JPS58207168A - Storage device of color picture information - Google Patents

Abstract

PURPOSE:To reduce the capacity of a picture information storage device, by converting the color picture information into the luminance and hue informations and then compressing the quantity of at least one both informations which are converted in response to the features of the picture information. CONSTITUTION:The picture information given from a reader 11 is added to a color separating part 14 and converted into a luminance information and two hue information to be fed to corresponding information reconstituting parts 15- 17. These parts 15-17 sample these informations with pitches that can be controlled independently of each other to compress the informations and store these compressed informations to storage parts 18-20. The read informations of the parts 18-20 are converted into original luminance and hue informations via information reconstituting parts 21-23 and then displayed to a display 12 via a color synthesizing part 24 and an interface 13. The information compressing factors of the parts 15-17 are properly decided by the using original pictures. With a document, etc. for instance, the quality is never deteriorated for the reproduced pictures despite a large compressing factor of the color information.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical field to which the invention pertains] This invention relates to a color image storage device that stores color images with high efficiency.

[Technical background of the invention and its problems]

Color images have the advantage of being easy to view, rather than presenting a large amount of information to humans. Black and white photo turned into color photo
As with the development of X, images handled electronically are also becoming colored. Supporting this progress is progress in electronics.

Furthermore, in the field of computer-based recognition processing, voice recognition has taken the lead, but image processing, particularly single image recognition, is finally gaining ground. One reason for this delay in image processing is that the image contains too much information.

However, these obstacles are gradually being overcome due to the increase in memory capacity and the increase in computing speed and processing power of computers.

Among these trends, there is a noticeable movement toward handling color images. In this case, since the color image is made up of information of a plurality of colors, the amount of information i to be handled becomes extremely large. No matter how large the size of memory becomes, and the speed of computer calculations increases, the problem of having a large amount of information, for example, will lead to larger storage capacity, and the size of the front and back will become thicker. It is had.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a color image information storage device k which eliminates the above drawbacks and requires a small capacity.

[Summary of the invention]

This FJA converts color image information into luminance information indicating brightness and hue information indicating hue, and compresses the amount of at least one of the luminance information and hue information depending on the characteristics of the image being handled. , it's something to remember. The image to be handled is an image to be read or an image to be displayed.

〔Effect of the invention〕

For example, 1"
:11 If the hue change is small, compress the entire amount of hue information,
Since the amount of luminance information is maintained, the storage capacity is greatly reduced. The same thing happens when there is little change in brightness. These effects will be better understood when working with actual documents.

This is because there is almost no hue change in an actual document.

[Embodiments of the invention]

Next, an embodiment of the present invention will be described with reference to the drawings. As shown in FIG. 1, the device in this embodiment can read color images: red, green, and blue (hereinafter abbreviated as R, G, and B).
Reading that outputs the digital component signal for each! l(Lυ and
An input/output interface (131 and this input/output interface
JG via 3.

A color separation unit α which is supplied with a component signal of B and converts it into luminance information and first and second hue information (and each of the luminance information and first and second hue information that can be output from this color separation unit I) A first luminance information reconstruction unit (151, first and second
Hue information reconstruction unit αe.

αD and these reconstruction units tt51. first and third storage units storing signals from each of σ and aη, (l
i, (a) and this first and third memory &1i (18
1, α1. (The second luminance information reconstruction unit η, the third and fourth Hue information reconstruction unit (2) and these reconstruction units.

A color synthesis unit 8 (Foundation) that recombines the signals @a, c+, and B from the receiver, and a control unit (E) that sequentially controls the color separation unit I to the color synthesis unit (to).
It consists of However, the output from the color composition department (foundation) is
The display device (1) is connected via the input/output interface section σ3).
It is visualized in 2.

In the reading device circle, reflected light from, for example, a document (not shown) is decomposed into R, G, and B, and digitized signals corresponding to the respective intensities are output. Here, the pixel of the original t is specified as (i, j), and the respective signals are R(i, j) G(ij), B(i, j
).

For such a signal, L(i,
j) = a(1,1)R(i,j) 10a(1,2)
G (i, j) + a (1, 3) B (i, j) 0 + (i, j) = a (2, x) R (i, j) + a (
2,2)()(i,j)-(1) Pass a(2,3)B(t
+J) 02(i,j) = a(3,1)R(i,D +a(
3,2)G(i,j)+a(3,3)B(i,j). Equation (1) is a type of coordinate transformation.

Here, L (i, D is the luminance information of the pixel specified by (i, j), c, (1, 2) and C2 (1, 2) are the hue information.

a(1,1), a(i,z),...+ a(
3,3) are well-known conversion coefficients. In this example, values that can be adopted in the NTSO system as shown in Table 1 below are used.・Only-Memory (Only M
memory (hereinafter abbreviated as ROM).

R (i, j), G (i, j), B (i, j,)
is, for example, 4 bits D22-16, so R
A 12-bit signal is given to OM. This one
The 2-pit signal is connected to the R and OM address lines.

These R(ilj), o(i, j), B(i, j
), the address in OM is R(i,
j), G(ilj).

For B(i,j), L(+,J), 01(1,J), C'2(1,J) calculated according to formula (1)
Store the 6 values of .

Therefore, R, OM, i.e., the variable-converted image information L(i,''), C+(
t, J), C2(i.

j) is output. However, R(i,'')+G(t,J)t
B('+J) and L(i, j), 0s(i, J), 0
The sampling frequency ridge (output frequency) is the same as that of 2(i,j). In this way, image information is converted into luminance information and first and second hue information.

The inventors investigated color images, especially text*, and found that it is easier to read documents, etc. in color images, and there are advantages such as being able to quickly grasp important points. be. Dad, in this color image, there are many changes in brightness, but there are almost no changes in hue.

Specifically, there are only a few places where colors other than black are used. This makes sense when you consider the process of creating Matametsu. This is because colors other than black are used to make the document easier for the author to see, and not so many colors that it becomes visually confusing.

The inventors further analyzed color images such as photographs and confirmed that the hue does not change in a very small range. At the same time, the linear density during reading or displaying has a performance of 10 μ/μ and 16 μ/μ.On the other hand, in the case of photographs and other images, the change in hue requires approximately 10 μ. As far as hue changes are concerned, there is a lot of information that there are even 8 pieces per basket.

Therefore, in this embodiment, the luminance information is reduced to the amount of information supplied via the input/output interface, and the first and M2 hue information are compressed. That is, the first and second hue information add/blend the original image information. In the first brightness information reconstruction unit a9, the first and second hue information reconstruction units σ phrase, and ση, the rate at which all information is compressed is variable. In this embodiment, the compression rate in the first luminance information reconstruction unit 0 group is maintained at 1, that is, the amount of information, and the compression ratio in the first and second hue information reconstruction units uel, (17) Let be 1 or more.

The method of compressing information in this embodiment is performed by the geometric mode conversion method described below. This geometric mode conversion method is a type of a-density conversion, as disclosed in Japanese Unexamined Patent Publication No. 90375/1983.

As shown in Fig. 3, the read pitch p by the reading device is d, and the signal read at the point indicated by the circle mark in the color image is converted into the signal of the pitch d' indicated by the X mark. do.

The principle of this conversion is the weighted average. That is, point Q after conversion
For the area S occupied by 1, the original signal contained within this area S (for each information component, αe.

(O input signal to 171) P 11 + "!2 +
P2. ＋P! 2. SII + S1
2 * S□, 522 (However, S = S,, +S□2+
S21+522), the average pixel value at point Q1 is W+I'S.

In such a geometric mode conversion method, the point after conversion (4
As the area S occupied by 1 increases, the number of points PIJ before conversion included in this area S increases. That is, since the average of more points is calculated, the compression ratio increases.

This geometric mode conversion method consists of a first means of determining the position of each pixel in the converted image in the original image, and a method of determining the position of pixels surrounding the position determined by the first means from the original image. A second means for selectively outputting information and the principle f'? :
This can be realized by storage means that stores pixel information obtained by performing interpolation calculations using appropriate distances according to equation (1).

The first means specifies the position indicated by the X mark in Fig. 3, that is, the position after conversion. The painting information is selected. This pixel information and the position information specified by the first means specify the address of the storage means, specifically KldROM. The output from this ROM is directly converted into a geometric This was achieved using the mode conversion method.

In this example, as the variable Pl in equation (2), luminance information L(l, j) is used for (i, j) that specifies one pixel.
, first and second hue information 01(i,j), 02(i
, j) is used. It should be noted that, as shown in FIG. This is performed for each part of the set of hue information C2 (iIJ), and does not even convert the variable space. Simply the same
It merely changes the number of samples within the set. This is different from the conversion in the color separation section a.

However, if the geometric mode conversion method whose principle is shown in equation (2) is used, the information changes more smoothly and the quality of the original image is not degraded.

In this embodiment, based on the above-mentioned consideration, the first luminance information reconstruction unit (151) maintains the number of samplings in the input information L("l J ), that is, outputs the output information L'(i
, j). First and second hue information configuration h1~
Q6J, αη, input information C+(t, D, 02(t−
For example, if the number of samples of +j) is reduced to 1 out of 10,
That is, the area S shown in the stripe 3 diagram is made 10 times the original information,
As shown in FIG. 4, the output information 0+'(l, j), C
2'(t,j) is output.

The information obtained in this way L'(i, J) C+'(
i, J) C2' (1, J) CffR1 to 71-
3 Memory section α8yacJ, c! I can remember 01vc.

Now, when reading 9 of an A4 size image at 8 lines/n, luminance information L (1, j), first and second hue information 01 (i, j
), 02(i, j) is 4 Mbits per A4 size image. The total is 12 Mbits. However, in this embodiment, 0+(i,j),
Since the amount of information for 02(j+j) is only 0.4M bits each, the total is 4M+〇, 4 x 2M = 4.8M
(bit) is fine. That is, the amount of information is reduced to 40%, and the storage portions t in the first to third storage portions α~, Q9, and @ are only 40% of the conventional size.

As a result, the image information is compressed and stored.

This compressed information can be restored as follows.

In principle, it would be good to follow the processes from N1 to the third memory section, α mark, and 001 in reverse. First, in the second luminance information reconstruction section C/111, the third and fourth hue information reconstruction sections 2, [23], the first luminance information reconstruction section α gloss, the first
and the second hue information reconstruction unit [6] performs the reverse processing in [171].

Referring to Figure 3, from the arrangement of points Q indicated by X marks, ○
Find Pij f indicated by the mark.

Each of these reconstruction units (two swords, (22), r23) includes a first means for determining the position in the original image for each pixel of the image after conversion, and a position determined from the original image by the first means. A second method that selectively outputs pixels around the position.
and storage means for storing pixel information obtained by performing calculations in advance.

That is, first, the position of point PIJ indicated by the circle in FIG. 3 is determined by the first means. Next, the second means outputs the pixel at the 0 point Q around the point P1j obtained from the first means (・τ) to the selection FFJ. The pixel values at the positions designated by the first means can be determined by proportional distribution based on the position and circle relationship with the multiple strokes obtained by the second means, so these values are stored in the storage means. The storage means is, for example, ROM.
The address is specified by the position of the nine points P1 and the pixel values of a plurality of points on the simple side.

In this embodiment, the second luminance information reconstruction unit (2I)
The input information L'(i, j) is output as is. ”(',
'') is the same as L(i,j), the second luminance information reconstruction unit (21) outputs all the same information as L(i,j). In the third and fourth hue information reconstruction units @ and @, the first
and a second hue information reconstruction unit α0. In αη, corresponding to the compression of the information amount, that is, the expansion of the linear density, the information amount is expanded, that is, the linear density is completely reduced, and C1″(i, D, 0
2" (i, j) is output.

The information from each of these reconstructor assistants, @, and (c) has zero information for the same point in the original image, and the amount of information is the same.

Such information L(i, j), O,'(i, j),
C='(i, D are inversely converted into brightness information of the three primary colors R, G, and B in the color synthesis unit r24J. In other words, they can be inversely converted according to the following formula. However, this conversion is also NTS
Based on the C method, the coefficients B(i, j) are shown in Table 2.

The color composition section (1) is composed of a ROM, similar to the color separation section (a), and the calculation of equation (3) is performed in advance and stored in the ROM. L(i, j), 0
+ '(i, D.

C2''(i, j)it are both 4-bit signals, and with these signals, that is, with 12-bit signals, R
Specify the address of OM.

In the end, the signal from the color combining section (support)) is transmitted to the display device a.
The brightness and darkness of the three primary colors R, G, and B suitable for display 1:
#Notice.

In the following embodiments, the control section (c) controls all sequential operations between the constituent elements.

, [Other Embodiments of the Invention] In the embodiments described above, the amount of information in the first luminance information reconstruction unit σ9 is not compressed, that is, the number of samplings is maintained, and the first
and a second hue information reconstruction unit (16), (L71
d information was compressed by x to reduce the number of samplings.

This is based on the fact that there is little change in hue in documents used in office work.

The example here is based on the recent improvement in the performance of display devices.
This is an example in which brightness information is compressed and stored using the same configuration as the previous embodiment, focusing on the fact that the colors that can be expressed have become more diverse and it has become possible to create color images with various hues. The display device in this embodiment is such a high-performance device, and is intended for original images in which luminance changes are small and the hue changes minutely and frequently.

In this case, the first and second hue information reconstruction units (t6
), σD is not compressed, and the information amount in the first 14 degree information reconstruction unit t15) is compressed. Corresponding to this, the second luminance information reconstruction unit 0 In this way, according to the present invention, the amount of information that changes little is compressed depending on the nature of the image being handled, and the amount of information that changes frequently is not compressed in principle. However, even with this frequently changing information, it is optional to compress the amount of information as long as the characteristics of the original image are not lost.

Furthermore, the compression of the amount of information is not limited to the geometric mode conversion method as in the above-described embodiments. For example, when the hue change of the original image is small, the first and second hue information reconstruction units αe and σ convert one piece of information, for example every 10 cycles, to the information that is sent one after another synchronously. You can also do it like this.

Accordingly, the third and fourth hue information reconstruction units (16
1σγj, the information sent by each station 10 to may be output for a time t0 at synchronization t0.

Furthermore, whether or not to press the news is based on the following two points.

Firstly, it depends on the characteristics of the original image being read, and secondly on the display capabilities of the display device. The former compresses information that is essentially unnecessary for image formation among the information handled as described above. In the latter case, since there is no point in storing information that cannot be displayed on a display device, such information may be compressed.

In short, any modifications may be included in this invention as long as they do not depart from the spirit of this invention.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing the configuration of the embodiment, Fig. 2 is a diagram showing the color separation section a4) in Fig. 1, Fig. 3 is a diagram for explaining the geometric mode conversion method, and Fig. 4 is a diagram showing information compression. 1 α condolence...Color separation unit uj...B1 luminance cancellation re-establishment part O...First hue information reconstruction unit aη...Second hue information reconstruction unit α... 1st memory section (1 bell... 2nd memory section assistant... 3rd memory section (2)... Color composition department agent Patent attorney Noriyuki Chika (and 1 other person) ζ 11 1st Figure 2 for Figure 1

Claims (3)

[Claims] (1) A color separation unit that converts color image information into brightness information indicating brightness and hue information indicating hue; a brightness information reconstruction unit that compresses the amount of brightness information from the color separation unit; and the color a hue information reconstruction unit that compresses the amount of hue information from the decomposition unit; a storage unit that stores information from the first hue information reconstruction unit and the first luminance: official gazette reconstruction unit; a color synthesis section that configures information from the storage section to be suitable for image display; A color stroke number information storage device characterized by changing the compression rate of the amount of hue information. (2) A color image storage device according to claim 1, wherein the color image information is composed of red, green, and blue components. (3) When the hue change of the image to be handled is small, increase the compression ratio of the amount of hue information in the hue information reconstruction unit, and (4) When the luminance change is small, increase the compression ratio of the amount of luminance information in the luminance information reconstruction unit, and (5) increase the amount of luminance information in the hue information reconstruction unit. (6) The hue information reconstruction unit is calculated based on the fineness of the hue information.