JP2746897B2 - Image processing method - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method, and more particularly to an image processing method for converting an original image into an image having a special effect by reducing the gradation.

[Prior Art] With respect to digital image processing, various devices and various methods have been developed in various fields, and natural images such as photographs are input as digital image data to perform some processing to create creative images. It is possible to produce. One of the processes is gradation conversion.

Conventionally, in this gradation conversion, the gradation of the original image is divided at a certain ratio, and all the data in each divided gradation are divided by one.
There is a method of replacing with a representative value (generally called posterization).

In addition, a method of normalizing an output gradation in which a gradation part with a high input frequency is increased and a gradation part with a low input frequency is reduced in output gradation in accordance with the frequency of the input gradation is known. It is known that a rich gradation surface is possible.

[Problems to be Solved by the Invention] However, in the posterization method in the above-mentioned conventional example, there are many input gradations due to the method of equally dividing the gradations. In other words, there are many similar colors or many parts with similar brightness.
Therefore, the input image has a problem that the area where the gradation changes is small and the special effect of lowering the gradation, which is the original purpose, is lost. That is, when the frequency of the predetermined gradation is high, the portion where the gradation changes is reduced, and the effect of posterization is not seen.

Further, when the cumulative frequency value corresponding to each input tone is normalized using the cumulative frequency distribution, the output tone for a portion having a high input frequency is subdivided, and the portion having a low frequency is conversely largely divided. Therefore, although the above-mentioned problem does not occur, the data in which the input gradation is concentrated is expanded in the output gradation, so that the original color and brightness of the input image are changed. There is a problem that the difference between the images is too large.

The present invention has been made to solve the above problems,
An object of the present invention is to provide an image processing method capable of satisfactorily realizing the effect of posterization when performing a process of lowering the gradation of input image data.

It is another object of the present invention to provide an image processing method that preserves the original color and brightness of any original image and can output a posterized image without a sense of incongruity.

It is another object of the present invention to provide an image processing method capable of outputting a posterized image in which features of an original image are stored.

It is another object of the present invention to provide an image processing method capable of outputting a posterized image in which a truly characteristic portion of an original image is stored.

Further, the present invention provides an image processing method capable of outputting a posterization image in which features of an original image are stored.

Further, the present invention provides an image processing method capable of outputting a posterization image independent of an original image.

[Means and Actions for Solving the Problems] In order to solve this problem, the present invention provides an image processing method for lowering the gradation of image data, comprising the steps of: Then, the gradation is divided for each predetermined cumulative frequency, a representative value is selected from the divided gradation area, and output as output image data.

In the image processing method for reducing the gradation of image data, a continuous gradation region having a frequency equal to or less than a predetermined value is divided into one gradation from a gradation histogram of the input image data. Is output as output image data.

Also, there is provided an image processing method for lowering the gradation of image data, wherein a gradation histogram of the input image data is created, and when the input image data is converted into image data having a designated number of gradations, the gradation is calculated. By determining a gradation area in which the occurrence frequency is equal to or more than a predetermined value in the histogram, and comparing the accumulated frequency in the determined gradation area with a predetermined number, a representative value of the gradation area is output as output image data. Determine whether or not.

Embodiment First, in order to clarify the features of the present embodiment, a posterization according to a conventional example will be described in detail.

6 and 7 show an example of a conventional posterization method. In the figure, the horizontal axis indicates the input gray scale, and the vertical axis indicates the output gray scale. FIG. 6 shows a normal case in which an input image is output as it is, and FIG. 7 is a diagram in which when the input is 32 gradations (0 to 31), the output is 4 gradations. Output value is 0 for input value (0-7), output value is 8 for input value (8-15)
Means to replace with a representative value.
This is the general posterization method.

For example, consider a case where an image is input at 256 gradations (0 to 255) and output at 8 gradations. FIG. 12 shows a histogram of the input image, in which the input gradation is equally divided into eight. The ratio of the area of each divided histogram is equal to the ratio of the area of each gradation of the output image. Therefore, the portion of the input gradation value 160 to 191 is, for example, the representative value 160
In this case, most of the output image has a gradation value of 160, the portion where the gradation changes is reduced, and the effect of posterization is not seen.

FIG. 8, FIG. 9, FIG. 10, and FIG. 11 show an example of a method of normalizing the output gradation in accordance with the frequency of the input gradation.
In FIG. 8, the horizontal axis represents the gradation and the vertical axis represents the frequency, and the frequency of certain data for each gradation is shown. In FIG. FIG. 9 is showing an input gradation and output gradation of the same data and FIG. 8, the histogram of Figure 8, that data between the input tone value a ₁ of a ₂ are concentrated Can be read. This results in that the data are concentrated between In Figure 9 the output tone value b ₁ of b _2, the gradation is not richly expressed. Therefore, seeking cumulative frequency as shown in FIG. 10, by changing placed on the output tone value as intact Fig. 11 the cumulative frequency value corresponding to each input tone value, from high input frequency a ₁ input gradation portion between a _2, which is converted into an output gradation from c ₁ to c _2.

Also, using the cumulative frequency distribution, the cumulative frequency value corresponding to each input tone is normalized to 0 to 255, the value is divided into eight equal parts, and the minimum tone value in the divided area is determined as a representative value of each divided area. FIG. 13 is a diagram showing the relationship between the input gradation and the output gradation when used.
The output gradation for the part with high input frequency is subdivided, and the part with low frequency is conversely divided into large parts, so that there is no problem mentioned above, but the input gradation is concentrated around 160 to 191 Since the data is spread between 64 and 223 in the output gradation, the original color and brightness of the input image change, and the difference between the input image and the output image becomes too large. Occur.

First Embodiment Hereinafter, a first embodiment of the present invention will be described in detail.

FIG. 1 is a block diagram of an image processing apparatus of the present embodiment for realizing the image processing method of the present invention. FIG. 2 is a block diagram of the hardware of the image processing apparatus of the present embodiment.

An image input unit 1 inputs an original image. In FIG. 2, an image input device 11 (for example, a color scanner,
Camera) and the A / D converter 12. Reference numeral 2 denotes a histogram calculation unit that creates a histogram of the input image. Reference numeral 3 denotes a division unit that divides the histogram at every certain cumulative frequency. Reference numeral 4 denotes an output gradation specifying section for specifying the output gradation of the input gradation with respect to the input gradation, which corresponds to the character value input method 16 in FIG. Reference numeral 5 calculates a representative value from the range of each gradation value of the histogram divided by the output gradation calculation unit. Reference numeral 7 denotes an image output unit for outputting a processed image, which corresponds to a D / A converter 17 and an image output device (for example, a color CRT, a photo printer, etc.) in FIG.
8, 9, and 10 indicate signal lines of image data, and indicate that input image data is handled as RGB luminance data.

2 is executed by the CPU 14 shown in FIG. 2 according to a program stored in the ROM 15b in the memory 15, and the input image data is stored in the image memory 13. Is stored in RAM15a and ROM in CPU14
The RAM 15a stores information input from a character and numerical value input device 16 (for example, a keyboard, a digitizer, etc.), and the ROM 15b stores a processing program of the CPU 14.

 Next, the contents of the processing will be specifically described.

Image memory 1 from image input device 11 through A / D converter 12
For the RGB image data input to 3, the histogram calculator 2 creates a histogram for each of the RGB. FIGS. 3A to 3C show examples of the created histograms. The horizontal axis indicates the gradation, and the vertical axis indicates the frequency.

Here, the desired number of tones of the output image is input from the output gradation specifying unit 4, and the division unit 3 divides the histogram by (total frequency) ÷ (number of tones). By expressing a frequently-used part using many luminance (gradation) values, an effect that looks like the color and lightness of the original image smoothly without an unpleasant color change on the output image is aimed. It is a thing. FIGS. 4 (a) to 4 (c) show the state of equally dividing.

Here, a representative value of the output gradation is calculated within the range of the input gradation values divided by the output gradation calculation unit 5. In this embodiment, the average value is used as the representative value of the divided area. However, similar results can be obtained by using the minimum value, the maximum value, and the like. For example, in FIG. 4A, the representative value is (a + b) ÷ 2 in a certain divided area (shaded area in the figure). As described above,
An output image having a desired gradation with respect to the input image can be obtained. 5 (a) to 5 (c) show the gradation relationship between the input image and the output image when the histogram of the input image is shown in FIG. 3 and the output image has 8 gradations.

Second Embodiment Next, a second embodiment will be described. As in the embodiment of FIG. 1, an image is input from the image input unit 1, a histogram is created by the histogram calculation unit 2, and the histogram is generated by the division unit 3 in accordance with the number of gradations designated by the output gradation designation unit 4. In this case, the histogram is divided by (total number of pixels) / (specified gradation).

Here, a threshold is obtained from a threshold generator (not shown). In the present embodiment, a preset value is stored in the memory 15 in the CPU 14.
, And the value is called, but the threshold value can be determined interactively via the character / numerical value input device 16 while observing the status of the output image. The threshold is used for extracting a portion indicating a feature in the histogram, and is represented by a frequency value.

As shown in FIG. 14, the output tone calculating unit 5 calculates a representative value of the output tone within the range of each divided value, as shown below, by focusing on the portion of the histogram equal to or higher than the threshold value. I do.

Input tone values output tone values _{0 ~ (d 1 -1) (} d 1 -1) / 2 d 1 ~ (d 2 -1) (d 1 + d 2 -1) / 2::: d 5 ~ ( _{d 6 -1) (d 5 +} d 6 -1) / 2 d 6 ~ 255 (d 6 + 255-1) / 2 or less similarly to determine a representative value of an output gradation of each divided region. As can be seen from FIG. 14, there are eight regions divided by the division unit 3, but the representative value of the output gradation is reduced to seven by using the threshold. When a histogram as shown in FIG. 15 is calculated, the representative value is 10.

As described above, it is possible to automatically change the output gradation depending on the scene of the input image and to perform more preferable posterization processing. Although the average value between the input tone values was used as the representative value, the maximum value and the average value were used as in the first embodiment.
Even if the minimum value is used, the same effect can be obtained.

Third Embodiment Next, a third embodiment will be described. As in the first embodiment, an image is input from the image input unit 1, the desired number of gradations is input from the output gradation designation unit 4, and a histogram is created by the histogram calculation unit 2. Here, a threshold is obtained from a threshold generator (not shown). In the present embodiment, a preset value is called from the memory as in the second embodiment. However, it is also possible to interactively determine the threshold value via the character / numerical value input device 16 while viewing the status of the output image. It is. The threshold value is a frequency value as in the second embodiment.

FIG. 16 is a diagram for explaining the present embodiment, in which the number of desired output gradations is eight. The horizontal axis shows the input gradation and the vertical axis shows the frequency. Paying attention to the part above the threshold in the histogram,
First, a portion where the value equal to or smaller than the threshold value is continuous is counted. No. 16
In the figure, (φ ~ e ₁ ), (e ₄ ~ e ₈ ), (e ₉ ~ e ₅ ), (e ₇ ~ 25
5) 4 sections. Next, the total number of frequencies equal to or greater than the threshold is counted. The division number of the portion exceeding the threshold value is obtained by subtracting the count number of the portion where the value equal to or less than the above-mentioned threshold value continues from the designated desired output gradation number, and the division number is 4.

Next, from the value obtained by dividing the total frequency equal to or higher than the threshold value by the number of divisions, each part (e _{1 to} e ₄ ), (e _{5 to} e ₉ ),
In (e ₅ ~e ₇₎ of the sum of the frequency of that portion removing smaller than a predetermined value. This is for ignoring fine parts such as a part X surrounded by a dotted circle in FIG.

By repeating the above, the histogram has a number of peaks roughly equal to or larger than the threshold, and this peak is calculated by subtracting the number of valleys (3) from the desired output gradation (8). The division in (5) is as shown in FIG. In this case, as (e ₁₀ ~e ₄₎ and the (e ₅ ~e _6), it occurs that should moiety originally one region will be divided. In the present embodiment, the smaller one is ignored except for the one with the larger total frequency of the divided parts. Therefore, the divided input tone values are
(Φ to e ₁ ), (e _{1 to} e ₂ ), (e _{2 to} e ₃ ), (e _{3 to} e ₁₀ ),
(E _{10 to} e ₅ ), (e _{5 to} e ₆ ), (e _{6 to} e ₇ ), (e _{7 to} 255), and the representative value of the output gradation within the range of each input gradation value is output. The tone calculation unit 5 calculates the tone. In this manner, an output image of posterization that does not depend on the scene of the input image with a desired output gradation can be obtained.

In the first, second, and third embodiments, the input image and the output image are used as luminance data. However, density data may be used.

As described above, by changing the gradation number of the original image using the histogram, a desired gradation image can be obtained without being influenced by the scene of the original image. No gradation conversion image can be obtained.

[Effects of the Invention] According to the present invention, it is possible to provide an image processing method capable of satisfactorily experimenting with the effects of posterization when performing a process of lowering the gradation of input image data.

Further, it is possible to provide an image processing method in which the original color and brightness of any original image are preserved, and a posterization image without a sense of incongruity can be output.

Further, it is possible to provide an image processing method capable of outputting a posterization image in which features of an original image are stored.

Further, it is possible to provide an image processing method capable of outputting a posterization image in which a truly characteristic portion of the original image is stored.

[Brief description of the drawings]

FIG. 1 is a block diagram of an image processing apparatus according to the present embodiment which realizes the image processing method of the present invention, FIG. 2 is a block diagram of the hardware of the image processing apparatus according to the present embodiment, and FIGS. c) is a diagram showing an example of a histogram of the input image, FIGS. 4 (a) to (c) are diagrams in which the gradation is divided into eight according to frequency, and FIGS. 5 (a) to (c). FIG. 3 is a diagram showing the gradation relationship between input and output when the data used in FIG. 3 is processed, FIG. 6 is a diagram showing the gradation relationship between normal image input and output, and FIG. 8 to 11 are diagrams showing a conventional method for normalizing an output tone corresponding to the frequency of an input tone, and FIGS. 12 and 13 are diagrams showing a conventional input tone. 14 and 15 are diagrams for explaining the second embodiment, and FIGS. 16 and 17 are diagrams for illustrating a third embodiment of the present invention. Explain example It is a diagram. In the figure, 1 ... image input unit, 2 ... histogram calculation unit,
3 division unit 4 output gradation designation unit 5 output gradation calculation unit 7 image output unit 11 image input device 12
… A / D converter, 13… Image memory, 14… CPU, 15… Memory, 15a… RAM, 15b… ROM, 16… Character / numeric input device, 17… D / A converter, 18 ... Image output device.

Continuation of the front page (72) Inventor Hiroyuki Kimura 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc. (72) Inventor Toshihiro Kojima 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc. (72) Inventor Shigeki Yamada 3-30-2 Shimomaruko, Ota-ku, Tokyo Inside Canon Inc. (56) References JP-A-61-172541 (JP, A) Adaptive self-service with a limited number of representative colors Color image display Kenji Kira, Seiki Inoue, Kazuo Fukui IEICE Technical Report IE 83-92 (February 2, 1984) Fundamentals of Image Engineering 12-16 Takeshi Yasui and Masayuki Nakajima co-authored by Shokodo Co., Ltd.

Claims (3)

(57) [Claims] In an image processing method for lowering the gradation of image data, a gradation is divided for each predetermined cumulative frequency from a gradation histogram of input image data based on the number of divisions of the gradation. An image processing method, wherein a representative value is selected from within a gradation area and output as output image data. 2. An image processing method for reducing the gradation of image data, wherein a continuous gradation region having a frequency equal to or less than a predetermined value is divided as one gradation from a gradation histogram of the input image data. An image processing method comprising: outputting a representative value of a gradation area as output image data. 3. An image processing method for lowering the gradation of image data, comprising: generating a gradation histogram of input image data; and converting the input image data into image data having a designated number of gradations. A tone region whose occurrence frequency is equal to or more than a predetermined value is determined in the tone histogram, and a cumulative value in the determined tone region is compared with a predetermined number, so that a representative value of the tone region is used as output image data. An image processing method comprising determining whether or not to output.