JPH08275009A - Color image processing unit - Google Patents

Abstract

PURPOSE: To obtain the color image processing method that reproduces images with delicate color difference in an excellent way and the color image processing unit realizing the method. CONSTITUTION: The processing unit is made up of an A/D converter 102, an image memory 103, a RAM 105, an output selector 106, plural ROMs 107a-c, plural output ports 108a-c, plural printers 109a-c corresponding to them, and a parameter ROM 111 which are connected to a CPU 104 controlling image processing via a bus 110. Color conversion processing to process an input color image from a camera 101 including a color signal in excess of a color reproduction range of a printer being an output device of the input color image from a camera 101 is conducted by obtaining a conversion amount of a color signal based on, e.g. a distribution state of color signals in a color space so as to set the reproduced color signals within the color reproduction range thereby attaining image reproduction with fidelity, and also the color difference between colors close to each other is maintained in this color processing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image processing method for processing an input color image signal and outputting it as a color image, and in particular, it provides a good image for a color original beyond the color reproduction area of an image output device. The present invention relates to a color image processing device that is obtained.

[0002]

2. Description of the Related Art When a color image signal is processed and output as a color image, there is no problem as long as the target is a color image signal within the range of a color reproduction area of an image output device.
When a color image signal that exceeds the color reproduction area of the image output device is targeted, a color area in which the color is crushed as it is occurs, so that the image quality may be significantly deteriorated. Therefore, in order to solve this problem, a method has been developed in which, for a color image signal that exceeds the color reproduction area of the image output device, data processing is performed so that the color image signal falls within the color reproduction area.

As a conventional color conversion method of a color image for that purpose, for example, as taught in Japanese Patent Application No. 4-180347, a color range of an input image is detected and a color reproduction range of a print engine is used. An example in which an optimum curve table is selected from the density conversion tables depending on how wide it is, and an image signal is converted into data using this table and given to an image output device, or, for example, Japanese Patent Application No.
As taught in Japanese Patent No. 186968, a color image exceeding the above color reproduction range without converting the color component data of the color image data corresponding to most of the color reproduction range of the color image forming apparatus. There is also an example in which the data is converted into color image data corresponding to the position of the closest outermost edge in the space of the color reproduction range. However, any of the methods described above is a method of converting all the color signals that can be taken by the input image so as to be within the color reproduction area of the image output device.

[0004]

As described above, in the conventional color image processing apparatus, since all the color signals that the input image can take are converted into the color reproduction area of the image output apparatus, There are issues to be solved. That is, when the range of colors that the color image signal can take is wide, the entire color range of this color image signal is set within the color reproduction range of the image output device, so that all colors are used evenly. Unless the image is a good image, the colors that are used in a proportionate amount in the image are sacrificed to ensure the reproducibility of other unused colors and colors that are rarely used, and as a result, The point is that the color reproducibility of the image deteriorates.

In other words, depending on the image to be subjected to color image processing, in many cases, it should be the most important,
There is a problem in that the color reproduction characteristics of colors that are often used in an actual input image are sacrificed for color reproduction of colors that are rarely used.

Therefore, an object of the present invention is to ensure that an input color image signal including a color signal that exceeds the color reproduction area of an output device for outputting a color image falls within the color reproduction range of the output device. It is possible to perform color conversion so that color differences between colors that are close to each other can be retained, and a color image processing method that enables good reproduction of subtle color differences, It is to provide a color image processing device that realizes this.

[0007]

Therefore, the present invention takes the following means in order to solve the above problems and achieve the object. A color image processing device that groups input color image signals in a color space, converts each group, and outputs as a color image to a predetermined output device. For each group, color signals belonging to the same group are output. Means for checking the distribution state of the pixels in the image,
A color image processing apparatus, comprising: a means for obtaining a conversion amount of a color signal according to a distribution state of pixels having color signals belonging to the same group in the image, for each group. .

That is, the configuration 1 corresponding to the first and second embodiments of the present invention is as follows. As means for checking the distribution state of pixels having color signals belonging to the same group in each image for each group, the means in the first embodiment is as follows:
A list of color signals included in a group and their representative values, a list of so-called "runs", which are connected components of pixels having color signals belonging to the same group, and a table 1 in which the area of the run is recorded. In the example, the list of color signals included in the group and their representative values, the list of runs,
The table 4 in which the area of the run and the perimeter of the run are recorded corresponds.

Also, for each group, parallel movement (that is, the first method) is used as means for obtaining the color signal conversion amount according to the distribution state of pixels having color signals belonging to the same group in the image. Table 3 and 2 in the first embodiment for giving threshold values for switching between color conversion by the color conversion and color conversion by the movement toward the center (that is, the second method).
Table 5 in the embodiment corresponds.

The configuration 2 corresponding to the first and second embodiments is as follows. As a conversion method in which the relative positional relationship in the color space between the color signals included in the group is invariable, conversion by parallel movement (first method) is also used. As a conversion method in which the positional relationship is not saved, movement toward the center (second method) corresponds to each conversion method.

[0011]

By taking such means, the color image processing apparatus of the present invention has the following operations. (Part 1) When a color signal included in an input image is converted into another color signal and then output, the subjective degree of deterioration in image quality due to the conversion of the color signal depends on the pixel having the converted color signal. , It depends on the distribution in the image. That is, even with the same color reproduction error amount, depending on the appearance state of the pixel having the color signal in the image, it may or may not be easily perceived by the observer. Therefore, for each group, it is recognized how pixels having color signals belonging to the same group, that is, pixels of similar colors are distributed in the image, and according to the distribution,
To reduce the color reproduction error that is easily perceived by the observer,
The conversion amount of the color signal included in each group is determined. Then, the input color image is converted and output based on the conversion amount.

(Part 2) For each group, it is recognized how pixels having color signals belonging to the same group, that is, pixels of similar colors are distributed in the image. Then, according to the distribution, in a color signal of a group in which it is easy to perceive whether or not a subtle color difference is reproduced, the relative positional relationship in the color space between the color signals included in the group is A color signal conversion amount that increases the color signal conversion amount by the conversion method that does not change is determined.

(Part 3) For each group, the area of the connected component of pixels having color signals belonging to the same group, that is, the area of a region in which pixels of similar colors appear in a cluster is calculated. It Then, in each group, the connected component having the largest area is extracted, and the color signal of the group having the larger value has a similar relative positional relationship in the color space between the color signals included in the group. The color signal conversion amount is set to a large value by a different conversion method, and the color signal is converted into a color signal that can be reproduced by the printer.

(Part 4) For each group, the area of the connected component of pixels having color signals belonging to the same group, the complexity of the shape of the connected components of pixels having color signals belonging to the same group, that is, similarities The area and shape complexity of the region in which the pixels of different colors appear in a cluster are calculated. And the larger the area of the connected components belonging to the group,
Further, as the shape is less complicated, the color signal conversion amount is set to a larger value by a conversion method such that the relative positional relationship in the color space between the color signals included in the group does not change, and the color that can be reproduced by the printer is set. Converted to a signal.

[0015]

Embodiments of the present invention will be described below with reference to the drawings. (First Embodiment) FIG. 1 shows the functional arrangement of a color image processing apparatus according to this embodiment.

The camera 101, the A / D converter 102, and the image memory 103 are connected in series. The CPU 104, which controls this apparatus, via the bus 110, in addition to the above-mentioned image memory 103, the RAM 105 and the output selection apparatus 10
6, ROM (A to C) 107a to 107c, and output port
(A to B) 108a to 108c and parameter ROM 111
Connected to and respectively. In addition, the above-mentioned output port
(A to B) 108a to 108c are printers 109
a to 109c.

A camera 101 is a camera for taking a video and converting it into a color image signal, and an A / D converter 102.
Is a circuit for converting this color image signal into digital data, and 8 (8) for each color component of RGB (Red, Green, Blue).
Convert to bit format data.

The image memory 103 is a memory means capable of storing this digital data in a predetermined frame unit, and is also used for holding the digital data converted into color value data. obtain.

The CPU 104 plays a central role in controlling the entire system of the device of the present invention, controls each component of the device, and controls the video signal data (that is, RGB data) stored in the image memory 103. , Lab-based color values L ^* , a ^* , b ^* in the "uniform color space" (each of which has an 8-bit configuration), and the image memory 1 is read again.
It has a function of storing the data in 03.

The RAM 105 is a general-purpose memory device used for the work area of the CPU 104, temporary holding of data, and the like, and is also used as a resident memory for processing programs and the like as necessary.

The output selection device 106 is a selection means for selecting a desired printer from the plurality of printers 109a to 109c for output, and a switch that can be manually set and its set value are set via the bus 110. It is composed of a general-purpose port and the like wired so that it can be referred to from the CPU 104.

The ROMs (A to C) 107a to 107c are read-only memory means corresponding to each of these printers 109a to 109c, and are non-color registers in which color information of areas in which the printer cannot reproduce color is registered. The reproduction area color table is stored for reference.

The output ports (A-B) 108a-108c are
Ports for connecting the system to external devices, which are dedicated to connecting the printers 109a to 109c, respectively.

As the type of printer, an ink ribbon type, an ink jet type, an electrostatic recording type and the like can be applied. The bus 110 is wired so as to efficiently transmit a control signal transmitted / received by the CPU 104 for centrally controlling the system, a data signal, and the like between predetermined components.

The parameter ROM 111 is a means for storing and holding various constants used in a series of processes of the present invention so that they can be read as parameters. Next, the processing operation performed by each of the above-described components will be described. First, all image data is taken in and data conversion is performed.
That is, when a video signal (especially a color image signal) of a subject (not shown) photographed by the video camera 101 is input, the A / D converter 102 digitally converts the video signal to obtain an image memory. Record in 103.

Next, the RGB signals stored in the image memory 103 are sent to the CPU for each pixel via the bus 110.
L ^*, which is read out by 104 and is a uniform color space ^.
The value is converted into a ^* b ^* color system value and recorded again in the image memory 103 via the bus 110. In addition, this R
Since the conversion from the GB color system to the L ^* a ^* b ^* color system is a well-known technique, its detailed description is omitted here.

After this processing is completed, the color signals of the image recorded in the image memory 103 are transferred to the printers 109a-1.
09c, the process of creating a conversion table used when converting to a color signal that can be output to any one of 09c.

FIG. 2 shows a distribution chart of the color signal in the color space. Generally, the distribution of color signals included in a color image in the color space is not uniform and forms some so-called "lumps" as shown in FIG. Therefore, the color signals are grouped for each lump, that is, for each similar color.

When the pixels of the original image are grouped in the same group, the original image is divided into regions. The division state of the area depends on the image, and an area composed of pixels belonging to one group is divided by pixels belonging to another group.
It may be divided into multiple parts. Therefore, for each group, the area of a "run (figure)" that is composed of pixels that belong to the same group and are connected is checked.

It is determined that the color signal C of a certain pixel of interest in FIG. 2 is included in the group i, for example, ΔEi ≤
This is the case when the relationship of D 1 holds. (Note that Δ here
Ei is a color difference ΔE between a color signal included in group i and a color signal having the smallest color difference from C. Also, D
Is a constant previously stored in the ROM 111. ) The color signals of the first pixel unconditionally form a group. After that, the target pixel is sequentially changed, and the color signal C of the target pixel is compared with the color signals already grouped. Then, for example, if there is a color signal Cg whose color difference ΔE is less than or equal to the threshold value D, it is assumed that the color signal C of the pixel of interest belongs to the group g to which this color signal Cg belongs.

On the other hand, among the color signals grouped,
If the color difference ΔE is satisfied by ΔE> D even when compared with any color signal, the color signal C of the pixel of interest is a new group.
g '(not shown) is assumed to be formed.

That is, the color signal C ((
(White dots), other already grouped color signals (black dots)
When the color differences ΔE1 Δ, E2, ... ΔEi, ΔEj with respect to are all larger than D, the color signal C of the pixel of interest forms a new group k.

FIG. 3 exemplifies a table showing a group of colors. The specific grouping is performed by the CPU 1 based on the color signals read from the image memory 103.
By the action of 04, the table 1 as shown in FIG. 3 is created in the RAM 105. Table 1 is composed of a list of color signals included in the group, their representative values, a list of runs, and an area of the runs.

Here, the concept of "run" will be described with reference to FIG. In FIG. 12, the hatched pixels are pixels having color signals belonging to the same color signal group i. Of these, the pixel columns connected to each other are referred to as “connected components”, and a figure having a certain shape is formed. Hereinafter, such a figure will be referred to as a "run".
Then, a number is assigned to each run for distinction.
In other words, the group number and the run number can identify the area where similar colors appear in the image.

Further, the "connection" includes, for example, "4 connection" and "8 connection", but in the present embodiment, description will be made mainly based on "4 connection". In addition, the above-mentioned 4
As with the connections, the device according to the invention can be modified with 8-connected components.

For example, the pixels on the upper, lower, left, and right sides of a certain pixel of interest (hereinafter referred to as "pixel of interest") are referred to as "four neighboring" pixels, but in particular, the pixel of interest among these pixels. Those having color signals belonging to the same group are referred to as "four connected pixels".

FIG. 13 shows "run area" and "perimeter".
FIG. Among the pixels in the drawing, some of the pixels in the four neighborhoods (that is, the upper, lower, left and right) belong to other groups. This is called a "boundary point" pixel. On the other hand, there are some pixels in which four neighboring pixels belong to the same group. This is called a "non-boundary point" pixel.

The "peripheral length" of the run is the number of darkly hatched pixels in the figure. Here, when the “run” is illustrated, pixels having color signals belonging to the group i are connected and continuous to each other to form an elliptic shape as a whole, that is, a hatching (dark hatching, thin hatching). Both) are applied.

The area of the run is the number of pixels forming the run, that is, the number of darkly hatched pixels (boundary points) and the thinly hatched pixels (not boundary points). ) Is the sum of the numbers.

4 (a) to 4 (e) exemplify conceptual views of buffers for registering groups of pixels and codes of runs. The combination of numbers marked in each box in each figure (for example, "1-1") is "group number"-
It means "run number".

The table 1 is created by, for example, a two-dimensional (in FIG. 4) in the RAM 105 in FIG. 1 for recording the code of the group to which each pixel belongs and the run as shown in FIG. 4 (a). By using the buffer structure of (2 rows), the process of grouping the color signals and calculating the area of the run is sequentially performed for each pixel from the upper left pixel to the lower right pixel of the image to be processed. Done.

Below, how the group number and the run number of the target pixel (thickly hatched) to be processed is determined will be described. FIG.
In (a), the color signal of the pixel of interest is not included in any of the groups so far, so a new group 6 is created and registered as the first run.

In FIG. 4B, the group is different from the left side of the pixel of interest, and the upper side is the same group 1, so the same run number as the upper side is added and registered as 1-1. In FIG. 4 (c), the left and upper sides of the target pixel are in the same group 1, the run number is 2 on the upper side and 1 on the left side.
So, take the smaller one and register it as 1-1.

In FIG. 4 (d), it was found that run 1 and run 2 were connected, so the run numbers of the other pixels forming run 2 were changed, and these two runs were integrated. Make a series of runs 1-1.

In FIG. 4E, the color signal of the target pixel belongs to group 6. Since the upper side and the left side are different groups, as a new run 2 for group 6,
Register 6-2.

After this, L ^* a ^* b ^* tristimulus values (L ^*
_i , a ^* _i , b ^* _i ) is denoted by Ci. The color signal C (L ^* , a ^* , b ^* ) of the pixel is read from the image memory 103 of FIG. 1 and is represented by the following equation 1 from the color signal representative value Cmi of each group already registered in Table 1. Calculate the color difference Di. The color signal representative value Cmi of each group is a color signal composed of an average value of L ^* , a ^* , and b ^* components of all color signals Cij included in the color signal list of the corresponding group. ^{Di = {(L * - L} * mi) 2 + (a * - a * mi) 2 + (b * - b * mi) 2} 1/2 ··· ( Equation 1) group Di is minimized g The color difference Dgj from the color signal Cgj included in the color signal list of 1 is sequentially calculated. If there is a color signal Cgj in which Dgj <= D in the color signal list, it is determined that the color signal C belongs to the group g. here,
D is a constant for determining whether or not it is considered that the two color signals belong to the same group, and is predetermined and recorded in the parameter ROM 111. This group g is compared with the group numbers gu and gl of the pixels on the upper side and the left side of the relevant pixel recorded in the buffer, and if there is an equal one, the run number r recorded there is compared with the run number r of the relevant pixel. Name it as a run number (see Fig. 4 (b)), and add 1 to the area of the corresponding run in Table 1 (increment). For example, in the case of FIG.
1 is added to the area s11 of.

If the same group number is recorded on both the upper and left sides and different run numbers are recorded, the smaller run number rs is adopted (see FIG. 4 (c)).
In addition, the larger run number recorded in the buffer
Replace rl with the smaller run number rs (Fig. 4 (
Refer to (d)), and update the area of the run corresponding to rs and rl in Table 1 as follows. That is, Sgrs and Sgrl are added, and 1 is added to this to make a new Sgrs, and Sgrl
To 0.

When the group g is different from the group number of any of the upper and left pixels, the highest run number belonging to the group number g in Table 1 is 1
Add to the run number and record in the buffer (Fig. 4
(See (e)). Further, a new run is added to the table 1 and 1 is registered in the corresponding area.

Next, a color signal whose color difference Dgj is 0, that is,
Check the color signal list to see if there is a color signal equal to C. If not, add C to the color signal list of group g,
After recalculating the representative value of the color signal, the process proceeds to the next pixel.

When the color signal Cgj satisfying Dgj ≤ D is not in the list, the group whose color difference Di from the representative value is the next smallest
The above process is performed for g '. Hereinafter, the process is repeated in the order of smaller color difference Di from the representative value to compare all the color signals of the groups registered in Table 1 with D
If there is nothing that satisfies ij ≤ D, the number g "obtained by adding 1 to the maximum group number registered in Table 1
Add the group of to the color signal list and color signal representative value.
Register C and set the area of run number 1 to 1 (Fig. 4 (a)
See).

Also, the group number is set at the corresponding position in the buffer.
g ", the code of run number 1 is recorded, and the process proceeds to the processing of the next pixel. Next, the stored contents of the second line of the buffer are copied to the first line, and then the contents of the second line are cleared. Then, the process shifts to the process related to the next row of the image memory 103.

The above series of processing is repeated until the last pixel. Next, the area of each run in Table 1 is normalized by the area of the image, that is, the total number of pixels of the image. Hereinafter, all areas are normalized areas.

By these processes, the table 1 is completed, the color signals included in the image are divided into groups, and the area of the run of pixels belonging to the same group is grasped for each group.

It is easy for an observer to perceive whether or not a subtle difference in color is reproduced, because pixels having these colors appear in a specific area in a cluster in the original image. If there is. In other words, the area of the run that occupies the maximum area in each group is considered to be a measure for determining the reproducibility for subtle color differences when converting the color signals of that group, and the run with a large area belongs to it. It is considered that the subjective color reproducibility can be improved in the group by the process for improving the reproducibility for the subtle color difference. Therefore, a conversion table for converting the color signal of each pixel recorded in the image memory 103 into a color signal that can be output to the printer is
Follow the procedure below to create.

FIG. 10 shows a flow chart of the determination processing relating to the color reproduction of the present invention, and this series of processing procedures will be described with reference to FIG. 1 related to this flow chart.

First, the first step (S1) is executed, but the process does not branch to the end and the process proceeds to the next step S2. That is, it is checked whether or not the color signals belonging to each group exceed the color reproduction range of the printer, and it is determined whether or not the color signals that can be output by the printer need to be converted (S2). Specifically, all color signals included in the color signal list for each group are compared with the color reproduction range of the printer, which is written and stored in the ROM (107a, 107b, or 107c) corresponding to the printer. However, if any color signal that cannot be reproduced by this printer is included, conversion processing is performed on all the color signals belonging to that group, and the color signals that can be reproduced by the printer are converted. Processing is performed (details will be described later).

Further, since the conversion processing as described above is performed independently for each group, it is determined whether or not all the existing groups have been processed in order to shift to the processing for the next group (S10). There is a first step (S
Return to 1).

Each ROM for storing the color reproduction range of the printer (that is, 107a, 107b, 107 in FIG. 1).
c) is arranged in each of the printers (i.e., 109a, 109b, 109c in FIG. 1) connected in correspondence with them, and according to the selection of the desired printer made by the user at the time of operation, It is switched and used by the CPU 104.

Color signals exceeding the color reproduction range of the printer are
It is converted into a color signal that can be output by the printer as follows. Hereinafter, the region on the color space occupied by the color signals belonging to the group will be described on the assumption that the color signals belonging to the corresponding group are represented by “spheres” having the smallest volume.

A process for converting a color signal that exceeds the color reproduction range of the printer into a color signal that can be output by the printer in the color region represented by a sphere is shown in the conceptual diagrams of FIGS. 5 (a) to 5 (b). As shown, the following two processing methods are adopted.

That is, the "first method" is a process of moving all the color signals in parallel in the direction of the arrow toward the fixed point C on the color space (FIG. 5 (a)). Also, "Second
The method "is a method of moving toward the center (o) of the sphere in proportion to the distance from the center (o) (FIG. 5 (b)). FIG. 5 shows the color space two-dimensionally,
The actual processing is performed in the three-dimensional space.

In this first method, since the relative relationship between the color signals in the group does not change, a subtle difference in color in the same group can be expressed as it is. If this is done, the color changes overall in a certain direction, causing so-called "color shift", and subjective color reproducibility deteriorates.

On the other hand, in the second method, the subtle difference in color within the same group cannot be distinguished, and the subjective color reproducibility is deteriorated. However, this method causes an overall "color shift". Absent.

By combining the above two processing methods with each other and converting into a color signal which can be printed out by a desired printer, subjective color reproducibility can be improved. Specifically, a color signal that exceeds the color reproduction range of the printer is converted to a color signal that falls within the color reproduction range of the printer while repeatedly making a slight change to determine whether or not it falls within the color reproduction range of the printer. It

Therefore, first, by the first method, the color signal is gradually added to the fixed point C on the color space (in the color reproduction range).
When the total amount of movement that has moved exceeds a predetermined threshold value, the second method described above is then used to move concentrically toward the center (o) of the sphere. In this process, it is better to reproduce subtle color differences better in terms of subjective color reproducibility, that is, as described above, for groups with large area runs. It is considered that the subjective color reproducibility is improved by setting a large threshold value of the movement amount by the first method.

The following series of processing is performed for a group including color signals exceeding the color reproduction range of the printer.
It is done for each group. Here, a specific description will be given of the method described above.

First, a sphere including all the color signals Cgj included in the group and having a minimum volume is determined in the color space, and the coordinates (L ^* C, a ^* C, b ^* C) of the center thereof are RA.
It is recorded and held in M105. All color signals Cgj included in the group are moved in parallel toward the fixed point C (that is, the first method) by the movement amount (ΔL, Δa, Δb) determined by d, as shown in the following Expression 2. (S3).

Then, it is judged whether or not it falls within the color reproduction range of the printer (S4). However, the fixed point C is also a predetermined color signal, and it is desirable that the fixed point C is a color signal representing a gray of a medium brightness. The value of the color signal representing the fixed point C is stored and held in the parameter ROM 111.
The initial value of d is 1. And the original color signal C
The table 2 in FIG. 7 shows gj (L ^* , a ^* , b ^* ) and the conversion destination color signal Cgj '(L ^* ', a ^* ', b ^* ') converted by the following equation 2 as one set. save.

L ^* _gj ′ = L ^* _gj + ΔL a ^* _gj ′ = a ^* _gj + Δa (Equation 2) b ^* _gj ′ = b ^* _gj + ΔbΔL = (L ^* _O− L ^* C _{) / P O × d Δa =} (a * O -L * C) / P O × d Δb = (b * O -L * C) / P O × d PO = ((L * O -L * C) ² + (a ^* _O -a ^* C) ² + (b ^* _O -b ^* C) ² ) ^1/2 ΔE = (ΔL ² + Δa ² + Δb ² ) ^1/2 (Equation 3) If If all the color signals have been converted into the color reproduction range of the printer by the parallel movement of Expression 2, the process of the next group is performed (S11). If there is at least one color signal that exceeds the color reproduction range of the printer, d is incremented by 1 (S5), the total movement amount ΔE shown in Equation 3 is calculated, and the threshold value is calculated. Compare with the value ΔEth (S
6).

Here, ΔEth is a threshold value of the movement amount for switching the conversion by the first method and the second method, and is a monotonically increasing function of the area of the run. This value is set as a value corresponding to the area of the run in a table as shown in Table 3 of FIG. 6, and is stored in the parameter ROM 111. In other words, the larger the area of the run,
Since the large threshold value ΔEth is adopted, it is possible to increase the movement amount in the first method, and as a result, the delicate color difference within the same group is well preserved. Actually, first, of the runs belonging to the group to be processed, the area of the run having the maximum area is read from Table 1, and then the value of ΔEth corresponding to the area of this run is read from Table 3. It is read from the parameter ROM 111 and used.

If ΔE is smaller than ΔEth, all the color signals included in the group are translated (that is, the first method) (S3) by the processing represented by the above equation (2).
Then, a process (S4) of determining whether or not it falls within the color reproduction range of the printer is performed.

If ΔE becomes larger than ΔEth, the color signal conversion method is switched to the method of moving the sphere representing the group toward the center (that is, the second method). Therefore, first, a sphere including all the conversion destination color signals Cgj ′ currently stored in the table 2 and having the minimum volume is determined in the color space, and the coordinate of the center (L
^(* C, a ^* C, b ^* C) are held in the RAM 105. Then, according to the processing shown in the following Expression 4,
Convert all Cgj 'held in
The corresponding column of the conversion destination color signal Cgj 'is overwritten and recorded (S7). However, here, ΔT is a so-called “step” of a predetermined reduction rate, and is stored in the parameter ROM 111. Further, e means the number of repetitions, and its initial value is 1.

After that, it is checked whether or not all the Cgj's are converted so as to be within the color reproduction range of the printer (S8). If Cgj 'is within the color reproduction range of the printer, the process proceeds to the next group (S12).

On the other hand, if it is not within the color reproduction range, e is incremented by 1 (S9) and the same processing is repeated from step S7. L ^* _gj '' = L ^* _gj '+ ΔT × (L ^* _C− L ^* _gj ′) × e a ^* _gj ″ = a ^* _gj ′ + ΔT × (a ^* _C− a ^* _gj ′) × e .. (Equation 4) b ^* _gj '' = b ^* _gj '+ [Delta] T * (b ^* _C- b ^* _gj ') * e By the above processing, the color signal Cij of the given image can be output by the printer. A table for converting to the color signal Cij ′ is created in Table 2.

The image stored in the image memory is output to the printer as follows. The color signal C read out pixel by pixel from the image memory is converted into a conversion destination color signal C ′ corresponding to the color signal C in the table 2,
Finally, it is sent to the corresponding printer and printed out.

In the above first embodiment, the color signals included in the image are divided into groups in the color space, and the color signal conversion processing is performed only on the groups to which the color signals exceeding the color reproduction range of the output device belong. In doing so, when converting to a color signal that can be reproduced by the output device, the size of the run of pixels having color signals belonging to the same group is checked for each group, and a group having a larger run produces a slightly different color difference. By performing color conversion processing so that it is saved, subtle color differences are reproduced as they are, in areas where similar colors appear as lumps in the image, resulting in good subjective color reproducibility. A printed matter can be obtained as output.

(Second Embodiment) The configuration of the color image processing apparatus according to the second embodiment of the present invention is basically the same as the basic configuration illustrated in the first embodiment. Therefore, description of each element of the basic configuration is omitted.

However, in the above-described first embodiment, when determining the reproducibility of a subtle color difference, processing is performed using the "run area" which is composed of connected components of pixels having color signals belonging to the same group. On the other hand, the present embodiment is characterized in that the processing is performed in consideration of not only the "area" of the run but also the "shape complexity" of the run.
The reason why the processing method having such characteristics is necessary is that if the figure occupies the same area (that is, the run), the simpler the shape is, the more subtle color difference is reproduced. It is nothing but because it is easily perceived by the observer.

Therefore, in the second embodiment, a table 4 having a format as shown in FIG. 8 is created as a table corresponding to the table 1 of the first embodiment. That is, in the table 4, the run, the area, the perimeter, the color signal list, and the color signal representative value are set in a table format as shown for each color group. That is, the point that this table 4 is different from the above-mentioned table 1 is that in the table 4, the "peripheral length of the run" is measured and the value is held.

For the perimeter of the run described above, the buffer table consisting of two lines for recording the group number and the run number illustrated in FIGS. 4 (a) to 4 (e) in the first embodiment is further expanded, In this embodiment, the calculation can be performed by using a buffer table composed of three rows as shown in FIG. Specifically, paying attention to the pixels in the second row of the buffer table of these three rows, the pixels in the “four neighborhoods” (that is, in the four different directions) are examined, and when they are connected to different runs, The pixel is regarded as a "boundary point" (that is, a pixel with a dark hatching in the figure), and the counter Pij for the corresponding perimeter is incremented by one.

In this case, all the pixels located at the outermost periphery of the image are treated as boundary points. Moreover, in FIG.
As shown in d), when the runs are combined by connecting the runs with the same group number but different run numbers, as in the calculation of the area of the runs, Also add the perimeter. For example, when the run 1 of group 1 and the run 3 of group 1 are integrated,
The perimeter p11 of the new run 1 is set to p11 + p13, and the perimeter p13 of the new run 3 is set to 0. These respective processes are repeated for all pixels to complete the predetermined table 4.

The "shape complexity" of a run is the area S of the run.
And the perimeter P of the run, the degree of dispersion V calculated based on the following equation 5 can be used for evaluation. V = P ² / S (Equation 5) Therefore, the normalized area of the run in the first embodiment is
Corresponding to the measure for determining the reproducibility for the subtle color difference, in the present embodiment, the reproduction for the subtle color difference of each group is performed by the importance W _ij calculated by the following equation 6. Determine sex. However, S _ij is j of group i
The area of the th run, S ₀ is the total number of pixels in the image, k is a predetermined constant, and Vi is the degree of dispersion of the j th run of group i. Importance W of each of these runs
Among _ij , the maximum value W ^max i in each group is used as a “measure” for determining the reproducibility of the color signals of the group with respect to subtle color differences.

W _ij = S _ij / S ₀ + k / V _ij (Equation 6) Further, corresponding to the table 3 according to the first embodiment, in this embodiment, the table 5 shown in FIG. 11 is used. , ΔEth is obtained as the threshold value of the amount of movement due to the parallel movement. Table 5
Is the maximum value W ^max i and ΔE of the importance in each group.
It represents the correspondence of th, and the predetermined value is
It is stored in the parameter ROM 111.

Except for the matters described above, a conversion table for converting the color signal Cij of the input image into the color signal Cij 'that can be output to the printer by performing the same processing as in the first embodiment ( That is, (corresponding to the table 2 according to the first embodiment) is created. The output of the image to the printer is performed in the same manner as in the first embodiment.

According to the second embodiment described above, it is observed whether or not a subtle color difference is reproduced by considering not only the areas of runs of similar colors but also the complexity of the shape of the runs. It is possible to more accurately grasp the color signal in the area that is easily perceived by a person, and to perform conversion processing within the color reproduction range of the printer used. As a result, it is possible to achieve better subjective color reproducibility than in the past.

(Modified Embodiments) In addition to the first and second embodiments described above, various modifications of the color image processing apparatus of the present invention are possible without departing from the scope of the present invention. It can be implemented.

[0087]

According to the present invention, the color signal of the input image is converted so that the color reproduction error that is easily perceived by the observer of the color image is reduced, so that the same color reproduction error amount is obtained. An output with little deterioration in image quality can be obtained. In addition, the color signal in the region where it is easy to perceive whether or not a subtle color difference is reproduced is converted with a smaller change in the relative positional relationship in the color space, and thus a subtle color difference is also good. Output is obtained.

Although the description has been given based on the embodiments of the present invention, the present invention includes the following inventions. (1) A color image processing device that groups input color image signals in a color space, converts each group, and outputs a color image to a predetermined output device,
For each group, a means for examining the distribution state of pixels having color signals belonging to the same group in the image, and for each group, pixels having color signals belonging to the same group in the image And a means for obtaining the conversion amount of the color signal in accordance with the distribution state of the color image processing apparatus. This (1) corresponds to, for example, the first and second embodiments.

(Operation 1) When the color signal included in the input image is converted into another color signal and output, the subjective degree of image quality deterioration due to the conversion of the color signal is It depends on what distribution the pixels have in the image. That is, even with the same color reproduction error amount, depending on the appearance state of the pixel having the color signal in the image, it may or may not be easily perceived by the observer. Therefore, for each group, it is recognized how pixels having color signals belonging to the same group, that is, pixels having similar colors are distributed in the image, and the observer is notified according to the distribution. The conversion amount of the color signals included in each group is determined so that the color reproduction error that is easily perceived is reduced. Then, the input color image is converted and output based on the conversion amount.

(Effect 1) Since the color signal of the input image is converted so that the color reproduction error that is easily perceived by the observer is reduced, the subjective deterioration of the image quality is reduced even with the same color reproduction error amount. be able to.

(2) The means for obtaining the conversion amount of the color signal, according to the distribution state of the pixels having the color signal belonging to the same group in the image for each of the groups,
A color signal is obtained by a combination of a conversion method in which the relative positional relationship between the color signals included in the group is invariable in the color space and a conversion method in which the relative positional relationship in the color space is not preserved. The color image processing device according to (1), characterized in that the conversion amount is calculated. This (2) corresponds to, for example, the first and second embodiments.

(Operation 2) For each group, it is recognized how pixels having color signals belonging to the same group, that is, pixels having similar colors are distributed in the image. Then, according to the distribution, in a color signal of a group in which it is easy to perceive whether or not a subtle color difference is reproduced, the relative positional relationship in the color space between the color signals included in the group is A color signal conversion amount that increases the color signal conversion amount by the conversion method that does not change is determined.

(Effect 2) Since the color signal in the region where it is easier to perceive whether or not a subtle color difference is reproduced, the conversion is performed so that the change in the relative positional relationship in the color space is reduced. Subtle color differences are reproduced well, and subjective deterioration of image quality can be reduced.

(3) The means for investigating the distribution state in the image is a means for obtaining, for each of the groups, the "area" of the connected component of pixels having color signals belonging to the same group. The larger the area of the connected components of the pixels having the color signals belonging to the group, the larger the conversion amount of the color signals by the conversion method such that the relative positional relationship in the color space becomes invariant. The color image processing device according to (2). This (3) corresponds to, for example, the first embodiment.

(Operation 3) For each group, the area of the connected component of pixels having color signals belonging to the same group, that is, the area of a region where pixels of similar colors appear in a cluster is calculated. It Then, in each group, the connected component having the largest area is extracted, and the color signal of the group having the larger value has a similar relative positional relationship in the color space between the color signals included in the group. The color signal conversion amount is set to a large value by a different conversion method, and the color signal is converted into a color signal that can be reproduced by the printer.

(Effect 3) Color signals that exceed the color reproduction range of the output device are converted into color signals within the reproduction range and output. At this time, if the subtle color differences are not reproduced in the region where pixels of similar colors appear in a cluster, the image quality perceived by the observer is deteriorated. On the other hand, when pixels with similar colors appear everywhere in the image, the subjective image quality is not significantly deteriorated even if a subtle difference in color is not reproduced. Therefore, conversion is performed such that a color signal of a group in which a connected component of pixels having color signals belonging to the same group has a larger area, the relative positional relationship in the color space between the color signals included in the group remains unchanged. By setting a large amount of color signal conversion by the method, subtle color differences can be reproduced without being crushed, for example, in areas where pixels with similar colors appear in a cluster in one place. To be done. As a result, good subjective color reproducibility can be obtained.

(4) The means for checking the distribution state in the image is, for each of the groups, the "area" of the connected component of the pixels having color signals belonging to the same group and the "shape complexity of the connected component". The larger the area of the connected component of the pixels having the color signals belonging to the same group and the smaller the complexity of the shape of the connected component, the more the relative positional relationship in the color space is invariant. The color image processing apparatus according to (2), characterized in that the conversion amount of the color signal by such a conversion method is increased. This (4) corresponds to, for example, the second embodiment.

(Operation 4) For each group, the area of the connected component of pixels having color signals belonging to the same group, the complexity of the shape of the connected components of pixels having color signals belonging to the same group, that is, similar The area and shape complexity of the region in which the pixels of different colors appear in a cluster are calculated. And the larger the area of the connected components belonging to the group,
Further, as the shape is less complicated, the color signal conversion amount is set to a larger value by a conversion method such that the relative positional relationship in the color space between the color signals included in the group does not change, and the color that can be reproduced by the printer is set. Converted to a signal.

(Effect 4) It is easy to perceive whether or not a subtle color difference is reproduced in a region in which pixels of similar colors appear in a cluster, and the subjective shape of the region. The simpler the shape, the easier it is to perceive whether or not a subtle color difference is reproduced.
Affects subjective image quality. Therefore, the area of the connected component of pixels having color signals belonging to the same group is large,
By increasing the color signal conversion amount by a conversion method that makes the relative positional relationship between the color signals included in the group in the color space invariant in a group that includes simple connected component shapes. , similar color pixels, for example, in areas such as have appeared waiting masses one places, because it is reproduced "without collapse" in subtle differences in color, good subjective color reproducibility obtained To be

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a color image processing apparatus according to this embodiment.

FIG. 2 is a distribution diagram of color signals in a color space.

FIG. 3 is a table 1 showing a table of color groups.

4A to 4E are conceptual diagrams of a two-line buffer for recording a color group number and a run number, and a recording example thereof.

5A is a conceptual diagram of a first method for converting a color signal within a color reproduction range of a printer, and FIG. 5B is a conceptual diagram of a second method thereof.

FIG. 6 is a table 3 showing the relationship between the area of color runs and the threshold value ΔEth of the movement amount of a predetermined function.

FIG. 7 is a table 2 showing the correspondence between color signals and conversion destination color signals for each color group.

FIG. 8 is a table 4 showing a table of runs, areas, perimeters, color signal lists, and color signal representative values for each color group.

FIG. 9: 3 for recording group number and run number
A conceptual diagram of a row buffer.

FIG. 10 is a flowchart of determination processing relating to color reproduction according to the present invention.

FIG. 11 is a table 5 showing the relationship between the importance and the threshold value ΔEth of the movement amount.

FIG. 12 is an explanatory diagram of labeling processing for color signals.

FIG. 13 shows boundary points and non-boundary points, and
Explanatory drawing about the area and circumference of a run.

[Explanation of symbols]

101 ... Camera, 102 ... A / D converter, 103 ... Image memory, 104 ... CPU, 105 ... RAM, 106 ... Output selection device, 107a ... ROM (A), 107b ... ROM
(B), 107c ... ROM (C), 108a ... Output port
(A), 108b ... Output port (B), 108c ... Output port (C), 109a ... Printer, 109b ... Printer, 1
09c ... Printer, 110 ... Bus, 111 ... Parameter ROM.

Continuation of front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display location H04N 1/46 H04N 1/46 Z 9/79 9/79 H

Claims (3)

[Claims] 1. A color image processing device for grouping input color image signals in a color space, converting each group and outputting as a color image to a predetermined output device, wherein each group has the same group. Means for investigating the distribution state of the pixels having the color signal belonging to the same image in the image, and the conversion of the color signal according to the distribution state of the pixels having the color signal belonging to the same group in each image for each group A color image processing apparatus comprising: a means for obtaining an amount. 2. The means for obtaining the conversion amount of the color signal includes, for each of the groups, a color signal included in the group according to a distribution state of pixels having color signals belonging to the same group in the image. Determining a conversion amount of a color signal by a combination of a conversion method in which the relative positional relationship in the color space remains unchanged and a conversion method in which the relative positional relationship in the color space is not preserved. The color image processing apparatus according to claim 1, wherein: 3. The means for checking the distribution state in the image is a means for obtaining, for each of the groups, an “area” of connected components of pixels having color signals belonging to the same group. A group having a larger area of connected components of pixels having color signals belonging to the group has a larger conversion amount of the color signal by the conversion method such that the relative positional relationship in the color space becomes invariant. The color image processing apparatus according to claim 2.