JPH01288087A - Limited color picture processor - Google Patents

Abstract

PURPOSE:To interpolate even a picture not entirely depending on a brightness to be displayed with excellent picture quality by counting the frequency of occurrence of each representative color of a real surrounding picture element corresponding to an interpolated picture element, extracting the representative color with the maximum frequency of occurrence and using it as the pixel for the interpolated pixel. CONSTITUTION:A pixel interpolation means 17 is provided with a picture segmentation means 14 reading each pixel of a real surrounding pixel comprising plural real picture elements located around the interpolated pixel to decide a representative color code of the pixel of the interpolated pixel, a frequency count means 15 counting the frequency of occurrence of each representative color of the real surrounding pixel and a maximum frequency of occurrence pixel extracting means 16 extracting a representative color having the maximum frequency of occurrence from the representative color of the real surrounding pixel and using its code as the pixel of the interpolated pixel. Thus, the pixel interpolation means 17 outputs selectively the representative color suitable for the pixel of the interpolated picture element based on the statistic property. Thus, the interpolation with excellent picture quality is attained even for a picture whose pixel is entirely independent of the brightness displayed actually such as a limited color pixel.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a limited color image processing device, particularly a limited color image that is composed of a plurality of pixels and has code information indicating a representative color instead of an RGB value as a pixel value. This invention relates to an improvement in pixel interpolation processing in a color image processing device.

[Background Art] In recent years, various image processing devices have been put into practical use as remote monitoring and even videophones. In such an image processing apparatus, it is required to reduce the amount of data to be transmitted as much as possible, and to be able to determine the content of the transmitted image at an early stage of transmission.

To this end, various high-efficiency encoding methods and hierarchical transmission methods have been put into practical use as methods for compressing image information. On the other hand, there is a method of performing pixel interpolation using the transmitted pixel values.

The pixel interpolation method described above reduces the number of pixels by half.
If the data is thinned out and transmitted, it has the effect of halving the amount of data.

Furthermore, if pixels are thinned out according to a sampling pattern that takes into account the human visual characteristics and reduces only the resolution in diagonal directions, which are not visually sensitive, it is possible to eliminate visual discomfort. It is possible to reproduce images without causing the user to notice any deterioration in image quality.

Hereinafter, conventional image interpolation processing will be explained based on the drawings.

Figure 5 shows the schematic configuration of the image interpolation device disclosed in "Image Information Processing" (Corona Publishing, pp. 131-133, edited by Hajime Enomoto), and Figure 6 shows the pixel sampling pattern. It is shown.

As shown in FIG. 5, the device includes an image input means (11) for manually inputting pixel values sampled according to the sampling pattern shown in FIG. an image storage means (hereinafter referred to as image memory) (12) for storing images, a line memory (31) for storing the image of three horizontal video lines, an interpolation filter (32) corresponding to the sampling pattern; a convolution calculation unit (33) that calculates the pixel value of the inserted pixel; a display (18) that displays interpolated image information;
It consists of

Next, the operation of this device will be explained.

First, sampled pixel values are input from the image input means (11). At this time, mark O (
bs Cs ds e, etc.) are actual pixels whose pixel values have been input, and the X marks (a, etc.) are interpolated pixels that have been thinned out.

Then, the input pixel values are written into the image memory (12). At this time, 0 is written in the image memory (12) at the position of the interpolated pixel.

Then, in order to perform interpolation processing, that is, to calculate the pixel value of the interpolated pixel, the contents of the image memory (12) are read into the line memory (31) every three lines of the image memory (12).

Next, based on the pixel value of the actual pixel read out in the line memory (31), a convolution operation between the pixel value and the interpolation filter (32) is executed by the convolution operation device (33).

The result is then written to the image memory (12).

That is, the pixel value of the interpolation pixel a is determined by the convolution calculation device (33) using the interpolation filter (32) as shown in equation (1).

a= (b+c+d+e)/4 (1)
Furthermore, the values of the actual pixels (b%C%d, e) are not changed.

The result of the convolution operation as described above being performed on the entire image, that is, the interpolated image, is an image that approximates the image before being subsampled, and is displayed on the display (18).

A conventional image interpolation device is configured as described above, and thinned out pixels are interpolated from actual pixels.

Further, in the case of a color image, a color image interpolation device can be configured by performing the above interpolation processing on three RGB images.

[Problems to be Solved by the Invention] As described above, the conventional image interpolation device utilizes the correlation of pixel values.

That is, using the fact that the pixel value represents the brightness level and that the value does not change significantly locally, the interpolated pixel (a) is converted to the surrounding real pixels (b%c, d, e etc)
By the convolution operation, i.e., a-conv (b, c, d%e-old -> (
2) by performing linear interpolation (for example, interpolation as in (1) above).

However, with conventional pixel interpolation devices, it is not possible to apply linear interpolation as described above to pixel interpolation of pixel-thinned images for limited color images that have code information indicating representative colors rather than brightness as pixel values. There was a problem.

The present invention has been made to solve the above problems, and provides a limited color image processing device that can perform pixel interpolation of a limited color image whose pixels have been thinned out according to a desired sampling pattern. The purpose is to

[Means for Solving the Problems] The limited color image processing device according to the first claim is a pixel interpolation means for interpolating thinned out pixels between actual pixels of a limited color sampling image, Count the frequency of occurrence of each representative color of real peripheral pixels consisting of multiple real pixels located in the surrounding area,
A pixel interpolation means is provided which uses the pixel value of the interpolation pixel as a representative color code having the maximum occurrence degree of the actual surrounding pixels, and performs pixel interpolation of the sampled limited color image.

The limited color image processing device according to the second claim is a pixel interpolation means for interpolating pixels thinned out between real pixels of a limited color sampling image, the pixel interpolation means interpolating pixels thinned out between actual pixels of a limited color sampling image, and comprising interpolating pixels and pixels located around the interpolating pixels. Each representative color is weighted according to the distance from each actual pixel,
A pixel interpolation unit that calculates an interpolation color that is a pixel value of an interpolation pixel, and sets a representative color that is most similar to the interpolation color as the pixel value of the interpolation pixel, and performs pixel interpolation of a sampled limited color image. It is.

[Operation] According to the invention of claim 1, the statistical properties of the pixel values of real pixels around the interpolated pixel are measured, and the interpolated pixel is determined from the surrounding pixel values (representative color) based on the statistical properties. By selectively outputting the pixel values of , pixel interpolation of the limited color image is performed.

Next, according to the invention of claim 2, each representative color is weighted according to the distance between the interpolation pixel and each real pixel located around the interpolation pixel, and the interpolation that is the pixel value of the interpolation pixel is Pixel interpolation of a limited color image is performed by calculating a color and outputting a representative color that is most similar to the interpolated color as a pixel value of an interpolated pixel.

[Embodiments] Hereinafter, preferred embodiments of the limited image processing device according to the present invention will be described based on the drawings.

FIG. 1 shows a preferred embodiment of a limited color image processing apparatus according to the first aspect of the invention.

In the figure, the same parts as those in the conventional example shown in FIG.

As shown in FIG. 1, the limited color image processing device according to the present embodiment processes a limited color sampling image sampled according to a predetermined sampling pattern (for example, a sampling pattern as shown in FIG. 6). Image input means (1) for inputting information into the device
1) and an image memory (12) that stores input image information.
), a representative color memory (13) that stores the representative color code applied to the limited color sampling image and the RGB signal of the representative color corresponding to the representative color code, and a space between the actual pixels of the limited color sampling image. A pixel interpolation means (17) for interpolating the subtracted pixels, and a display (18) for displaying the limited color image interpolated by the pixel interpolation means (17).
It's empty.

The feature of the present invention is that the pixel interpolation means (17)
In order to determine the pixel value (representative color code) of the interpolation pixel, the pixel interpolation means (17) calculates each pixel of real peripheral pixels consisting of a plurality of real pixels located around the interpolation pixel. Image cutting means (14) for reading values (representative color code)
a frequency counting means (15) for counting the frequency of occurrence of each representative color of the actual surrounding pixels; and a frequency counting means (15) for counting the frequency of occurrence of each representative color of the actual surrounding pixels; and maximum frequency pixel extraction means (16).

Therefore, the pixel interpolation means (17) selects and outputs a representative color suitable for the pixel value of the interpolated pixel based on statistical properties from among the representative colors of the actual surrounding pixels.

Next, the operation of the limited color image processing device will be explained.

First, the image input means (11) inputs a limited sampling image and stores the input image information in the image memory (11).
12).

The image memory (12) stores the input pixel value at the position of the actual pixel whose pixel value has been entered manually, and stores a code indicating an interpolated pixel at the position of the thinned out pixel, for example, "0".
Write and memorize. At this time, the image memory (12
), pixels are interpolated between actual pixels.

Next, the image cutting means (14) cuts out an image made up of a plurality of real pixels around the interpolated pixel.

That is, in the case of the sampling pattern shown in FIG. 5, the actual pixels S, C5dse, and C5dse are extracted.

Then, the frequency counting means (15) counts the degree of occurrence of each representative color in the real peripheral pixels from the pixel value of the cut out real pixel, and outputs it to the maximum frequency pixel extracting means (16). .

Next, maximum frequency pixel extraction means (16) extracts a pixel value (representative color) with the highest frequency of occurrence among the actual surrounding pixels based on the output of the frequency counting means, and converts the extracted pixel value into It is output to the image memory (12) as a pixel value of the interpolated pixel.

The pixel interpolation means (17) then uses the image memory (12)
The above-described one-pixel interpolation process is performed for all interpolated pixels.

On the other hand, the image human input means (11) stores a representative color code constituting the manually generated limited color sampling image and a representative color indicated by an RGB signal corresponding to the representative color code into a representative color storage means (13). Store in.

After the pixel interpolation is completed, the display (18) sequentially reads each pixel value stored in the image memory (12) as an RGB signal via the representative color storage means (13), and uses the read RGB signal. Image display is performed based on.

In this embodiment, an example is shown in which the image cutting means (14) is used as a means for reading out the pixel values of the real peripheral pixels of the interpolation pixel, but the line memory (14) shown in the conventional example is used.
31) can also produce similar effects.

Further, the means for reading out the pixel values of the real peripheral pixels of the interpolated pixel is as follows:
Actually, no special hardware is required, and any means for reading out the pixel values of real pixels located around the thinned out pixels can produce the same effects as the present embodiment.

Next, another embodiment according to the invention of claim 1 will be described.

FIG. 2 shows a system configuration of a limited color image processing apparatus according to another embodiment.

As shown in FIG. 2, the apparatus applies the communication port (21) as the image input means shown in the embodiment of FIG. 1, and stores the representative color as the representative color storage means (13). A color map (22) for converting and outputting representative color codes read from an image memory (12) into RGB signals is applied.

Furthermore, the device uses a CPU as the interpolation means (17).
(23) is applied, and the CPU (23) performs the above-mentioned search interpolation process according to the control program stored in the memory (24), that is, the frequency of counting the representative colors of real pixels around the interpolated pixel. It performs counting, extracts the maximum opening degree pixel value, and writes the pixel value to the image memory (12), and also controls various devices.

Therefore, even if the apparatus is constructed as shown in FIG. 2, the same effects as in the embodiment shown in FIG. 1 can be obtained.

That is, according to the embodiments of FIGS. 1 and 2, the pixel value of the interpolated pixel is interpolated based on the frequency distribution of the actual surrounding pixels, so the interpolated pixel value is determined based on the local correlation of the image. This is a very appropriate value.

Next, the invention shown in the second claim will be explained in detail.

FIG. 3 shows an embodiment of a limited color image processing apparatus according to the second aspect of the invention.

As shown in the figure, the features of this embodiment include an image memory (12), an image cutting means (14), an interpolation color calculation means (15a), and an interpolation representative color extraction means (16a).
and a pixel interpolation means (17), which weights each representative color according to the distance between the interpolation pixel and each real pixel located around the interpolation pixel, and calculates the interpolation value that is the pixel value of the interpolation pixel. The method is to calculate the color and set the representative color that is most similar to the interpolated color as the pixel value of the interpolated pixel.

The image cutting means (14) is means for reading out the pixel values of each real pixel located around the interpolated pixel.

The interpolation color calculation means (15a) converts each representative color of the real peripheral pixels read out by the image cutting means into an RGB signal, and then converts each representative color according to the distance between the interpolation pixel and each real pixel. , weighting, and calculating an interpolation color which is a pixel value of an interpolation pixel.

Then, the interpolation color representative color extraction means (16a) extracts the representative color that is most similar to the calculated interpolation color.

Further, the pixel interpolation means (17) is means for writing the output result of the interpolation color extraction means (16a) into the image memory (12) as a pixel value of an interpolation pixel.

Next, the operation of this embodiment will be explained.

First, the image manual means (11) inputs a limited sampling image and stores the input image information in the image memory (11).
12).

The image memory (12) stores the input pixel value at the position of the actual pixel whose pixel value has been entered manually, and stores a code indicating an interpolated pixel at the position of the thinned out pixel, for example, "0".
Write and memorize. At this time, the image memory (12
), pixels are interpolated between actual pixels.

Next, the image cutting means (14) cuts out an image made up of a plurality of real pixels around the interpolated pixel.

That is, in the case of the sampling pattern shown in FIG. 4, real pixels bSc, d, and e are extracted.

Then, the interpolation color calculating means (15a) calculates the pixel value of the interpolation pixel from the representative color indicated by the pixel value of the actual pixel located around the pixel read out by the image cutting means (14) as follows. Calculate as follows.

First, in the case of a limited color sampling image whose pixels have been thinned out according to the sampling pattern shown in FIG.
C, , , C8 binding, when the interpolation color C of the interpolation pixel a is, the interpolation color C is -a
adan 8-(see, +9°+9.+9°)/4
It is given by (3). However, Danyu, C,, C0, Mi6, C are three-dimensional vectors representing colors in the three-dimensional color space, -
In other words, it is a color value expressed using RGB.

Further, the sampling pattern is not limited to that shown in FIG. 6, and a pattern such as that shown in FIG. 4, for example, can also be used.

When the interpolation color of the interpolation pixel a is determined using the representative color indicated by the values of the actual pixel c, dSe, the following procedure is performed.

That is, the interpolation color calculation means (15a) determines the weight W6 etc. for each pixel used for interpolation, and calculates C'' (wbCb +
VV oC, +W dQ d+wC)/4 -e (wb'+wo+wd 10w.) Calculate the interpolated color using (4).

At this time, the weight for each pixel is generally the reciprocal of the distance between the interpolated pixel and the actual pixel used for interpolation on the pixel plane Wb-1/db w-1/do (5) wd-1
/dd w -1l11/d.

Use.

That is, in the - film, the representative color indicated by the pixel value used for interpolation is Cs (i-0, 1, 2...), and the weight for each pixel is W [(t-0, 1, 2...). ), then the interpolated color is calculated by C-ΣwC/Σw 1 (6).

Next, since the interpolation color 9 calculated by the interpolation color calculation means (15a) does not necessarily match the representative color, the interpolation color 9 is calculated by the interpolation representative color extraction means (16a). The distance between C and each representative color (distance #t on the three-dimensional vector),
That is, the difference is calculated and the closest (most similar) representative color is extracted.

Then, the pixel interpolation means (17) writes the code information of the representative color extracted by the interpolation representative color extraction means (16a) into the position of the interpolation pixel in the image memory (12).

When the above pixel interpolation process is executed for all interpolated pixels, an interpolated image is completed, and the result is stored in the representative color storage means (1
3), the image is displayed as a color image on the display (18). At this time, the representative color storage means (13) stores the number of the representative color stored in the image memory (12) in RGB.
It plays the role of converting into a value.

According to this embodiment, each representative color is weighted according to the distance between the interpolation pixel and each real pixel located around the interpolation pixel, and the interpolation color that is the pixel value of the interpolation pixel is calculated. Since the representative color most similar to the interpolation color is used as the pixel value of the interpolation pixel, any sampling pattern may be used, and the position of the pixel used for interpolation is not particularly limited.

In this embodiment, an example is shown in which the image cutting means (14) is used as a means for reading out the pixel values of the real peripheral pixels of the interpolation pixel, but the line memory (14) shown in the conventional example is used.
31) can also produce similar effects.

Furthermore, even if the apparatus of the second claim of the present invention is constructed as shown in the embodiment of FIG. 2, and the CPU (23) is configured to perform the pixel interpolation as described above, It has the same effect as.

That is, the CPU (23) calculates an interpolated color from real pixels around the interpolated pixel, extracts a representative color that is most similar to the interpolated color, and extracts the representative color from the actual pixels surrounding the interpolated pixel, according to the control program stored in the memory (24). Image memory of code information (1
2) may be configured to execute writing to and control various devices.

Therefore, according to the embodiment according to the second claim of the invention,
Each representative color is weighted according to the distance between the interpolation pixel and each actual pixel, an interpolation color that is the pixel value of the interpolation pixel is calculated, and the representative color that is most similar to the interpolation color is determined as the pixel value of the interpolation pixel. Therefore, the interpolated value becomes a very appropriate value in view of the local correlation of the image.

[Effects of the Invention As explained above, according to the invention according to the first claim,
In the subsampled limited color image, the occurrence frequency of each representative color of the real surrounding pixels corresponding to the interpolated pixel is counted, and the representative color with the maximum frequency is extracted and used as the pixel value of the interpolated pixel. This has the effect of enabling interpolation with good image quality even for images such as color images where pixel values are completely unrelated to actually displayed luminance values.

Further, according to the invention of claim 2, as explained above, in the subsampled limited color image,
By providing an interpolation color calculation means, an interpolation representative color extraction means, and a pixel interpolation means as a means for executing pixel interpolation processing, a representative color that is most similar to the interpolation color calculated from the actual surrounding pixels corresponding to the interpolation pixel can be obtained. Since the interpolated pixels are interpolated with , the effect is that it is possible to interpolate with good image quality even in images where pixel values have no relation to the actually displayed brightness values, such as limited color images. be.

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram showing one embodiment of the limited image processing device according to the invention of the first claim, and FIG. 2 is a system configuration diagram of another embodiment of the invention of the first and second claims. , 3rd
The figure is a block configuration diagram of an embodiment according to the invention of claim 2,
FIG. 4 is an explanatory diagram of the sampling pattern, FIG. 5 is a block diagram of a conventional example, and FIG. 6 is an explanatory diagram of the sampling pattern. In the figure, (12) is image memory, (13) is representative color storage means, (14) is image cutting means, (15) is application text counting means, (16) is maximum frequency pixel extraction means, (17) is
) is a pixel interpolation means, (15a) is an interpolation color calculation means, (1
6a) is an interpolation representative color calculation means. In addition, in the figures, the same reference numerals indicate the same or equivalent parts. Agent: Patent attorney Masuo Oiwa (and 2 others) Figure 1 Figure 3 XXXX: XXXXX xxxx4xxxx oxxx6xxx. xxxxxxxxxx Figure 4 ○ xoxox ○ e xoxoxox oxoxox. Figure 6 Procedural amendment (voluntary)

Claims (2)

[Claims] (1) In a limited color image processing device that performs image processing of a limited color image consisting of a plurality of pixels and having code information indicating a representative color as a pixel value, limited color sampling is performed in which pixels are sampled according to a predetermined sampling pattern. image input means for inputting pixel values of an image into the apparatus; image storage means for storing the input pixel values of the limited color sampling image; A pixel interpolation means for interpolation, which counts the frequency of occurrence of each representative color of real peripheral pixels consisting of a plurality of real pixels located around the interpolated pixel, and sets the pixel value of the interpolated pixel to the maximum of the real peripheral pixels. What is claimed is: 1. A limited color image processing device, comprising: pixel interpolation means that uses a representative color code having a frequency of occurrence, and performs pixel interpolation of a sampled limited color image. (2) In a limited color image processing device, an image input means for inputting into the device a pixel value of a limited color sampling image whose pixels have been sampled according to a predetermined sampling pattern; and a pixel value of the input limited color sampling image. and pixel interpolation means for interpolating the thinned out pixels between the real pixels of the limited color sampling image, the interpolation pixel and each real pixel located around the interpolation pixel. pixel interpolation means that calculates an interpolation color that is a pixel value of an interpolation pixel by weighting each representative color according to the distance from the interpolation pixel, and sets the representative color that is most similar to the interpolation color as the pixel value of the interpolation pixel; What is claimed is: 1. A limited color image processing device that performs pixel interpolation of a sampled limited color image.