JPH0591339A - Data compressing device and data expanding device - Google Patents

Abstract

PURPOSE:To provide a picture of high quality by using a maximum value and a minimum value of gradations as a first reference value and a second reference value respectively and expanding and reproducing a picture having a middle gradation from the mixture proportion value of first and second reference values. CONSTITUTION:A background picture signal and a superimpose picture signal including characters and symbols are compressed or encoded independently of each other. Picture data of characters, shadow, or the like is expressed with plural gradations by a data converting means 112 and is handled as a multilevel picture. A maximum value and a minimum value of gradations are encoded to Huffman codes in a Huffman encoding part 116 by run length, and a middle value is encoded by assignment of a code or the like provided in the Huffman code system. In this case, the middle value of picture data on boundaries of characters, shadow, or the like is expressed as the mixture proportion (mixing value) of the maximum value of gradations (picture data of characters or the like as the first reference value) and the minimum value of gradations (background picture data as the second reference value).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression apparatus and its expansion apparatus suitable for image compression when, for example, characters or symbols are superimposed on a color moving image, and more particularly to characters or symbols. The halftone data generated at the boundary between the background and the background is also compressed and expanded to provide a high quality image.

[0002]

2. Description of the Related Art Conventionally, when data compression processing is performed on an image signal of an image created by mixing a background image with a superimposed image signal including characters and symbols, as shown in FIG. Before the compression process, characters and symbols are superimposed on the background image in advance and mixed, and then the compression process is performed. In the figure, the background image signal is input to the video mixing device 90 together with the superimpose image signal, and mixing of the two is performed here. As a result, characters and symbols are superimposed on the background image. Then, the image signal after superimposing is supplied to the image compression device 92, where data compression processing is performed and a compressed image signal is output.

However, in such a conventional technique,
Since the superimposed image signal including characters and symbols is compressed by the image compression device 92 like the background image signal, there is a disadvantage that deterioration occurs. For example, when the data compression rate in the image compression device 92 is high, the image quality correspondingly deteriorates from the high frequency component. Superimposed image signals often include high-frequency components in the edge portions of characters and the like, so that the deterioration is more noticeable than in the background image.

Further, the superimposed image data is
If only the background image and the portion such as characters are distinguished, it can be represented by binary data. However, in such binarized data, a line (0 ° <α <45 °) slightly inclined from the horizontal shown in FIG. 7 (A), a circle PA shown in FIG. 7 (B),
In each part of PB, PC, PD, staircase noise is generated as shown in the figure. Superimposed image data including characters and symbols includes many such conditions. For this reason, the binarization causes the staircase noise in the edge portion to be considerably conspicuous. Therefore, in order to obtain a high quality image, it is necessary to handle the superimposed image data including characters and symbols as multi-valued data.

In practice, if the superimpose image data is treated in advance as multivalued data, the above-mentioned staircase noise becomes visually inconspicuous. If 8 to 10-value data is used, 25 of the luminance signal of the video signal of the background image is used.
It is almost impossible to distinguish from 6 values (8 bits).

Therefore, the method of G3 is independently used for each bit plane of the superimposed image data,
A method of performing compression by this is conceivable. However, with this method, the amount of information is increased by a factor of 8 or more compared to compression by simple binarization. On the other hand, there is a limitation on the recording of image data on the medium, and there is also a demand that superimposed image data must be recorded within the range of such restrictions.

As a response to these requests, there is one already proposed by the applicant as Japanese Patent Application No. 3-106720. This is a data compression device and decompression device as shown in FIGS. 8 (A) and 8 (B). In FIG. 3A, the background image signal is input to the image compression apparatus 10, and the superimpose image signal is input to the superimpose image compression apparatus 12. The output sides of the image compression device 10 and the superimpose image compression device 12 are connected to the input side of a multiplexer 14, respectively, and the image data after the compression processing is output from the multiplexer 14 as a multiple bit stream. ing. In FIG. 3B, the compressed multiple bit stream obtained by the data compression device is input to the demultiplexer 20. The output side of the background image signal of the demultiplexer 20 is connected to the input side of the image decompression device 22, and the output side of the superimpose image signal is connected to the input side of the superimpose image decompression device 24. There is. The output sides of the image decompressing device 22 and the superimpose image decompressing device 24 are connected to the input side of the video mixing device 26 so that the video mixing device 26 outputs the image signal for display. Is becoming

[0008]

However, when processing image data that is shaded around a character or symbol, there are three types of data areas: background, shade, character and symbol. Since such an image has double edges, it often contains a higher frequency component, which causes a problem that the background image is extremely deteriorated. Furthermore, if the colors of the background images are exchanged on the decoder side, there is a problem that the intermediate gradation portion becomes unnatural.

[0009]

In order to solve the above-mentioned problems, the present invention relates to a data compression device for compressing an image having an intermediate gradation, and a data conversion means for expressing the image in a plurality of gradations, and this data conversion. With respect to the image data obtained by the means, the data of the maximum value (background image signal) and the minimum value (superimpose image signal) of the gradation are directly compression-coded and the intermediate value of the gradation is Data is,
There is provided a data compression device including an encoding unit that performs compression encoding as a mixing ratio value of the maximum value and the minimum value of the gradation. Further, in a data decompression device for decompressing image data having an intermediate gradation compressed by the data compression device, a means for expanding / decompressing the compressed image data and a maximum value and a minimum value of the gradation are 1 (background image signal) and second reference value (superimposed image signal)
And a means for decompressing and reproducing an image having an intermediate gradation from the mixture ratio value of the first background image signal and the second reference value.

[0010]

According to the present invention, the background image signal and the superimposed image signal containing characters and symbols are separately compressed or coded. Image data such as characters and shadows are expressed in multiple gradations by a data conversion means and treated as a multi-valued image. Then, the maximum value and the minimum value of the gradation are subjected to the Huffman coding by the run length, and the intermediate value is coded by assigning the code provided in the Huffman code system. Therefore, even if compared with data compression based on the G3 standard, 1.5 to
The code amount is approximately doubled, and despite the high compression rate, good quality is maintained by the multilevel expression.

Further, an intermediate value of image data generated at a boundary such as a character and a shadow is a maximum value of gradation (that is, image data of a character which is a first reference value) and a minimum value of gradation (that is, It is expressed as a mixing ratio (mixing value) with the background image data which is the second reference value. For this reason, when changing the color of the background image, it is sufficient to change the second reference value (minimum value), and it is easy to change the color.

Further, even when three types of data areas of shadow, characters and symbols are required, only the mixing ratio (mixing value) needs to be transmitted. Therefore, it is possible to simultaneously encode characters, symbols, and shadows without complicating the encoding algorithm.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a data compression apparatus and its decompression apparatus according to the present invention will be described below with reference to the accompanying drawings. <Embodiment 1> (Data compression apparatus) This embodiment provides one of effective compression methods for image data such as superimposed characters.

FIG. 2 shows the structure of the data compression apparatus. In the figure, superimposing image data of characters and symbols is input to the pre-processing unit 110. The output side of the pre-processing unit 110 is connected to the input side of the run length counter 112, and the output side of the run length counter 112 is connected to the input side of the run length encoding unit 114. The output side of the run length encoding unit 114 is connected to the input side of the Huffman encoding unit 116, and the Huffman encoding unit 116 outputs the bit stream of the compressed image data. It has become.

Of the above units, the pre-processing unit 110 is
For example, if the input superimpose image data is 3
This is a pre-process for quantizing into 8 bits and 8 gradations. For example, in an image scanner or the like, an image is scanned and read in a plurality of gradations, but the same operation as that is performed in the pre-processing unit 110. When the input superimpose image data is completely binarized, appropriate filtering is performed to pseudo-increase the gradation to 6 to 8, and quantization is performed. Use gradation.

FIG. 1A shows 6 intermediate gradations included in 8 gradations from a black level to a white level.
Those levels are ESC1, ESC2, ESC3, ...
..., each represented by ESC6. The ESC level is the data of the intermediate value of the gradation, and the maximum value (white as the superimposed image) and the minimum value (black as the background image) of the gradation.
And a mixing ratio value (mixing value). By providing the ESC with a mixing value corresponding to the halftone of black and white, in other words, the mixing ratio of white, which is a superimpose image, and black, which is a background image, the brightness and A high-quality image can be obtained even if colors or the like are freely changed when expanding.

Next, returning to FIG. 2, the run length counter 112 determines the run length (run length) in the input data, that is, the continuous degree of white and black, which are the minimum and maximum values of the above-described 8 gradations. It is for counting. The run length coding unit 114 also performs coding corresponding to the count value of the input run length. As a method of compressing image data, there is a method which is put into practical use in a facsimile or the like. According to this, the binary image data is compressed by Huffman coding the run length (run length) of white data and the run length of black data. As is well known, such a technique has been clarified as "CCITT Recommendation T.4, standardization of a grape facsimile apparatus for document transmission", and is generally called G3 standard. Next, the Huffman encoding unit 116
Is for Huffman coding of white and black of the input data, and assigning a predetermined code to the other data of the mixing value represented by ESC.

Next, the operation of the data compression apparatus configured as described above will be described. The superimpose image data is input to the pre-processing unit 110, and the pre-processing described above is performed here. As a result, the superposed image data is represented by 3 bits and 8 gradations and is output from the pre-processing unit 110.

Here, it is assumed that the form of the character represented by the superimposed image data is the alphabet "A" as shown in FIG. 4, and this is indicated by the arrow FA in the figure.
It is assumed that the data scanned in the direction of is input to the pre-processing unit 110. The background is black and the characters are white. Scanning in the direction of arrow FA leads to a background portion → boundary portion → character portion → boundary portion → background portion.
Looking at this on a pixel-by-pixel basis, the image data is black → black → black → black → ESC1 → ESC3 → ESC6 → white → white → white → white → ESC6 → ESC4 → ESC2 → black → black → black → black → Will be lined up. The transition of the intermediate level displayed by ESC at the boundary portion differs depending on the image. This image data consists of "black" part, "E
It will consist of three parts, the "SC" part and the "white" part.

(First Compression Encoding Method) With respect to the output of the pre-processing unit 110, for the black part and the white part, the run length counter 112, the run length encoding part 114, the Huffman encoding part By 116, run length coding and Huffman coding similar to the G3 standard described above are performed. Since the run lengths of the black portion and the white portion are both “4”, the image data string can be displayed as black 4 → ESC1 → ESC3 → ESC6 → white 4 → ESC6 → ESC4 → ESC2 → black 4. For the codes specifically assigned to black 4 and white 4, refer to the code tables shown in Tables 1 and 2, for example.

[0021]

[Table 1]

[0022]

[Table 2]

The codes in these tables are Huffman codes created in advance as statistically efficient ones for various images, and short codes are assigned in order of the run length with the highest occurrence probability. For example, the black 4 is coded as "1011" with reference to the background run length "4" in Table 1. Also, white 4
Is encoded as "011" by referring to "4" in the character part run length from Table 1. The same applies to image data of other run lengths.

Next, for the halftone ESC1 to ESC6, codes are defined as shown in Table 3 as a part of the Huffman code system shown in Tables 1 and 2, and encoding is performed using this code. Is done.

[0025]

[Table 3]

For example, ESC1 is "0000000.
It is encoded to "1000". The same applies to the other cases.
Such encoding is performed by the Huffman encoding unit 116,
A bit stream of code data will be output.

In the above example, the code data is output in the order of 1011 → 000000001000 → 00000001101 → 00000010100 → 011 → 00000010100 → 000000010010 → 00000001100 → 1011. The decompression of the compressed data may be performed by reversing the compression process.

As described above, according to this embodiment, superimposed characters and the like are expressed in binary gradation or more,
Since the boundary between the background and the characters is represented by a special code representing a halftone, the staircase noise as shown in FIG. 7 disappears, and good quality can be obtained.
By expressing a character or the like that is superimposed with a gradation value or more, there will be almost no visual difference from the background image having a gradation of 256 values, and even if data compression is performed, it will be very Good image quality can be obtained.

(Second Compression Encoding Method) Next, an example of a method in which compression methods for mixing values are different will be described.
As shown in FIG. 1B, halftones of image data such as superposed characters of 3 bits and 8 gradations, which have been preprocessed in the preprocessing unit 110, are represented by L1 and L as shown in FIG.
2, ..., L6.

As in the above-described embodiment, the arrow FA in FIG.
The case where the character “A” is scanned in the direction of will be described as an example. Similarly, the image data will be arranged as follows: black → black → black → black → L1 → L3 → L6 → white → white → white → white → L6 → L4 → L2 → black → black → black → black →. .. This image data is composed of three parts: a "black" part, a halftone "L" part, and a "white" part.

With respect to the output of the pre-processing unit 110, the run length counter 112 and the run length encoding unit 114 perform the run length encoding similar to the above-mentioned G3 standard for the black portion and the white portion. Done. The run length of the black part and the white part are both "4"
Therefore, the image data string is as follows: black 4 → L1 → L3 → L6 → white 4 → L6 → L4 → L2 → black 4.

In this case, in the present embodiment, only one Huffman code is provided to identify L indicating the halftone L. Specifically, a code “11” corresponding to L shown in Table 6 is provided as a part of the Huffman code system shown in Tables 4 and 5.

[0033]

[Table 4]

[0034]

[Table 5]

[0035]

[Table 6]

For the number of consecutive L codes, for example, the Huffman code of the character portion run length shown in Tables 4 and 5 is used. Then, after that, a specific 3-bit fixed-length code for each halftone level is assigned. Specifically, as shown in FIG. 8, for each halftone level L1, L2, ..., L6, “001”,
"010", "011", "100", "101",
A code of "110" is assigned to each. The black level is “000” and the white level is “111”.

Therefore, in the above example, black 4 → L → white 3 → 001 → 011 → 110 → white 4 → L → white 3 → 110 → 100 → 010 → black 4.

The data thus run-length coded is supplied to the Huffman coding section 116, and Table 4,
Huffman coding is performed with reference to the code tables shown in Tables 5 and 6. The codes in these tables are Huffman codes created in advance as statistically efficient for various images, and short codes are assigned in order of the run length with the highest occurrence probability. For example, the black 4
Is encoded as “1011” by referring to “4” in the background run length from Table 4. Also, white 4
Is encoded as "011" by referring to "4" in the character part run length from Table 4. The same applies to image data of other run lengths.

In the above example, the code data is output in the order of 1011 → 11 → 10 → 001 → 011 → 110 → 011 → 11 → 10 → 110 → 100 → 010 → 1011.

With this method as well, the same effect can be obtained with the only difference in the compression method. In addition, in both cases, the amount of code after compression is about 1.5 to 2 times as compared with the case where a superimposed image is treated as a binary image and is encoded by the G3 standard method. It is possible to provide superimpose image data including high quality characters and symbols.

(Data Decompression Device) Further, the decompression device can be constructed by the procedure reverse to the above-mentioned compression encoding method. When decompressing, the brightness and color data of the character part, the background part are arbitrary, and for example, like the decompressing device A of FIGS. 1A and 1B, the white part is 100%, ESC
6 is 90% ... ESC1 is 15% and the black part is 0%, the mixing ratio value (mixing value) is set, and the intermediate value levels are combined. However, in consideration of image fidelity, it is more preferable to set the ESC level to a level ratio when multi-valued (for example, 170 is 170/256).

Also, like the decompressor B, the white part has a first reference value (for example, blue which is a superimpose signal) 1.
00%, ESC6 90% ... ESC
It is also possible to set the mixing ratio value (mixing value) with 1 being 15% and the black part second reference value (for example, the background red) being 0%, and reconstructing the intermediate value level to change the color. ..

<Second Embodiment> (Data Compressing Device) Next, another embodiment will be described.
In the above-described embodiment, two systems are used and the output data is alternately multiplexed every other line.
FIG. 3 is a block diagram of the data compression device. As shown in FIGS. 5 (A) and 5 (B), character information and information in which a desired shade is applied around the character information are prepared in advance. Applies white. Of course, the boundary part includes multivalued information for the purpose of high quality. These two types of images are sequentially encoded in the scanning direction by the method of the above-described embodiment. Both data are alternately multiplexed by the MUX 121 at the time of output as shown in FIG. The line counter 120 sends a control signal to the Huffman encoder 116 and the MUX 121 so that multiplexing can be performed alternately.

(Data Decompression Device) This is a mixture ratio value (mixing value) of the maximum value and the minimum value of the gradation corresponding to the intermediate value of each image of both the expanded images after expanding the image data. Depending on the, it is a method of mixing and playing. This will be described with reference to FIG. The two pieces of data in the figure show a state in which the data is expanded on the line memory.
Although the maximum value is set to the white level and the minimum value is set to the black level for both the shadow and the character, the mixing values of two levels are actually shown in FIG. 6 so that they are expanded to two line memories by using these values. Mix the data. The background, shadow, brightness and color of characters are arbitrarily determined during expansion. As an example, assuming that the data corresponding to the background, shadow, and character areas are H, I, and M, the data in areas 1 to 7 in FIG. 6 are calculated as follows.

Region 1 H Region 2 H × (100-ESC1) + I × ESC1 Region 3 H × (100-ESC3) + I × ESC3 Region 4 I Region 5 I × (100-ESC2) + M × ESC2 Region 6 I × ( 100−ESC4) + M × ESC4 region 7 M In this way, even image data having a shadow outline around a character can be highly compressed with high quality.

[0046]

As described above, according to the data compression apparatus and the decompression apparatus thereof according to the present invention, it is possible to compress image data with a high compression rate and high quality. In addition, by compressing multiple images that have been converted into data with multiple tones using the method described above, and multiplexing the output data for each preset unit, you can Shades can be encoded together. Furthermore, since the mixing value corresponding to the intermediate value is encoded, any brightness or color level can be set during expansion, and good characters, symbols, or shadows can be generated by using the mixing value. .. Also, characters, symbols, and shadows can be encoded at the same time without complicating the encoding algorithm.

[Brief description of drawings]

FIG. 1 is a diagram illustrating a method of processing halftones in a data compression device and a data expansion device according to the present invention.

FIG. 2 is a configuration diagram showing a data compression apparatus according to a first embodiment of the present invention.

FIG. 3 is a configuration diagram showing a data compression device of a second embodiment which is an example of multiplexing.

FIG. 4 is an explanatory diagram of character source data.

FIG. 5 is an explanatory diagram of source data of characters and shadows.

FIG. 6 is an explanatory diagram showing a halftone level processing method according to a second embodiment.

FIG. 7 is an explanatory diagram showing a problem caused by binarization of image data.

FIG. 8 is a configuration diagram of a conventional example.

FIG. 9 is a configuration diagram of a conventional example.

[Explanation of symbols]

110 ... Pre-processing unit, 112 ... Run length counter,
114 ... Run length coding unit, 116 ... Huffman coding unit, A, B, C ... Multiplexed bit stream, ESC
1, ESC2, ESC3, ESC4, ESC5, ESC
6, L1, L2, L3, L4, L5, L6 ... Halftone level.

Claims (7)

[Claims] 1. A data compression device for compressing an image having an intermediate gradation, wherein data conversion means for expressing an image in a plurality of gradations and image data obtained by the data conversion means are converted into one of the gradations. The maximum value and the minimum value data are compression-encoded as they are, and the intermediate value data of the gradation are encoded as a mixture ratio value of the maximum value and the minimum value of the gradation. A data compression device characterized by being provided. 2. A data decompression device for decompressing image data having an intermediate gradation, which is compressed by the data compression device according to claim 1, and means for expanding / decompressing the compressed image data, A maximum value and a minimum value as first and second reference values, and means for expanding and reproducing an image having an intermediate gradation based on a mixture ratio value of the first and second reference values. Data decompression device. 3. A data compression apparatus for compressing data having an intermediate gradation in consideration of run length, and a data converting means for expressing an image with a plurality of gradations, and image data obtained by the data converting means. On the other hand, the maximum value and the minimum value data of the gradation are subjected to the run length coding and the variable length coding, and the data of the intermediate value of the gradation is the maximum value and the minimum value of the gradation. And a coding means for coding with a code assigned to the variable length coding system as a mixed value of the data compression apparatus. 4. A data compressing device for compressing data of an image having an intermediate gradation in consideration of run length, and data converting means for expressing the image in a plurality of gradations, and image data obtained by the data converting means. On the other hand, the maximum value and the minimum value data of the gradation are subjected to the run length coding and the variable length coding, and the data of the intermediate value of the gradation is the maximum value and the minimum value of the gradation. A code provided in the variable-length coding system is assigned as the mixed value of, and a coding means for coding by adding the run length of this mixed value and the code indicating each level is provided. Data compression device. 5. The data compression apparatus according to claim 1, wherein the image data having the intermediate gradation is a background image signal and a superimposed image signal. 6. The data decompression device according to claim 2, wherein the background image signal value is used as a first reference value and the superimposed image signal value is used as a second reference value. 7. A data compression apparatus according to claim 1 or a data decompression apparatus according to claim 2, wherein a plurality of image data having halftones are processed for each set unit. apparatus.