JP2882986B2 - Image conversion system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-value input system in which a display of a natural image requiring more than 16 million colors is processed using an image display device having a limited number of display colors. It is used for displaying without deteriorating the naturalness of the image.

[0002] With the spread of multimedia technology, compression storage and transmission of natural moving images are actively performed. The present invention resides in a system having these image compression / decompression techniques, and prepares color data on a numbered table by selecting an arbitrary number of colors from a large number of colors, such as a natural image. This is used when displaying numbers by arranging them.

[0003]

2. Description of the Related Art Heretofore, many moving image compression methods have been proposed. Image data expanded using these methods requires a large amount of information to represent one color. For example, if one natural color is represented by a luminance signal and a color difference signal, an information amount of 24 bits is required. As described above, there is a limited color display in order to reproduce an image composed of colors requiring a large amount of data using a relatively inexpensive display device. This limited color display selects multiple representative colors from an image composed of many colors,
The image is reconstructed with the selected representative color. For example, a limited color is preferentially extracted from a designated emphasized area of image data given from 16.770,000 colors, and the total 256 colors are repeatedly considered while taking into account the ratio with the number of remaining colors assigned outside the emphasized area. Is an example of this limited color display (Japanese Unexamined Patent Application Publication No. 2-168685, hereinafter referred to as Conventional Technique 1).

[0004]

A processing procedure for displaying an image reconstructed by using the above-described method is to convert a signal in a color space dependent on a reconstructed decompression algorithm into a color space dependent on a display device. , And from there, a candidate color must be determined using the method described in Prior Art 1. This requires a large amount of calculation for converting the color space, and causes a heavy load when processing is performed by a processor having a low processing capacity.

In order to reduce the amount of calculation,
There is also a method in which an optimum representative color is assigned to an image before compression using a conventional technique, and redundancy is removed by using the correlation between consecutive numbers of numbers indicating the representative color to remove redundancy. However, in this method, the relationship between the arranged entries is small, and generally no effective compression efficiency is obtained.

It is an object of the present invention to provide an image conversion system capable of shortening the conversion processing time of image data having many pixels.

Another object of the present invention is to provide an image conversion system capable of converting an image with high color reproducibility.

[0008]

According to the present invention, a plurality of color information which are connected to an image input device and an image output device and have a plurality of color information respectively representing a plurality of colors limited depending on the image output device are provided. An image conversion system for converting image input data input from the image input device into image output data using the color information, and outputting the converted image output data to the image output device. A second table having a plurality of information numbers respectively indicating a plurality of color information; and referring to the second table, converting image input data from the image input device into image intermediate data composed of the information numbers. first conversion means for, by referring to the first table, a second conversion means for converting the image intermediate data to the image output data, the image input
A first table for creating the first table using data;
Table creation means, and the second table from the first table.
And a second table creating means for creating a table.
A featured image conversion system is obtained.

According to the present invention, in the image conversion system, the first conversion means compresses the image input data, expands the image input data according to a predetermined expansion algorithm, and complies with a color space dependent on the expansion algorithm. An image conversion system is provided, comprising: data processing means for generating image processing data in a format; and means for converting the image processing data to the image intermediate data with reference to the second table.

According to the present invention, when reproducing a moving image in the image conversion system, the color information in the target frame in the image input data is compared with the color information in the first table. , An image conversion system having a determination unit for determining whether the first table is appropriate.

[0011]

That is, in the present invention, the image data represented by the color space coordinates given by the image input device (this is referred to as color space A) is used, and the color space coordinates (which are As a means for displaying a limited color in the color space B), information of any one color in the first table including the color information in the color space B of the already selected candidate color is
A multidimensional second table that can be directly referred to using the color space A data is used.

In the present invention, when a series of moving image data is input, a first table having an optimum candidate color from the image data of the first frame and an arbitrary one color in the first table are provided. A second table having a number indicating the candidate color data is created by referring to the second table, and the display is performed by converting the expanded data of the first and subsequent frames into a number indicating the candidate color data. Further, in the present invention, when the input image is moving image data, the frame to be processed is shifted to the next frame, and the first table provided with the color information of the previously created candidate color is an arbitrary frame. When the color reproduction is not optimal, the first table and the second table are automatically created and updated. Furthermore, Oite this onset Ming, the first table, when the input digital image signals independent of each other, creating a multi-dimensional table based on the data already color information of the candidate colors are selected Are searched for colors corresponding to all combinations of possible values of the input signal from the first table provided with the numbers, and the numbers of the colors are stored. Two tables, a first table of color information of the selected candidate color and a second table of a number indicating an arbitrary color in the first table, are arranged before a series of encoded image data. And memorize. When the first and second tables are updated, they are placed before the compression code at the time of updating.
The determination for updating the first and second tables is made for each frame. The first table currently used is
When compared with the total number of all pixels used in the target frame for each color, the color used most in the frame is the first color.
In this case, a part or all of the first table is newly created.

[0014]

In the present invention, coordinate conversion from the color space A to the color space B and assignment to candidate colors are simultaneously performed in each pixel of the input image by a table reference method. The processing time per hit is reduced.

Further, in the present invention, since a candidate color is selected for each input image, it is possible to convert the input image into an image having high color reproducibility by using the candidate color adaptively optimized for the input image.

Further, in the present invention, since a table is created corresponding to all combinations of possible values of the input signal, the conversion table (second table) created once is used for all unknown input pixels. It does not require a table setting for each frame for input of moving images and the like, and the processing time per pixel becomes shorter as the number of pixels requiring conversion increases. Also, if the table used is not optimal, it is automatically determined and some or all of the tables are updated.
The color reproducibility can always be kept optimal so that the image quality does not deteriorate.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS.

FIG. 1 shows an overall configuration diagram of an image conversion system according to an embodiment of the present invention, and also shows a data flow. In FIG. 1, an image conversion system 1 is connected to an image input device 2 and an image output device 3,
The compression / recording-side processing unit 4 performs image conversion including image compression / decompression processing using the image input data 16 as the first conversion means, and stores the compressed code data 14 and the table 12 necessary for image conversion as a file. Image intermediate input data 1
7, the file recorded in the compressed file 5 is read out as image intermediate output data 18 in the decompression display side processing unit 6 as a second conversion means, and the image is decompressed again to perform image conversion. .

The compression / recording-side processing section 4 includes, as data processing means, an image compression means 7 for compressing image data input from the image input device 2, a limited color selection means 9, and a conversion table (YCENT) creation means. 10 and
The image processing apparatus includes a compressed file creating unit 8 that converts data from the image compressing unit 7, the limited color selecting unit 9, and the conversion table (YCENT) creating unit 10 into image intermediate output data 17. The limited color selection means 9 is used to input the image input data 15
Based on the color information table (CL)
UT) 12 and compare the color information in the target frame in the image input data 15 with the CLUT 12 to judge the suitability of the created table. YCENT
The creating means 10 includes a CLUT 12 as a second table.
YCENT 13 is created based on Image compression means 7
Creates compressed code data 14 composed of n frames from image input data 15. The compressed file creation means 8
Compressed code data 14, CLUT 12, YCENT1
3 to record the intermediate image output data 17 in the compressed file 5.

The decompression / display-side processing section 6 has an image decompression means 11. The image decompression means 11 creates decompressed image data 15 consisting of n frames from the image intermediate output data 18 of the compressed file 5 and generates the decompressed image data 1
5 is referred to as CLUT12 and YCENT13,
This becomes image output data 19.

FIG. 2A is a diagram showing an example of a main part of the image conversion system according to the embodiment of the present invention. Referring also to FIG. 2A, the image conversion system 1 includes a CPU 21, image input data 2, an image output device 3, a file input unit 22, a file output unit 23, and a memory unit 24. ing. The CPU 21 controls the image input data 16, the image output data 15, the file input unit 22, the file output unit 23, and the memory unit 24, respectively.

At the time of image compression, the CPU 21 stores the image input data 16 input from the image input device 2 in the memory 24.
And image compression means 7 for performing image compression based on the image input data 16. If necessary, the CPU 21 selects a limited color selection unit 2 that selects a limited color representing the color of the entire image according to the limited number of colors.
2 and a conversion table (YCEN
T) Forming means 23. Here, a conversion table (YCENT) 13 is a table of numbers indicating any one of the color information tables (CLUT) 12 of the candidate colors selected for the image output data 15.

The CPU 21 also stores the image code data 14, the CLUT 12, and the YC
A compressed file creating means 8 for recording and creating image intermediate data of the compressed file 5 using the ENT 13 is configured. The output of the compressed file 5 is performed by the file output unit 23 in response to an instruction from the CPU 21.

The CPU21, at the time of image expansion, expand and CLUT12 and YCENT13 from the compressed file 5 read by the file input unit 12 in the memory 24
I do .

The CPU 21 composes the image decompression means 11 for decompressing the image intermediate data 18 from the file output section and generating the decompressed image data 15 of n frames. The expanded image data 15 is stored in the CLUT 12 and the YC
The data is converted into a data format that can be output to an arbitrary display device with reference to the ENT 13, and an image is output by the image output device 2 including, for example, a color CRT.

Here, programs and work areas necessary for the operation of the CPU 21 exist on the memory 24. The image input data 16 and the image output data 3 include
The operation procedure necessary for each operation is in the image input data 16 and the image output data 3, and the CPU 11 writes the image input data on the memory 24 only by issuing an operation instruction, and reads out the image input data when outputting.

FIG. 2B is a diagram showing a modification of the image conversion system shown in FIG. 2A. The image compression means 7 and the image expansion means 11 for performing image compression and image expansion processing are provided by the CPU 2.
These means read out necessary data from the memory unit 24 in accordance with an operation instruction of the CPU 21, perform respective processes such as image compression and image decompression, and then write the data into the memory 24 again.

Next, the operation of the image compression means 7 and the limited color selection means 9 of the image conversion system 1 will be described by taking the processing of the first frame of the moving image as an example, and further, the operation of the conversion table creation means 10 will be described in detail. I do.

FIG. 3A is a flowchart showing the operation of the image conversion system according to the embodiment of the present invention during the compression processing. Referring to FIG. 3A, a compression process of image data is started (S1), and a table of color information (C
It is determined whether LUT) creation is necessary (CLUT creation determination step, S2). If it is a CLUT creation frame, move to the next stage. Next CLUT creation stage (S
Reference numeral 3) denotes a table storing values of representative colors used when a multicolor image is displayed with a limited number of colors. As a commonly used CLUT, RGB 24-bit color information is used. 256 typical to reproduce the image
It is formed by selecting colors. Many methods have been proposed for selecting 256 color information. In this CLUT creation step (S3), 256 representative colors unique to the target image are selected by an appropriate method, and the CLUT1 in which RGB values are written as shown in FIG.
2 is performed. After creation, the process moves to the next image compression processing stage. In this image compression processing step (S5), a number of compression algorithms and means have been devised so far, and this system can use any of the compression methods, but is input as an example of the input data format. The image data will be described in a Y / Cr / Cb format.

Here, the step of creating the conversion table (S4) in FIG. 3A shows the process of creating the YCENT 13. The flow of this processing is shown in the flowchart of FIG. First, conversion table creation is started (S1).
1). Next, a search tree for a binary search is created based on the CLUT 12 created in the S4 stage (S12). Using this search tree, the range in which the independent parameters of Y / Cr / Cb can remain as Y = 255, Cr = 127, Cb = 127 in the order from Y = 0, Cr = −128, Cb = −128. In all combinations, Y
/ Cr / Cb is converted to RGB, and the RGB values are examined using the search tree created in the previous stage (S12) to find the closest color value in the already created CLUT (S13). ). (Here the closest colors are RGB
Is the shortest distance when comparing the distances in a three-dimensional space with each being the axis) CL
If the closest color in the UT is found, the entry of the color is sequentially stored in a three-dimensional table as shown in FIG. 5 (S14). It is determined whether the loop is over or not (S1
5) If it is not the end, the processing from the neighboring value search (S13) is repeated. In the case of the end, the table in which all the entries are stored in the table 13 is referred to as YCE.
The conversion table creation ends as NT13 (S1).
6).

Returning to FIG. 3A, the CLUT 12, YCENT 13, and compression code data 14 created as described above are arranged in order to create a compressed file 5 (S6). Note that CLUT12, Y in the compressed file 5
The CENT 13, the compressed code data 14a, 14b, and the like can be distinguished from each other by adding a unique key or address. At this time, the compressed file 5 has the C code before the compressed code data 14 of the first frame.
When the LUT 12 and the YCENT 13 are arranged, the CLUT 12 and the YCENT 13 can be referred to immediately after the image decompression processing, which is advantageous for speeding up the processing. Next, it is determined whether or not the frame is the last frame (S7). However, since it is the first frame, the process returns to the S2 stage again.

Next, the processing of the image of the second frame will be described. First, it is determined whether to create the CLUT 12 (S2). The image of the first frame was newly created because the CLUT 12 did not exist before that. However, two frames adjacent to each other in a general moving image have a high correlation with the used pixel data. C
Even if the LUT 12 is used, there is no practical problem. However, when the change in the image is large and the previously created CLUT 12 is not always optimal, it is necessary to newly create the CLUT 12. For this reason, in the CLUT creation determination step (S2), it is determined how effective the pixel data of the frame to be compared is with the currently valid CLUT. As an example of this determination method, a histogram of the pixel data of the target frame is taken, and it is checked whether or not the same color exists in the current CLUT in the descending order of the colors included in the target frame.

Here, one threshold value (for example, 50) is provided, and five (5) colors are included in the target frame in descending order.
If the same color is present continuously in the current CLUT up to the 0th color, it is determined that “the CLUT is not updated”, and if the same color does not exist, “the CLUT is updated”. Further, for example, the case may be divided into "CLUT full update" if all the data do not exist up to the 50th, and "CLUT partial update" which replaces only the non-existent one if there is no data up to the 20th. With such a determination method, for a frame that does not require updating of the CLUT, direct image compression processing (S5) is performed, and the compressed code data 14 is arranged and recorded in the compressed file 5 as shown at 14a and 14b in FIG. . Here, if the CLUT 12 is updated, S2 and S
After a three-stage process, a new CLUT 12 and a corresponding YCENT 13 created between 14a and 14b in FIG.
Place.

Next, image decompression means 1 of the image conversion system
The operation 1 will be described in detail.

FIG. 3C is a flowchart showing the operation of the image conversion system at the time of decompression processing. FIG.
Referring to (c), the image expansion (S24) expands the image compressed by the image compression (S5) during the compression processing (FIG. 3A). For further explanation, it is assumed that the data format after image expansion is Y / Cr / Cb. First, expansion processing is started (S21). Assuming that the target compressed file 5 has the data arrangement indicated by reference numerals 12, 13, 14a, and 14b, the file is read from the beginning, and it is first checked whether there is a CLUT 12 or YCENT 13 (S22). . If these tables exist, the CLU is stored in an arbitrary area on the memory 24.
Tables such as T12 and YCENT13 are set (S23). If the data is compression-coded data 26a, 26b, the image is expanded (S24) and displayed (S2).
5). Here, it is determined whether or not the frame is the last frame (S2).
6) If it is not the last frame, the process returns to step S22 again, and when the last frame is reached, the decompression process ends (S27).

Since the decompressed data is data in the Y / Cr / Cb format, the corresponding entry is read out from this value for each pixel by directly referring to the table 25 in FIG. 5 to obtain the value of that pixel. . The image output device 3 refers to the CLUT 12 as shown in FIG.
Output as B24 bit image data.

The above is an example of the operation of the image conversion system according to the embodiment of the present invention. The system performs only the conversion of the image signal without including the compression / expansion processing, or the image signal provided is Y / Cr / Cb. It can be applied to other systems.

[0038]

As described above, according to the present invention, in each pixel of an input image, coordinate conversion from a color space dependent on an image input device to a color space dependent on an image output device and allocation to candidate colors are simultaneously performed. Since the conversion is performed by the table reference method, the processing time per pixel becomes shorter as the number of pixels requiring conversion increases.

Further, according to the present invention, since a candidate color is selected for each input image, it is possible to convert the input image into an image having high color reproducibility by using the candidate color adaptively optimized for the input image. .

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating an overall configuration of an image conversion system according to an embodiment of the present invention.

FIG. 2A is a configuration diagram illustrating a main part of an image conversion system according to an embodiment of the present invention. (B) is a figure which shows the modification of (a).

FIG. 3A is a processing flowchart including a compression operation of the image conversion system. (B) is a processing flowchart including a decompression operation of the image conversion system. FIG. 3C is a processing flowchart including a conversion table creation operation of the image conversion system.

FIG. 4 is a conceptual diagram of a CLUT.

FIG. 5 is a conceptual diagram of YCENT.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Image conversion system 2 Image input device 3 Image output device 4 Compression / recording side processing unit 5 Compressed file 6 Decompression / display side processing unit 7 Image compression means 8 Compressed file creation means 9 Limited color selection means 10 YCENT creation means 11 Image extension Means 12 CLUT 13 YCENT 14, 14a, 14b, 14c Compression code data 15 Decompressed image data 16 Image input data 17, 18 Image intermediate data 19 Image output data

────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ identification code FIG06F 15/68 310A (56) References JP-A-3-27674 (JP, A) JP-A-4-190466 (JP, A) JP-A-5-134640 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06T 1/00 G06T 5/00 G06T 9/00 G09G 5/06 H04N 9/64

Claims (3)

(57) [Claims] 1. An image input device, comprising: a first table connected to an image input device and an image output device, the first table having a plurality of color information respectively representing a plurality of colors limited depending on the image output device; An image conversion system that converts image input data input from the image input data into image output data using the color information and outputs the image output data to the image output device, wherein a plurality of pieces of information respectively indicating a plurality of pieces of color information in the first table A second table having a number, first conversion means for referring to the second table, and converting image input data from the image input device into image intermediate data composed of the information number; Referring to table 1
A second conversion unit for converting the image intermediate data into the image output data; and a first conversion unit using the image input data.
First table creating means for creating a table of
A second table for creating the second table from the first table
And an image conversion system comprising:
Tem. 2. The image conversion system according to claim 1, wherein the first conversion unit compresses the image input data, expands the image input data according to a predetermined expansion algorithm, and follows a color space dependent on the expansion algorithm. An image conversion system, comprising: data processing means for generating image processing data in a format; and means for converting the image processing data into the image intermediate data with reference to the second table. 3. The image conversion system according to claim 1, wherein when reproducing a moving image, color information in a target frame in the image input data is compared with color information in the first table. An image conversion system comprising: a determination unit configured to determine whether the first table is appropriate.