel of one color), the code after the data compression processing Can effectively reduce the amount Wear.

The invention described in claim 8 is the image data compression method according to claim 7, wherein in the representative value determination step, among the pixel data of the partial region, the pixel for the quantization interval having a large luminance value is selected. The representative value of the data is taken as the value of the pixel data having the maximum brightness value in the partial area, and the representative value of the pixel data for the quantization section having the smaller brightness value in the pixel data of the partial area is the minimum brightness value in the partial area. It is characterized in that it is set to the value of pixel data having a value. With this configuration, especially for an image composed of characters and line drawings, the contour of the line is emphasized, and it becomes possible to leave a lot of contour information. It is possible to prevent the deterioration of the image quality of the compressed image data.

The invention according to claim 9 is the invention according to claim 6 or 7.
The image data compression method according to claim 1, wherein in the representative value determining step, a histogram generating step for generating a histogram of the brightness value of the pixel data of the partial region, and a representative value value in each quantization interval based on the histogram are determined. And a representative value extracting step, which is configured to generate a histogram of the brightness value of each pixel of the partial region, and to generate the representative value according to the distribution of the brightness values in the histogram. Therefore, the optimum representative value can be determined according to the color change state of the image in the partial area in the original image, and the deterioration of the image quality due to the data compression can be minimized.

The tenth aspect of the present invention is the image data compression method according to any one of the sixth to ninth aspects, wherein in the partial image extracting step, the size of the partial area for dividing the original image data is set. It is characterized by having a partial area size setting step to be set. With this configuration, it is possible to freely set the resolution for compression by adjusting the area size according to the fineness of the image to be compressed. Therefore, it becomes possible to perform data compression with an optimum resolution in conformity with the original image to be compressed.

A program according to claim 11 is configured such that the computer is made to function as the image data compression device according to any one of claims 1 to 5, and the program is executed by the computer. The same operation as the image data compression device according to any one of 1 to 5 is obtained.

A recording medium according to claim 12 is the recording medium according to claim 1.
The image data compression apparatus according to claim 1, wherein the program recorded in the recording medium is read by a computer to be executed, and the image data compression apparatus according to claim 1. Similar effects are obtained.

An embodiment of the present invention will be described below with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram of an image data compression apparatus according to Embodiment 1 of the present invention.

In FIG. 1, the original image data input from the image input device 1 is compressed by the image data compression device 2, and the compressed image data is output to the image output device 3. The image input device 1 is a device such as a scanner or a digital camera that captures an image and outputs it as image data. The image output device 3 is a device that outputs an image of a printer, a copying machine, a facsimile, etc. by displaying it on a video screen or by printing it.

The image data compression device 2 includes an image storage means 4, a partial image extraction means 5, a brightness value calculation means 6, a section calculation means 7, a representative value determination means 8, a quantization means 11, and a partial area size determination means 12. have.

The image storage means 4 stores the original image data input from the image input device 1 and the compressed image data, and is composed of a RAM, an external storage device and the like. The partial image extraction means 5 divides the original image data stored in the image storage means 4 into a plurality of partial areas, and extracts pixel data of each partial area. The brightness value calculation means 6 calculates the brightness value for each pixel data of the partial area extracted by the partial image extraction means 5. The section calculation unit 7 obtains the maximum value and the minimum value of the brightness value of each pixel data of the partial area output from the brightness value calculation unit 6, and divides the section between the maximum value and the minimum value into a plurality of sections. Output the quantization interval. In the present embodiment, the section calculating means 7 equally divides the section between the maximum value and the minimum value into two sections. That is, the section calculation means 7 calculates the average value of the maximum value and the minimum value, and uses the average value as a threshold for dividing the quantization section. Hereinafter, a section having a brightness value larger than the threshold is called a quantization section H, and a section having a brightness value not larger than the threshold is called a quantization section L.

The representative value determining means 8 determines a representative value of pixel data belonging to each quantization interval. In the present embodiment, the representative value determining means 8 comprises a histogram generating means 9 and a representative value extracting means 10. The histogram generating means 9 generates a histogram of the brightness value of each pixel data in the partial area of the original image data output from the brightness value calculating means 6. The representative value extracting means 10 determines the value of the representative value in each quantization interval based on the histogram generated by the histogram generating means 9.

The quantizing means 11 calculates the values of all the pixel data belonging to the quantization section in each quantization section,
By setting the representative value in the quantization section, each pixel data is quantized to be data-compressed and stored in the image storage unit 4.

The image compression method of the image data compression apparatus of the present embodiment configured as described above will be described below.

FIG. 2 is a flowchart showing the overall processing procedure of the image data compression method according to the second embodiment of the present invention, and FIGS. 3 to 6 are the representative value determining means determining the representative value in each quantization interval. 7 is a flowchart showing a process to be performed.

In FIG. 2, first, the image input apparatus 1 is
The original image data is captured and transmitted to the image data compression device 2. Here, in the present embodiment, as an example, it is assumed that the original image data is RGB color image data and the data of each color of each pixel is 1 byte. The image data compression device 2 temporarily stores the input original image data in the image storage means 4. Next, the partial image extracting means 5 divides the original image data of the image storing means 4 into L partial areas S _n (n = 1, 2, ..., L) of M × N pixels, and Pixel data of the area S _n {R _n (i, j),
G _n (i, j), B _n (i, j)} (i = 1, 2, 3,
4, j = 1, 2, 3, 4) is read from the image storage means 4 (S1). In the present embodiment, for convenience of description, as an example, M = 4 and N = 4, and each partial region S
The size of _n is 16 pixels.

FIG. 7 is a schematic diagram showing the relationship between the original image data and the image of the partial area extracted by the partial image extracting means. As shown in FIG. 7, the original image data consists of a set of three images, an R (red) image, a G (green) image, and a B (blue) image. The partial image extraction means 5 converts this image data into 4 × 4.
Divided into subregions S _n of the pixel, and extracts R, G, each of the pixel data of the partial region S _n of 4 × 4 pixels of the image of B.

Next, the brightness value calculating means 6 has M × N × 3 pieces of pixel data {R _n (i, j), G _n (i, j) of the partial area S _n extracted by the partial image extracting means 5. ), B _n (i,
j)}, M × N luminance values {Y _n (i, j)} are calculated (S2). Here, as a calculation formula of the brightness value Y _n (i, j), a commonly used formula for calculating the brightness value of the RGB color image Y _n (i, j) = 0.299R _n (i, j)
j) + 0.587G _n (i, j) + 0.114B _n (i,
j) is used.

FIG. 8 is a schematic diagram showing the relationship between the brightness value and the pixels of the partial area when the brightness value is calculated from each pixel data of the partial area. As shown in FIG. 8, 4 × 4 luminance values are obtained as a result of calculating the luminance value from the pixel data of the partial region S _n of 4 × 4 pixels of each of R, G, and B by the above calculation formula.

Next, the section calculating means 6 determines the brightness value {Y
_n (i, j)} is sorted to obtain the maximum value Y _{n, max} and the minimum value Y _{n, min} , and the threshold value Y _th = (Y _{n, min} +
Y _{n, max} ) / 2 is calculated (S3). This makes Y
_{If n} (i, j)> Y _th , the pixel data {R
_{n (i, j), G} n (i, j), B n (i, j)} is assumed to belong to the quantization interval H, the pixel in the case of _{Y n (i, j) ≦} Y th Data {R _n (i, j), G _n (i,
j), B _n (i, j)} belong to the quantization interval L.

Next, the histogram generating means 9 generates a histogram of luminance values {Y _n (i, j)} (S4),
The representative value determining means 8 determines the representative value r _H of the quantization section _H and the representative value r _L of the quantization section L based on the histogram (S5). The process by which the representative value determining means 8 determines the representative values r _H and r _L from the histogram will be described later in detail.

Next, the quantizing means 11 uses the pixel data {R _n (i, j), G _n (i, j), B of each pixel of the partial area S _n.
n (i, j)} is the pixel data of the quantization interval H having a large luminance value {Y _n (i, j)} and the quantization interval L having a small luminance value.
Image data {b _n (i, j)} that is binarized with 1-bit data, and the representative value r _{n, H} and the brightness of the pixel data of the quantization interval H having a large brightness value. The pixel value is converted into the representative value r _n, L of the pixel data in the quantization section L having a small value and stored in the image storage means 4 (S6).

FIG. 9 shows the luminance value and quantity calculated in FIG.
A binarized image generated by binarizing by the substituting means
It is a schematic diagram showing the relationship with data. In this case,
Maximum brightness value Y_{n, max}Is Y_n(3,3) = 255
Minimum value Y_{n, min}Is Y_n(0,0) = 0, and the luminance value {Y
_nThe distribution of (i, j)} is a distribution having two local maxima.
(See FIG. 8)_thIs 127, the representative value r_{n, H}
Is {R_n(3,3), G _n(3,3), B_n(3,
3)}, representative value r_{n, L}Is {R_n(0,0), G_n(0,
0), B_n(0,0)}. And the binarized image
The data is as shown in FIG. 9 (b).

All the partial areas S _n (n = n
1, 2, ..., L) is repeated (S7,
S8), all partial areas S _n (n = 1, 2, ...,
When the processing for L) is completed, the image data compression apparatus 2 causes the binarized image data {b _n (i, j): i = 1, 2, 3, 4, j = stored in the image storage means 4. 1, 2,
3, 4, n = 1, 2, ..., L} and representative values {r _{n, H} , r _{n, L} : n = 1, 2, ..., L} are output as image data as compressed image data. It is transmitted to the device 3 (S9).

By compressing the image data as described above, in the image data compression apparatus 2 of the present embodiment, the information amount after compression is calculated from the information amount of the pixel data in the partial area S _n of the original image data. It becomes possible to do. For example, the amount of information of pixel data of 1 pixel is 1 byte ×
If 3 (the number of colors) = 3 bytes, the information amount of the partial area of 4 × 4 = 16 pixels is 48 bytes. When this is data-compressed by the above method, the information amount of the binarized image data is 16 bits (2 bytes) and the information amount of the representative value is 6 bytes, so that the total is 8 bytes. Therefore, the code amount is 1 /
It becomes 6. As described above, in the image data compression apparatus 2 according to the present embodiment, since the code amount after compression can be theoretically predicted, even in a device having a small memory capacity margin, the image data has the memory capacity. It is possible to avoid the situation of exceeding.

Next, the process in which the representative value determining means determines the representative value in each quantization interval will be described in detail.

In step S4 of FIG. 2, the histogram generating means 9 of the representative value determining means 8 causes the brightness value calculating means 6 to operate.
Luminance value of the calculated partial region S _n by {Y _n (i,
j)} from the histogram {H
_n (Y)} is generated. In this histogram, the brightness value is divided into certain appropriate sections, the horizontal axis represents the brightness value section, and the vertical axis represents the number of pixels in each brightness value section.

Next, referring to FIG. 3, the representative value extracting means 10
Finds a maximum point (peak) of the histogram {H _n (Y)}, and the maximum points are divided, that is, there are two or more maximum points and they are at distant positions (not adjacent to each other). ) Or not (S11).

Here, when the maximum points are divided,
Since there is a strong possibility that the image of the partial region S _n includes a contour (edge) portion such as a character or a line drawing, the representative value extracting unit 10 causes the representative value extracting unit 10 to represent the representative value r _{H, n} of the pixel data in the quantization interval H. Is defined as the maximum luminance value Y _{n, max,} and the representative value r _{L, n} of the pixel data in the quantization interval L is defined as the minimum luminance value Y _{n, min} .

[0054] That is, the first representative value extracting unit 10 obtains the maximum value Y _{n, m ax} of luminance in the partial region S _n, obtains the pixel data having the maximum luminance value (S12). Here, when there is one pixel data having the maximum brightness value (S13), the representative value r _{n, H} of the pixel data in the quantization interval _H
Is determined by the value of the pixel data (S19). However, when there are two or more pieces of pixel data having the maximum luminance value (S13), the representative value extracting unit 10 firstly has a color that is visually most strongly recognized by a person among those pixel data. Pixel data having the largest green (G) component is acquired (S14). Here, when there is one piece of pixel data having the largest G component (S15), the representative value extracting means 10
Represents the pixel data with the largest G component in the quantization interval H.
The representative value r _{n, H} of the pixel data is determined (S19). However, when there are two or more pieces of pixel data having the largest G component (S15), the representative value extracting means 10 is red (R), which is the color visually most strongly recognized by a person next to green. The pixel data having the largest component of is acquired (S16). Here, when the pixel data having the largest R component is one (S17), the pixel data having the largest R component is determined as the representative value r _{n, H} of the pixel data in the quantization interval H (S1).
9). However, when there are two or more pieces of pixel data having the largest R component (S17), the representative value extracting means 10 acquires pixel data having the largest blue (B) component this time (S18), and the B The pixel data having the largest component is determined as the representative value r _{n, H} of the pixel data in the quantization interval H (S1
9). Next, the representative value extraction means 10 acquires the minimum value Y _{n, min} of the brightness in the partial area S _n , and acquires the pixel data having the minimum brightness value (S20). Here, when there is one pixel data having the minimum luminance value (S21),
The representative value r _n, L of the pixel data in the quantization section L is determined as the value of the pixel data (S27). However, if there are two or more pieces of pixel data having the minimum luminance value (S2
1) First, the representative value extracting unit 10 acquires pixel data having the smallest green (G) component, which is the color that a person visually perceives most strongly, from the pixel data (S2).
2). Here, when the number of pixel data having the smallest G component is one (S23), the representative value extracting unit 10 sets the pixel data having the smallest G component to the representative value r _{n, of the} pixel data of the quantization interval L. _It is determined to be _L (S27). However, if there are two or more pixel data with the smallest G component (S2
3) The representative value extracting means 10 acquires pixel data having the smallest component of red (R), which is the color visually most strongly recognized by a person next to green (S24). Here, when the pixel data having the smallest R component is one (S25),
The pixel data having the smallest R component is determined as the representative value r _{n, L} of the pixel data in the quantization section L (S27). However, when there are two or more pieces of pixel data having the smallest R component (S25), the representative value extracting unit 10 acquires pixel data having the smallest blue (B) component this time (S2).
6) The pixel data having the smallest B component is determined as the representative value r _{n, L} of the pixel data in the quantization section L (S27).
The representative value determining means 8 outputs the representative value (r _{n, H} , r _{n, L} ) determined by the above processing to the quantizing means 11, and ends the processing.

If the representative value extracting means 10 determines in step S11 that the maximum points of the histogram {H _n (Y)} are not divided, the representative value extracting means 10 next
It is determined whether the histogram {H _n (Y)} has only one maximum point (S30).

Here, when there is one maximum point, the image of the partial area S _n is likely to be a single color or to have almost no color change in the area S _n . Therefore, the representative value extraction means 10 sets the representative values r _{H, n} and r _{L, n} of the quantization interval H and the quantization interval L as the pixel data values within the brightness value interval of the maximum point.

That is, the representative value extracting means 10 first
Pixel data within the luminance value section of the maximum point of the luminance value distribution is acquired (S31), and among these pixel data, the green (G) component, which is the color visually most strongly recognized by a person, is the largest. Pixel data is acquired (S32). Here, when the number of pixel data having the largest G component is one (S3
3), the representative value extracting means 10 determines the pixel data having the largest G component as the representative values r _{n, H} , r _{n, L} of the pixel data in the quantization interval H and the quantization interval L (S37). However, when there are two or more pieces of pixel data having the largest G component (S33), the representative value extracting means 10 in this case has the color red (R), which is the color visually most strongly recognized by a person next to green. The pixel data having the largest component is acquired (S34). Here, when the pixel data having the largest R component is one (S35), the pixel data having the largest R component is represented by the representative values r _{n, H of} the pixel data of the quantization interval H and the quantization interval L, r _{n, L} is determined (S37). However, if there are two or more pieces of pixel data with the largest R component (S3
5) The representative value extracting means 10 obtains the pixel data having the largest blue (B) component this time (S36), and the pixel data having the largest B component is the pixel of the quantization interval H and the quantization interval L. Data representative values r _{n, H} and r _{n, L} are determined (S3
7). The representative value determining means 8 outputs the representative value (r _{n, H} , r _{n, L} ) determined by the above processing to the quantizing means 11, and ends the processing.

In step S30, when the representative value extracting means 10 determines that there is one or more maximum points in the histogram {H _n (Y)}, the representative value extracting means 10 then disperses the brightness values. σ ² is calculated by (Equation 1) (S3
8).

[0059]

[Equation 1]

Then, the representative value extracting means 10 determines whether or not the variance value σ ² of the brightness values is less than or equal to a predetermined threshold value σ _th ² (S).
39).

[0061] Here, when the variance value sigma ² of the luminance values of the predetermined threshold value sigma _th ² or less (S39), the image of the partial region S _n, the color change in the partial region S _n comparison It is likely to be small. Therefore, the representative value extracting means 10 determines the representative value r of the pixel data in the quantization interval H.
_{Let H, n be the} value of the smallest luminance value of the pixel data belonging to the quantization interval H, and let the representative value r _{L, n} of the pixel data of the quantization interval L be the value of the pixel data belonging to the quantization interval L. Has the maximum brightness value.

That is, first, the representative value extracting means 10
First, of the pixel data belonging to the quantization interval H, the pixel having the smallest luminance value is acquired (S40), and among the acquired pixel data, green (G), which is the color visually most strongly recognized by a person. The pixel data having the smallest component of is acquired (S4
1). Here, when the number of pixel data having the smallest G component is one (S42), the representative value extracting unit 10 sets the pixel data having the smallest G component to the representative value r _{n, of the} pixel data of the quantization interval H. _H is determined (S46). However, if there are two or more pixel data with the smallest G component (S4
2) The representative value extracting means 10 acquires pixel data in which the component of red (R), which is the color visually most strongly recognized by a person next to green, is the smallest (S43). Here, when the number of pixel data having the smallest R component is one (S44),
The pixel data having the smallest R component is determined as the representative value r _{n, H} of the pixel data in the quantization interval H (S46). However, when there are two or more pieces of pixel data having the smallest R component (S44), the representative value extracting unit 10 acquires pixel data having the smallest blue (B) component this time (S4).
5) The pixel data having the largest B component is determined as the representative value r _{n, H} of the pixel data in the quantization interval H (S46).
Next, the representative value extraction unit 10 acquires the pixel data having the largest luminance value among the pixel data belonging to the quantization interval L (S4).
7) Among the acquired pixel data, the pixel data having the largest green (G) component, which is the color visually recognized by the person most strongly, is acquired (S48). Here, when the number of pixel data having the largest G component is one (S49), the representative value extracting unit 10 sets the pixel data having the largest G component to the representative value r _{n, of the} pixel data of the quantization section L. Determine to _L (S5
3). However, when there are two or more pieces of pixel data having the largest G component (S49), the representative value extracting unit 10 detects red (R), which is the color visually most strongly recognized by a person next to green. The pixel data having the largest component of is acquired (S5
0). Here, when the number of pixel data having the largest R component is one (S51), the pixel data having the largest R component is determined as the representative value r _{n, L} of the pixel data of the quantization interval L (S53). . However, when there are two or more pieces of pixel data having the largest R component (S51), the representative value extracting unit 1
For 0, this time, the pixel data having the largest blue (B) component is acquired (S52), and the pixel data having the largest B component is determined as the representative value r _{n, L} of the pixel data of the quantization interval L ( S53). The representative value determination means 8 is
The determined representative value (r _{n, H} , r _{n, L} ) is output to the quantizing means 11, and the processing is ended.

[0063] In step S39, is larger than variance sigma ² is the predetermined threshold value sigma _th ² of luminance values, the image of the partial region S _n, the color change in the partial region S _n is relatively large It is likely to be a thing. Therefore, the representative value extracting means 10 sets the representative value r _{H, n} of the pixel data in the quantization section H as the value of the pixel data belonging to the quantization section H whose luminance value is the median value of the quantization section H. The representative value r _{L, n} of the pixel data in the quantization interval L is the value of the pixel data belonging to the quantization interval L whose luminance value is the median value of the quantization interval L.

That is, the representative value extracting means 10 obtains the smallest luminance value Y _{n, pk1} of the pixel data belonging to the quantization interval H (S60), and further, the pixel data belonging to the quantization interval H. The maximum luminance value Y _{n, max} is obtained (S61). Then, the representative luminance value Y _{n, 1} = (Y _{n, pk1} + Y _{n, max} ) / 2 of the quantization interval H is calculated, and the pixel data having the luminance value Y _{n, 1} is acquired (S62). Next, the representative value extraction unit 10 acquires pixel data having the largest component of green (G), which is the color visually most strongly recognized by a person, from the acquired pixel data (S63). Here, when the number of pixel data having the largest G component is one (S64), the representative value extracting unit 10 sets the pixel data having the largest G component to the representative value r _{n, of the} pixel data of the quantization interval H. _H is determined (S68). However, when there are two or more pieces of pixel data having the largest G component (S64), the representative value extracting unit 10 detects red (R), which is the color visually most strongly recognized by a person next to green. The pixel data having the largest component of is acquired (S65). Here, when the pixel data having the largest R component is one (S66), the pixel data having the largest R component is determined as the representative value r _{n, H} of the pixel data in the quantization interval H (S68). . However, if there are two or more pieces of pixel data with the largest R component (S6
6), the representative value extracting means 10 obtains the pixel data having the largest blue (B) component this time (S67), and sets the pixel data having the largest B component to the representative value r of the pixel data in the quantization interval H. _It is determined to be _{n and H} (S68). Next, the representative value extracting means 10 acquires the maximum luminance value Y _{n, pk0} of the pixel data belonging to the quantization section L (S69), and further, the luminance of the pixel data belonging to the quantization section L. The smallest value Y _{n, min} is obtained (S70). Then, the representative luminance value Y _{n , 0} = (Y _{n, pk0} + Y _{n, min} ) of the quantization section L
/ 2 is calculated, and pixel data having a luminance value of Y _{n, 0} is acquired (S71). Next, the representative value extraction unit 10 acquires pixel data having the smallest component of green (G), which is the color visually most strongly recognized by a person, from the acquired pixel data (S72). Here, when the number of pixel data with the largest G component is one (S73), the representative value extraction means 10
Represents the pixel data having the smallest G component in the quantization interval L.
The representative value r _{n, L} of the pixel data is determined (S77). However, when there are two or more pieces of pixel data having the smallest G component (S73), the representative value extracting means 10 is red (R), which is the color visually most strongly recognized by a person next to green. The pixel data having the smallest component of is acquired (S74). Here, when the number of pixel data having the smallest R component is one (S75), the pixel data having the smallest R component is determined as the representative value r _{n, L} of the pixel data of the quantization interval L (S7).
7). However, when there are two or more pieces of pixel data having the smallest R component (S75), the representative value extracting unit 10 acquires pixel data having the smallest blue (B) component this time (S76), and the B The pixel data having the smallest component is determined as the representative value r _{n, L} of the pixel data in the quantization section L (S7).
7). The representative value determining means 8 outputs the representative value (r _{n, H} , r _{n, L} ) determined by the above processing to the quantizing means 11, and ends the processing.

By the above processing, the representative value determining means 8
Is a representative value r _{n, H} , r _{n, L} of the quantization interval H and the quantization interval _L.
The brightness value distribution (histogram)
FIG. 10 shows the relationship between and the representative value.

FIG. 10 is a diagram showing the relationship between the distribution of luminance values and the values of representative values.

In FIG. 10, (a) shows the case where the maximum points (peaks) of the histogram {H _n (Y)} are divided. In this case, the image of the partial area S _n is
Since there is a strong possibility that the portion (edge) such as a character or a line drawing is included, the representative value extracting unit 10 sets the representative value r _{H, n} of the pixel data in the quantization interval H to the maximum luminance value Y _{n, max.} Then, the representative value r _{L, n} of the pixel data in the quantization interval L is set as the minimum value Y _{n, min} of the luminance.

In (b), the maximum point (peak) of the histogram {H _n (Y)} is not divided, and the variance value σ ^{2 of} luminance values is
There represents a case where the predetermined threshold value sigma _th ² greater than this case, the image of the partial region S _n is the possibility color change in the partial region S _n is relatively large is large. Therefore, the representative value extracting means 10 sets the representative value r _{H, n} of the pixel data in the quantization section H as the value of the pixel data belonging to the quantization section H whose luminance value is the median value of the quantization section H. , The representative value r _{L, n} of the pixel data in the quantization section L is
The luminance value of the pixel data belonging to the quantization interval L is the value of the median value of the quantization interval L.

In (c), the maximum point (peak) of the histogram {H _n (Y)} is not divided, and the variance value σ ^{2 of} luminance values is
There represents a case of predetermined threshold value sigma _th ² or less, in this case, the image of the partial region S _n is the possibility color change in the partial region S _n is relatively small is large. Therefore, the representative value extracting means 10 sets the representative value r _{H, n} of the pixel data in the quantization section H as the value of the pixel data belonging to the quantization section H having the smallest luminance value, and sets the representative value in the quantization section L. Let the representative value r _{L, n} of the pixel data be the value of the pixel data having the largest luminance value among the pixel data belonging to the quantization interval L.

(D) shows a case where the histogram {H _n (Y)} has one maximum point. In this case, the image of the partial area S _n has a single color in the area S _n . Or there is a high possibility that there is almost no color change. Therefore, the representative value extraction means 10 sets the representative values r _{H, n} and r _{L, n} of the quantization interval H and the quantization interval L as the pixel data values within the brightness value interval of the maximum point.

As described above, the representative value determining means 8 of the present embodiment can determine the optimum representative value according to the color change state of the image in the partial area in the original image.
It is possible to minimize deterioration of image quality due to data compression. Further, since it is possible to accurately leave the information of the contour of the image, it is possible to avoid the deterioration of the print quality due to the reduction of the information amount of the original image (irreversible compression).

In the present embodiment, the representative value determining means 8 is configured to determine the representative value based on the histogram in each quantization section, but the image data to be compressed is not limited to characters or line drawings. If there is, the representative value determining unit 8 determines the maximum representative value of the pixel data for the quantization interval H having a large luminance value among the pixel data of the partial region extracted by the partial image extracting unit 5 in the partial region. A pixel data value having a brightness value is set, and a representative value of the pixel data for the quantization section L having a small brightness value in the pixel data of the partial region is set as a pixel data value having the minimum brightness value in the partial region. It may be configured to do so. With such a configuration, the contour portion of the character or line drawing is emphasized, a large amount of contour information can be left, and it is possible to prevent deterioration of the image quality of the image data after data compression.

Further, in the present embodiment, the quantization interval is explained as two intervals, but it is also possible to set the quantization interval to three intervals or more depending on the purpose.

Further, in the present embodiment, as an example, the size of the partial area extracted by the partial image extracting means 5 is 4 × 4.
Although described as a pixel, the size of the partial area can be changed by the partial area size determining unit 12 according to the purpose.

[0075]

As described above, according to the image data compression apparatus of the first aspect of the present invention, since it is possible to theoretically predict the maximum code amount after compression, the memory capacity is increased. It is possible to avoid the situation where the image data exceeds the memory capacity even in a device with a small margin, and to reduce the amount of information of the original image because the information of the outline of the image can be left accurately. It is possible to provide an image data compression device capable of avoiding deterioration of print quality due to (lossy compression).

According to the second aspect of the present invention, it is possible to provide a data compression apparatus capable of efficiently reducing the code amount after data processing.

According to the third aspect of the present invention, particularly in the case of an image composed of characters and line drawings, a large amount of contour information remains in the image after the data compression processing, so that the deterioration of the image quality of the image data after the data compression is prevented. It is possible to provide an image data compression device capable of performing the above.

According to the invention described in claim 4, it is possible to determine the optimum representative value according to the color change state of the image in the partial area in the original image, and the deterioration of the image quality due to the data compression is minimized. It is possible to provide an image data compression device that can suppress the above.

According to the fifth aspect of the present invention, it is possible to provide an image data compression apparatus which is suitable for an original image to be compressed and can perform data compression with an optimum resolution.

According to the image data compression method of the sixth aspect of the present invention, since the size of the compressed image data can be predicted, it is a device such as a printer or a copying machine having a small memory capacity margin. However, it is possible to avoid the situation where the image data exceeds the memory capacity, and
Since it is possible to accurately leave contour information even after the data compression processing, it is possible to provide an image data compression method capable of avoiding deterioration of print quality due to lossy compression.

According to the invention described in claim 7, it is possible to provide a data compression method capable of efficiently reducing the code amount after data processing.

According to the invention as set forth in claim 8, a large amount of contour information remains in the image after the data compression processing, particularly for an image consisting of characters and line drawings, so that deterioration of the image quality of the image data after data compression is prevented. It is possible to provide a possible image data compression method.

According to the invention described in claim 9, the optimum representative value can be determined according to the color change state of the image in the partial area in the original image, and the deterioration of the image quality due to the data compression is minimized. It is possible to provide an image data compression method that can suppress the above.

According to the tenth aspect of the present invention, it is possible to provide an image data compression apparatus which can be adapted to the original image to be compressed and can perform data compression with an optimum resolution.

According to the program of claim 11,
By causing the computer to execute the program, the same effect as that of the image data compression apparatus according to any one of claims 1 to 5 can be obtained.

According to the recording medium of the twelfth aspect, the image data compression apparatus according to any one of the first to fifth aspects is provided by causing a computer to read and execute the program recorded in the recording medium. The same effect can be obtained.

[Brief description of drawings]

FIG. 1 is a block diagram of an image data compression device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing the overall processing procedure of the image data compression method according to the second embodiment of the present invention.

FIG. 3 is a flowchart showing a process in which a representative value determining unit determines a representative value in each quantization interval.

FIG. 4 is a flowchart showing a process in which a representative value determining unit determines a representative value in each quantization interval.

FIG. 5 is a flowchart showing a process in which the representative value determining unit determines a representative value in each quantization interval.

FIG. 6 is a flowchart showing a process in which a representative value determining unit determines a representative value in each quantization interval.

FIG. 7 is a schematic diagram showing a relationship between original image data and an image of a partial area extracted by a partial image extracting means.

FIG. 8 is a schematic diagram showing a relationship between pixels in a partial area and the brightness value when the brightness value is calculated from each pixel data of the partial area.

9 is a schematic diagram showing the relationship between the brightness value calculated in FIG. 8 and the binarized image data generated by binarizing by the quantizing means.

FIG. 10 is a diagram showing a relationship between a distribution of luminance values and a value of a representative value.

[Explanation of symbols]

1 Image input device 2 Image data compression device 3 Image output device 4 Image storage means 5 Partial image extraction means 6 Luminance value calculation means 7 section calculation means 8 Representative value determination means 9 Histogram generating means 10 Representative value extraction means 11 Quantization means 12 Partial area size determination means

Claims (12)

[Claims] 1. Original image data is divided into a plurality of partial areas,
Output from the partial image extracting means for extracting pixel data of each partial area, a brightness value calculating means for calculating a brightness value for each pixel data of the partial area extracted by the partial image extracting means, and the brightness value calculating means. Section calculating means for obtaining the maximum value and the minimum value of the brightness value of each pixel data of the partial area, and outputting a quantization section obtained by dividing the section between the maximum value and the minimum value into a plurality of sections, A representative value determining unit that determines a representative value of pixel data belonging to a quantization section, and all pixel data belonging to the quantization section in each of the quantization sections are set to the representative value in the quantization section. Quantization means for quantizing each pixel data,
An image data compression device comprising: 2. The section calculating means equally divides a section between the maximum value and the minimum value into two quantization sections, and the quantizing section calculates each pixel data of the partial area. , Binarized image data obtained by binarizing the pixel data in the quantization section having a large luminance value and the pixel data in the quantization section having a small luminance value with 1-bit data, and the quantization section having a large luminance value. 2. The image data compression apparatus according to claim 1, wherein the representative value of the pixel data is converted into the representative value of the pixel data of the quantization section in which the luminance value is small. 3. The representative value deciding means sets a representative value of the pixel data for a quantization section having a large luminance value in the pixel data of the partial area, the pixel data having a maximum luminance value in the partial area. Of the pixel data of the partial area, and the representative value of the pixel data for the quantization section having the smaller brightness value is the value of the pixel data having the minimum brightness value in the partial area. The image data compression device according to claim 2. 4. The representative value determining means generates a histogram of luminance values of pixel data of the partial area, and a representative value determining a value of the representative value in each of the quantization intervals based on the histogram. Extraction means,
The image data compression apparatus according to claim 1 or 2, further comprising: 5. The image according to claim 1, wherein the partial image extracting means has a partial area size setting means for setting a size of a partial area into which the original image data is divided. Data compression device. 6. The original image data is divided into a plurality of partial areas,
A partial image extracting step of extracting pixel data of each partial area; a brightness value calculating step of calculating a brightness value for each pixel data of the partial area extracted in the partial image extracting step; and a brightness of each pixel data of the partial area A section calculation step of obtaining the maximum value and the minimum value of the values and outputting a quantization section obtained by dividing the section between the maximum value and the minimum value into a plurality of sections, and a representative of pixel data belonging to each of the quantization sections. A representative value determining step of determining a value, and in each of the quantization intervals, all pixel data belonging to the quantization interval is set to the representative value in the quantization interval, and the quantization is performed to quantize each pixel data. Conversion step,
An image data compression method comprising: 7. In the section calculating step, a section between the maximum value and the minimum value is equally divided into two quantization sections, and in the quantizing step, each pixel data of the partial area is Binarized image data obtained by binarizing 1-bit data into the pixel data in the quantized section having a large luminance value and the pixel data in the quantized section having a small luminance value,
7. The image data compression method according to claim 6, wherein the image data is compressed into a representative value of pixel data in a quantization section having a large luminance value and a representative value of pixel data in a quantization section having a small luminance value. 8. In the representative value determining step, among the pixel data of the partial area, a representative value of the pixel data for a quantization section having a large brightness value is set as a pixel having a maximum brightness value in the partial area. Data value,
8. The representative value of the pixel data for the quantization section having the smaller brightness value in the pixel data of the partial area is set to the value of the pixel data having the minimum brightness value in the partial area. The image data compression method described in. 9. The representative value determining step includes a histogram generating step of generating a histogram of luminance values of pixel data of the partial area, and a representative determining a value of the representative value in each of the quantization intervals based on the histogram. The image data compression method according to claim 6 or 7, further comprising a value extraction step. 10. A partial area size setting step of setting the size of a partial area into which the original image data is divided, in the partial image extracting step, according to any one of claims 6 to 9. Image data compression method. 11. A program for causing a computer to function as the image data compression device according to claim 1. 12. A recording medium on which the program according to claim 11 is recorded.