JPH0723238A - Picture data compression and decoding device - Google Patents

Abstract

PURPOSE:To enhance the compression ratio of universal coding by selecting an exclusive dictionary corresponding to the difference from statistic property depending on a line pitch and a dot period of a dot picture with respect to the picture data compression and decoding device applying universal coding to 2-dimension picture data such as a character and dots after they are converted into a pattern run length code. CONSTITUTION:A space dictionary 16-1 and a non-space dictionary 16-2 are provided as dictionaries used for universal coding and whether or not plural lines of conversion objects to intermediate data are space is discriminated by a dictionary selection means 14 by line pitch information of a character picture and one of the dictionaries 16-1, 16-2 is selected and a pattern run length code by an intermediate coding means 10 is coded by a universal coding means 12. A dictionary is selected similarly by a decoder. Moreover, a specific dictionary is selected from the dot picture according to its dot period.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data compression and decompression device for encoding and compressing image data of a one-dimensional bit stream such as characters and halftone dots into a universal code and decoding the code data, and more particularly to 1. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data compression / decompression device that universally encodes a three-dimensional bitstream after converting it into intermediate data of a pattern and run length.

In recent years, office automation has been developed, and a document has been handled as a binary image information in a facsimile or an optical disk file system. When the document information is used as digital data, the data amount of the image information is much larger than that of the character image and is ten to several tens of times. Further, recently, in order to improve the image quality, in the facsimile, the conventional G
About 200 dpi for three aircraft, 300 dpi for the next G4 aircraft
The resolution increases to i, 400 dpi, and the amount of data tends to increase. Therefore, in order to handle image information efficiently for storage and transmission, it is essential to reduce the data amount by adding efficient data compression.

[0003]

2. Description of the Related Art Conventionally, as a compression method for a monochrome binary image,
A universal encoding method has been already proposed after performing conversion processing to intermediate data that captures the two-dimensional characteristics of an image. According to this method, two typical conventional methods, the MMR method and the prediction method, have been proposed. A compression ratio comparable to the encoding scheme is obtained.

FIG. 24 shows a conventional data compression apparatus, which includes a pattern run length conversion unit 10 and an LZW encoding unit 1.
It consists of 2. FIG. 25 shows a conventional image data restoration device, which is composed of an LZW decoding unit 18 and a pattern run length inverse conversion unit 20. FIG. 26 shows conversion of image data and pattern run-length code by taking monochrome image data as an example. Image data is usually input as a one-dimensional bit stream. This image data is divided into line data having a predetermined bit length corresponding to the actual image as shown in FIG. This two-dimensional original image data is converted into intermediate data represented by a combination of the number of consecutive black-and-white patterns and the same pattern, that is, run length information, as shown in FIG. . For the pattern run-length code converted in this way, universal coding for converting to, for example, an LZW code is applied.

To restore the code data, after decoding the pattern run length code from the LZA code, the inverse original transform shown in FIG. 26 is performed to restore the original original image data and finally output as a one-dimensional bit stream. . This method is 2
Since the two-dimensional information of the value image is taken in by preprocessing and the result is universally encoded, the data can be efficiently compressed by a simple circuit without depending on the type of the data.

That is, an input image is converted into intermediate data of a pattern and a run length of a plurality of lines, and the statistical properties such as continuity of the pattern and the run length are learned from the intermediate data by a universal coding method. The optimization is performed so that the images having various properties are efficiently compressed. Further, the universal code is originally an information storage type data compression method, and since the statistical property of the information source is not assumed in advance at the time of data compression, it can be applied to various data such as character codes and object codes. . In a document image, there are similarities in character outlines and character spacing. Also, the halftone dot image has halftone dot periodicity,
The dot shapes are similar in identity. The redundancy with this similarity can be reduced by universal coding, and effective compression can be performed.

Here, the LZW code which is one of the universal coding is taken up (TAWelch, "A Technique f
or High-Performance Data Compression ", Computer, Jun
e1984). In the LZW code, the next symbol is incorporated in the next subsequence so that the next symbol can be encoded only by the index. FIG. 27 shows the tree structure of the dictionary in the LZW code, and FIG. 28 shows the encoding of the character string in the LZW code. Further, detailed algorithms of encoding and decoding of LZW code are shown in FIGS. 29 and 30. The LZW encoding has a rewritable dictionary, divides the input character code and data into different character strings, adds the number (index i) to the character string in the order of appearance, and registers it in the dictionary. The encoded character string is represented by the number of the longest matching character string registered in the dictionary and is encoded.

In the LZW encoding process of FIG. 29, first, in step S1, a character string consisting of one character for all characters is registered in advance as an initial value, and then encoding is started. Step S2
Then, the input first character K is used as a reference number ω for dictionary search, and this is used as a prefix string. Next, in step S3, the next character K of the input data is read, and in step S4, the character string (ωK) obtained by adding the character K read in step S3 to the initial character string ω obtained in step S2.
Is searched for in the current dictionary.

If the character string (ωK) is found in the dictionary in step S4, the character string (ωK) is replaced with the reference number ω in step S5, and it is determined in step S5 whether or not the input data has ended. Returning to this, the search for the maximum matching length is continued until the character string (ωK) cannot be searched from the dictionary.
Next, in step S4, if the character string (ωK) is not in the dictionary,
Proceeding to step S7, the reference number ω of the character K obtained in step S2 is output as the code word code (ω).

A new reference number is added to the character string (ωK) and registered in the dictionary, and the input character K in step S2 is added.
Is replaced with a reference number ω, the dictionary address N is incremented, and after the check in step S5, the process returns to step S2 to read the next character K. Figure 3
1, the encoding will be described in detail with reference to FIG. Note that, for simplification of description, the case where data consisting of a combination of three letters abc is compressed is taken up. First, the input data of FIG. 31 is read from left to right. When the first character a is input, the output code (reference number ω) is output as a code word because there is no matching character string in the dictionary other than a.

The reference character 4 is added to the expanded character string ab.
And add it to the dictionary. The actual registration is a character string (1b)
Will be in the form of. Then, the second b becomes the beginning of the character string.
Since there is no matching character string other than b in the dictionary, reference number 2 is output as a code word, and the expanded character string ba is actually added to reference number 5 in the form of 2a and registered in the dictionary. The third a is the beginning of the next character string. Hereinafter, this processing is similarly continued.

The decoding process of FIG. 30 performs the reverse operation of the encoding of FIG. First, in step S1, similarly to the encoding, a character string consisting of one character for every character is registered in the dictionary in advance as an initial value, and then decoding is started. First, step S2
The first code (reference number) is read with and the current CODE
Is the OLD code, and the first code corresponds to any one-character reference number already registered in the dictionary, so the character code (K) that matches the input code CODE is searched for and the character K is output. The output character (K) is set in char for exception processing later.

Next, in step S3, the next code is read and set in CODE as NEWcode. Next, in step S4, it is checked whether the code CODE input in step S3 is defined (registered) in the dictionary. Normally, since the input codeword is registered in the dictionary by the processing up to the previous time, the process proceeds to step S5 and the code CO
The character string code (ωK) corresponding to DE is read from the dictionary, and the character string K is temporarily stacked in step S6.
The reference number code (ω) is returned to step S5 as a new CODE, and the procedure of steps S5 and S6 is recursively repeated until the reference number ω reaches one character.

Finally, the process proceeds to step S7 and step S6.
The characters stacked at are popped up and output in the LILO (Last In Fast Out) format. At the same time, in step S7, a new reference number is added to the character string representing the previously used code ω and the first character K of the character string restored this time as a set (ω, K) and registered in the dictionary. Note that step S4
If the code is not registered in (which occurs when the immediately preceding reference number is referred to in encoding), step S9
, OLDcode to CODE, code (OLD
After returning (code, char) to NEWcode, the process proceeds to step S5.

The decoding process will be specifically described with reference to FIG. 33 as follows. In addition, in order to simplify the explanation a
The case of restoring data consisting of a combination of three characters bc is described. First, the first input character is 1, and since the characters a, b, and c are already registered in the dictionary as reference numbers 1, 2, and 3 as shown in FIG. 32, the reference to the dictionary matches the code 1. The character string a of the reference number is replaced and output.

Similarly, the following code 2 is replaced with the character b and output. At this time, the code processed previously is combined with the first character b decoded this time (1b) and is added to the dictionary with a new reference number 4. The third code 4 replaces 1b with ab by searching the dictionary and outputs the character string ab. Similarly, a character string 2a obtained by combining the previously processed code 2 and the first character a of the character string decoded this time
A new reference number 5 is added to (= ba) and registered in the dictionary. Similarly, this process is repeated thereafter.

However, there is the following exception processing in the decoding of FIG. This exception processing occurs in the decoding of the sixth input code 8. Code 8 is not defined in the dictionary at the time of decoding,
I can't decrypt. In this case, the character string 5b obtained by adding the first character b of the previously decoded character string ba to the code 5 processed last time is obtained and further replaced with 2ab and bab and output. Then, the reference number 8 is added to the character string 5b obtained by adding the character b of the character string decoded this time to the previous code word 5 to the output word of the character string and registered in the dictionary.

This exception processing is performed through the processing of steps S4 and S9 in the decoding processing flow of FIG.
Finally, in step S7, the character string is output and the new character string is registered in the dictionary with the reference number added. Incidentally, FIG.
9. The encoding process and the decoding process in FIG. 30 are performed while creating the same dictionary.

[0019]

However, the conventional method of converting the image data into the pattern run-length code and then the universal code, and the pattern-run length code decoded from the universal code into the original image data is reversely converted. The device had the following problems.

Usually, in the case of a character image created by a word processor or the like, blank spaces are periodically present because the line pitch between character lines is constant. In this case, the character portion and the blank portion can be separated, but the statistical properties of the image are different. Similarly, in the case of a halftone dot image, a similar pattern appears for each halftone dot period, but the statistical properties of the image are different in the line groups having different phases of the halftone dot period.

However, since the conventional method uses a single dictionary in the universal encoding part for converting to the LZW code, it is necessary to prepare a mixed dictionary for data having different properties, or to prepare for each data type. Since there are only two methods of clearing the dictionary and re-learning, there is a problem that the compression rate cannot be increased sufficiently. The present invention has been made in view of such a conventional problem, and classifies the data according to the difference in statistical properties depending on the line pitch of the character image and the halftone dot period of the halftone dot image to create a dedicated dictionary. It is an object of the present invention to provide an image data compression / decompression device that is associated with the compression rate of universal encoding.

[0022]

FIG. 1 is a diagram for explaining the principle of the first invention of the present application, which targets character images. First, the present invention is based on the intermediate coding means (pattern run length coding means) 10 and the universal coding means 1 shown in FIG.
An image data compression device provided with 2. Here, the intermediate encoding means 10 divides the image data composed of the one-dimensional bit stream into line data of a predetermined length and arranges them into two-dimensional data, and the two-dimensional data is divided by a predetermined number of lines in the division direction. It is converted into intermediate data that is a set of run lengths indicating the number of consecutive patterns and the same pattern in the line direction. The universal encoding means 12 encodes the intermediate data converted by the intermediate encoding means 10 into a universal code using a dictionary.

First of all for such an image data compression apparatus
The invention provides a blank dictionary 16-1 and a non-blank dictionary 16-2 as dictionaries used for universal encoding. Then, based on the line pitch information of the character image given in advance, the dictionary selecting means 14 judges whether or not the plurality of lines which are the conversion targets to the intermediate data are blank, and the two dictionaries 16-1 and 16 are provided.
It is characterized in that one is selected from -2.

The present invention is also directed to an image data restoration device provided with a universal decoding means 18 and an intermediate decoding means (pattern / run length inverse conversion means) 20 shown in FIG. 1 (b). Here, the universal decoding means 18
Uses a dictionary to decode two-dimensional intermediate data, which is a set of a pattern in the dividing direction and a run length indicating the number of consecutive identical patterns in the line direction, for each predetermined number of lines from the code data. Further, the intermediate decoding means 20 converts the two-dimensional intermediate data decoded by the universal decoding means 18 into 1
Inverse conversion into image data composed of a three-dimensional bit stream.

First of all, regarding such an image data restoration device
In the present invention, the blank dictionary 16-1 and the non-blank dictionary 16-2 are provided as the dictionaries used for universal decoding. Then, based on the line pitch information of the character image given in advance, the dictionary selection means 14 determines whether or not a plurality of lines to be converted into intermediate data are blank, and the two dictionaries 16-
The corresponding one is selected from 1, 16-2.

FIG. 2 is an explanatory view of the principle of the second invention of the present application, which is also intended for a character image. In the image data compression apparatus of FIG. 2A, a blank dictionary 16-1 and a non-blank dictionary 16-2 are provided as dictionaries used for universal encoding, and the dictionary selecting means 22 is a conversion target to intermediate data. It is characterized in that the frequency of appearance of the pattern is obtained for each of a plurality of lines to determine whether or not there is a blank, one of the two dictionaries is selected, and the selected dictionary number is encoded.

In the image data restoration device of FIG. 2B, similarly, a blank dictionary 16-1 and a non-blank dictionary 16-2 are provided as dictionaries used for universal decoding, and the dictionary selection means 24 uses the code data. It is characterized in that one of the two dictionaries is selected according to the dictionary number restored from. That is, in the second invention, the appearance frequency of the pattern of the intermediate data is learned and the dictionary is automatically selected.

FIG. 3 shows a third invention of the present application, which is intended for a halftone dot image. In the image data compression apparatus of FIG. 3A, a plurality of dictionaries 32-1 to 32-n corresponding to halftone dot periods are provided as dictionaries used for universal encoding. And
It is characterized in that the dictionary selecting means 26 selects one dictionary corresponding to a plurality of lines to be converted into intermediate data from a plurality of dictionaries 32-1 to 32-n according to halftone dot period information given in advance. To do.

In the image data restoration device of FIG. 3B, similarly, a plurality of dictionaries 32-1 to 32-n corresponding to halftone dot periods are provided as dictionaries used for universal decoding, and the dictionary selection means 26 It is characterized in that one dictionary corresponding to a plurality of lines to be converted into intermediate data is selected from a plurality of dictionaries 32-1 to 32-n according to halftone dot period information given in advance.

FIG. 4 is an explanatory view of the principle of the fourth invention of the present application, which similarly targets a halftone dot image. In the image data compression apparatus of FIG. 4A, a plurality of dictionaries 32-1 to 32-corresponding to halftone dot periods are used as dictionaries used for universal encoding.
n is provided, and the dictionary selection means 32 obtains the distribution of the appearance frequency of the pattern for each of the plurality of lines that are the conversion target to the intermediate data, and the plurality of dictionaries 32-1 are obtained from the difference with the appearance frequency obtained in the past. 32 to 32-n or a new created dictionary is selected and encoded together with the dictionary number of the selected dictionary.

In the image data restoration device of FIG. 4B, a plurality of dictionaries 32-1 to 32-n corresponding to halftone dot periods are provided as dictionaries used for universal decoding, and the dictionary selection means 34 causes the code data to be encoded. It is characterized in that one of the plurality of dictionaries 32-1 to 32-n is selected according to the dictionary number restored from, or a new created dictionary is selected for decoding.

Here, the image data compression or decompression device for the halftone image of FIGS.
For the division number n of 2-n, the least common multiple of the halftone dot period and the number of lines to be converted is obtained, and the value obtained by dividing the least common multiple by a plurality of lines to be converted is the number of divisions of the dictionary. That is, the number of divisions N = least common multiple (halftone dot period, number of lines) / number of lines.

[0033]

As described above, according to the present invention, in the case of a character image, two dictionaries 2 corresponding to a character and a blank are prepared, intermediate data such as pattern run length code is converted as a pre-process, and is subsequently given. A dictionary is selected according to the information such as the line pitch of the characters and encoded into a universal code. In addition, the pattern appearance frequency distribution is calculated for the two-dimensional data for each line that is converted to the pattern run length code as the intermediate data, and if it is completely blank, the dictionary corresponding to the blank is selected. Select and encode the corresponding dictionary.

On the other hand, in the case of a halftone dot image, a plurality of dictionaries corresponding to different phases of the halftone dot period are prepared, intermediate data such as pattern run length code is converted as preprocessing, and then given. A dictionary is selected according to the phase of the halftone dot period and encoded into a universal code. In addition, the distribution of the appearance frequency is obtained for the two-dimensional data for each of the plurality of lines to be converted into the pattern run length code as the intermediate data, and the appearance frequency distribution is compared with the previous appearance frequency distribution, and the plurality of lines with the smallest difference are obtained. The dictionary corresponding to is selected and encoded. At this time, if the difference from the distribution of the previous appearance frequency is equal to or more than a predetermined threshold value and no correspondence is obtained, a new dictionary is created and encoded.

[0035]

FIG. 5 shows an embodiment of the image data compression apparatus of the first invention for processing a document image. In this embodiment, the image data compression device uses a pattern
A run length converter 10 and a universal encoder 12 that performs LZW encoding are provided. Blank dictionary 16-1 as a dictionary used for LZW encoding in the universal encoder 12
And a character dictionary 16-2.

The dictionary selection unit 14 inputs the line pitch of characters from the document format information predetermined for the input character image, and the pattern run length code currently encoded by the universal encoder 12 is blank in the document image. Or character. If it is blank, the blank dictionary 16-1 is selected, and if it is a character, the character dictionary 16-2 is selected. Here, the pattern / run-length converter 10 is, for example, as shown in FIG.
, The character data input as a one-dimensional bit stream is divided into line data of a predetermined length and 2
For example, four lines are arranged in a dimension and converted into a pattern / run length code which is a combination of run lengths in which the type of pattern and the same pattern are repeated.

The universal encoder 12 is, for example,
The pattern run length code as intermediate data obtained by the pattern run length converter 10 according to the LZW encoding algorithm shown in FIG. 29 is converted into the blank dictionary 16-1 or the character dictionary 16-
2 is used to perform LZW encoding. FIG. 6 is a flowchart showing the encoding process of FIG. First, step S1
The dictionary used for universal encoding is the blank dictionary 16
-1 and the character dictionary 16-2. Then, in step S2, the current image data of the character image is input to the pattern / run length converter 10. This current image data is 1
Input as a dimensional bitstream. Subsequently, in step S3, the line image is divided into line data for each bit length corresponding to the number of characters of one line according to the document format of the character image to be processed, and the line data is arranged in the line direction to form the two-dimensional image shown in FIG. Convert to data. Then, with respect to the two-dimensional data, for example, four lines are set as one block, and conversion into a pattern run length code is performed which is a combination of a pattern in the row direction and a run length indicating repetition of the same pattern in the line direction.

Then, in step S4, a dictionary used for encoding by the universal encoder 12 is selected based on the line pitch information of the character at this time. Now, it is assumed that universal encoding is performed using a dictionary for a pattern run length obtained from a character image converted into two-dimensional data as shown in FIG. It is recognized from the character line pitch information that the character image converted into the two-dimensional data has a character portion and a blank portion each having four lines.

That is, the first block composed of the first four lines is recognized as a character portion from the character line pitch information,
Therefore, the character dictionary 16-2 is selected. Regarding the second block of the next four lines, it is recognized from the character line pitch information that it is a blank portion, and the blank dictionary 16-1 is selected. Below, for odd-numbered blocks, character dictionary 1
6-2 is selected, and the blank dictionary 16-1 is selected for even blocks.

Referring again to FIG. 6, when the dictionary is selected from the character line pitch information in step S4, step S5
The blank dictionary 16-1 or the character dictionary 16 selected in
-2, encoding and new dictionary registration are performed.
The encoding using this target dictionary is, for example, a process according to the AZW encoding algorithm shown in FIG. When the encoding for one block is completed in step S5, step S6
Then, the encoded data is output, the process returns to step S4, the same process is repeated for the next block, and the same process is repeated until all blocks are completed.

FIG. 8 shows an embodiment of the image data decompression device of the first invention, in which the code data obtained by the code data compression device of FIG. 5 is restored to the image data of the original document image. This image data restoration device includes a universal decoder 18, a pattern run length inverse converter 20 as an intermediate decoding means, a blank dictionary 16-1, a character dictionary 16-2, and a dictionary selection unit 14.
Composed of.

The universal decoder 18 uses a dictionary to
For example, the pattern run length code is restored from the code data according to the coding algorithm shown in FIG. At this time, the dictionary selection unit 14 uses the blank line dictionary 16-1 or the character dictionary 16-2 for each block of the pattern run length code to be restored based on the line pitch information of the characters of the document image to be restored which is given in advance. Select. The pattern run length inverse converter 20 is a universal decoder 1
The pattern run length code restored in 8 is inversely converted into two-dimensional data consisting of a pixel array of four lines, and then output as a one-dimensional bit stream.

FIG. 9 is a flow chart showing the restoration processing of FIG. First, in step S1, the dictionaries used for universal decoding are the blank dictionary 16-1 and the character dictionary 16.
-2. Next, in step S2, code data is input. Subsequently, in step S3, the blank dictionary 16-1 or the character dictionary 16- is created on the basis of the line pitch information of the character given in advance for the code data for one block.
Select one of the two.

Specifically, as shown in FIG. 7, since the code data of the first block is a character portion, the character dictionary 16-2 is selected, and the next second block is a blank portion. Select the blank dictionary 16-1 from the following, and similarly, for the odd blocks, the character dictionary 16-2,
For even-numbered blocks, the blank dictionary 16-1 is selected. When the dictionary is selected in step S3, the process proceeds to step S4, the selected target dictionary is referred to, decoding is performed to convert the code data into the pattern run length code, and new dictionary registration is repeated. When the universal decoding for one block is completed, the process proceeds to step S5, the pattern run length code is inversely converted into two-dimensional data according to the type of pattern and the run length, and the image data of the original document image is restored.

The processes of steps S3 to S5 are repeated for all blocks or a predetermined number of blocks. When the restoration processing for all blocks or a predetermined block unit is completed, the process proceeds to step S6, and the restored data restored to the two-dimensional image is output as a one-dimensional bit stream. FIG. 10 shows an embodiment of the image data compression apparatus according to the second invention. In this embodiment, the frequency of appearance of a pattern for each block composed of a predetermined number of lines to be converted into a pattern run length code, for example, 4 lines. Is characterized by detecting whether or not there is a blank. That is, the pattern / run length converter 10, the universal encoder 12, the blank dictionary 16-1 and the character dictionary 16-2 are the same as those in the embodiment of the first invention shown in FIG. A dictionary selection unit 22 for detecting is provided.

The dictionary selection unit detects the appearance frequency of the blank pattern for one block of the two-dimensional data converted by the pattern / run length converter 10. That is, when the appearance frequency of the blank pattern of the bit 0000 indicating a blank is obtained as an extremely high value with respect to other patterns, it is determined to be blank, the blank dictionary 16-1 is selected, and the universal encoder 12 is selected. Encode. On the other hand, with respect to the character portion, since there is the frequency of appearance of other bit patterns in addition to the blank pattern of all bits 0, it is judged that the character portion and the character dictionary 16-2 is selected and the universal encoder 1 is selected.
2 to be used for encoding.

FIG. 11 is a flow chart showing the processing operation of the image data compression apparatus using the dictionary selection unit 22 for performing the blank detection of FIG. In this case, the appearance frequency of the pattern is obtained in block units in step S3, and if there is only a blank, the blank dictionary 16-1 is selected, and if not, the character dictionary 16-2 is selected. Furthermore, in order to enable dictionary selection on the restoration side, the universal encoder 12
The dictionary number is first encoded prior to universal encoding of the run length code. The other processing is the same as the encoding in FIG.

FIG. 12 shows an embodiment of an image data decompression device for decompressing the image data of the original character image from the coded data obtained by the image data compression device of FIG. In this embodiment, the universal decoder 18 is provided with two blank dictionaries 16-1 and character dictionaries 16-2.
Select any one. The dictionary selection unit 24 decodes the dictionary number encoded at the beginning of the block portion of the coded data, and based on the decoded dictionary number, the blank dictionary 16-1
Alternatively, the character dictionary 16-2 is selected. In the universal decoder 18, the blank dictionary 16-1 or the character dictionary 16-
The pattern run length code decoded using 2 is
The pattern / run length inverse converter 20 restores the two-dimensional image data and finally outputs it as a one-dimensional bit stream.

FIG. 13 is a flow chart showing the restoration processing of the image data restoration device of FIG. 12, wherein the head encoded dictionary number of the block portion of the code data is restored in step S2, and based on the restored dictionary number. And blank dictionary 1
A feature is that 6-1 or the character dictionary 16-2 is selected. The rest is the same as the restoration process of FIG. 14
Shows an embodiment of an image data compression apparatus of the third invention for performing encoding on a halftone image. This image data compression apparatus is composed of a pattern / run length converter 10, a universal encoder 12, a dictionary selection unit 26, and a dictionary unit 28.
The dictionary unit 28 divides the dictionary into the dictionaries 30-1 to 30-n based on the relationship between the halftone dot period given in advance and the number of lines forming one block to be converted.

FIG. 15 specifically shows the dictionary division for the halftone image. In this halftone dot image, the halftone dot period is 12, and when converted into a pattern run length code every four lines, there are three blocks having different phases of the halftone dot period. Therefore, the patterns that are likely to appear in each of the three blocks are different, and therefore it is necessary to divide the dictionary into three dictionaries 30-1 to 30-3 and perform universal coding for each corresponding block of 4 lines.

That is, a value obtained by dividing the least common multiple of the halftone dot period and the number of lines in one block to be converted by the number of lines in one block is the number of divisions n in the dictionary. In the case of FIG. 15, the least common multiple of the halftone dot period 12 and the number of lines in one block is 12, and when this is divided by the number of lines in one block, the division number n.
= 3. Referring again to FIG. 14, the dictionary selection unit 26
Is a pattern run length converter 10 that converts halftone dot image data into a pattern run length code in block units, and then a relationship between a halftone dot period given in advance and a plurality of lines of the block currently being converted. From the dictionary section 2
The target dictionary is selected from the eight dictionaries 30-1 to 30-n, and the universal encoder 12 performs the encoding.

FIG. 16 is a flow chart showing the processing operation of the image data compression apparatus for the halftone image of FIG. First, in step S1, one dictionary is divided into a plurality of dictionaries 30-1 to 30-in accordance with the difference between the halftone dot period information and the phase, that is, the division number n determined by the number of lines forming one block.
Divide into n. Then, the current image data of the halftone image is input in step S2, and the pattern / run length converter 10 is input.
Is converted into a two-dimensional pattern run length code.

Then, in step S4, the corresponding dictionary 30-is determined from the relationship between the halftone dot period given in block units and the plurality of lines of the block currently being converted.
Select i. Subsequently, in step S5, universal coding and dictionary registration are performed using the selected dictionary, and when coding of one block is completed, code data is output in step S6. The above steps S4 to S6 are repeated until a predetermined block or all blocks are completed.

FIG. 17 shows an embodiment of an image data decompression device for decompressing the original halftone dot image from the coded data obtained by the image data compression device of FIG. Also in this embodiment, as in the case of FIG. 14, the halftone dot period and the conversion target 1
The dictionary unit 28 is divided into the dictionaries 30-1 to 30-n with the division number n determined from the relationship with the number of lines of the block. Further, the dictionary selection unit 26 selects a target dictionary from the relationship between a plurality of lines corresponding to the block currently to be decoded by the universal decoder 18 and a halftone dot period given in advance.

FIG. 18 is a flow chart showing the restoration processing of FIG. In this restoration process, step S
16, the dictionary is divided by the number of divisions n determined by the number of lines in one block determined by the difference in halftone dot period and phase, as in the case of FIG. Further, in step S3, the corresponding dictionary is selected from the relationship between the plurality of lines corresponding to the block currently being restored and the halftone dot period given in advance.

FIG. 19 shows an embodiment of the image data restoration device of the fourth invention for a halftone image. In this image data restoration device, the dictionary selection unit 32 obtains the distribution of the appearance frequency of the pattern for the halftone dot image data of a plurality of lines forming the block to be converted, and the distribution of the appearance frequency is calculated as before. It is characterized in that the dictionary corresponding to the block having the smallest difference is selected and encoded by comparing with the distribution of the appearance frequency of the block. If the difference from the appearance frequency of the previous block is larger than a predetermined threshold value, a new dictionary is created and encoded.

FIG. 22 shows the distribution of the appearance frequencies of past blocks in the dictionary selection unit 32 and the appearance frequency distribution of the blocks currently being processed. 22A, A, B,
C is the distribution of the appearance frequency of the previous block. In such a state, it is assumed that the distribution D of the appearance frequency of the block currently being processed shown in FIG. 22B is obtained. When this is overlaid on the appearance frequency distributions A, B, and C, it becomes as shown by the broken line. In this case, since the appearance frequency distribution B has the smallest difference from the appearance frequency distribution B of the previous block, the dictionary used for encoding the block having the appearance frequency distribution B is selected. In addition, the dictionary selection unit 32 uses the dictionary 30 in the dictionary unit 28.
When any one of -1 to 30-n is selected, or when a new dictionary 30- (n + 1) is selected, the block is added to the head of the block before the universal run encoder 12 encodes the pattern run length code. Encode the selected dictionary number.

FIG. 21 is a flow chart showing an image selection process for selecting a dictionary from the distribution of the appearance frequencies of the patterns shown in FIG. The frequency is calculated, and the dictionary is selected based on the difference from the distribution of the previous appearance frequency. If the difference from the previous appearance frequency exceeds a predetermined threshold, a new dictionary is created. Then, the selected dictionary number or the newly created dictionary number is encoded first. This is repeated for each block.

FIG. 22 shows an embodiment of the image data decompression device of the fourth invention for decompressing the original halftone dot image from the coded data obtained by the image data compression device of FIG. The universal decoder 18, the pattern run length inverse converter 20, and the dictionary unit 28 divided into the dictionaries 30-1 to 30-n are shown in FIG.
7 is the same as the third invention, but a dictionary selecting unit 34 is newly provided.

The dictionary selection unit 34 decodes the dictionary number encoded at the head of the block of the coded data, and the dictionaries 30-1 to 30-in the dictionary unit 28 based on the decoded dictionary number.
Either select n or create a new dictionary. Figure 2
3 is a flowchart showing the restoration processing of FIG.
The dictionary number sent in step S2 is restored and the corresponding dictionary is selected from the plurality of dictionaries 30-1 to 30-n, or a new dictionary is created and decoded by the universal decoder 18. To do.

[0061]

As described above, according to the present invention, a plurality of dictionaries are used, and in the case of a character image, a blank part and a character part are separately selected and corresponding dedicated dictionaries are selected. In the case of an image, by dividing a plurality of dictionaries having different phases from a plurality of lines as a halftone dot period and a processing unit and selecting a dedicated dictionary corresponding to each, without considering a portion having different statistical properties, A higher compression rate can be obtained as compared with the case of encoding using a single mixed dictionary.

[Brief description of drawings]

FIG. 1 is an explanatory view of the principle of the first invention for character images.

FIG. 2 is an explanatory view of the principle of the second invention for character images.

FIG. 3 is an explanatory diagram of the principle of the third invention for a halftone image.

FIG. 4 is an explanatory view of the principle of the fourth invention for a halftone image.

FIG. 5 is a block diagram of an embodiment of an image data compression apparatus of the first invention.

FIG. 6 is a flowchart showing an encoding process of the embodiment of FIG.

FIG. 7 is an explanatory diagram of character images and dictionary selection.

FIG. 8 is a configuration diagram of an embodiment of an image data restoration device of the first invention.

FIG. 9 is a flowchart showing a decoding process of the embodiment of FIG.

FIG. 10 is a block diagram of an embodiment of an image data compression apparatus of the second invention.

11 is a flowchart showing the encoding process of the embodiment of FIG.

FIG. 12 is a block diagram of an embodiment of an image data restoration device of the second invention.

FIG. 13 is a flowchart showing a decoding process of the embodiment of FIG.

FIG. 14 is a block diagram of an embodiment of an image data compression apparatus of the third invention.

FIG. 15 is an explanatory diagram of halftone image and dictionary selection.

16 is a flowchart showing the encoding process of the embodiment of FIG.

FIG. 17 is a block diagram of an embodiment of an image data restoration device of the third invention.

FIG. 18 is a flowchart showing a decoding process of the embodiment of FIG.

FIG. 19 is a block diagram of an embodiment of an image data compression apparatus of the fourth invention.

FIG. 20 is an explanatory diagram showing a difference in pattern appearance frequency distribution for each of a plurality of lines used for dictionary selection.

FIG. 21 is a flowchart showing the encoding process of the embodiment of FIG.

FIG. 22 is a configuration diagram of an embodiment of an image data restoration device of the fourth invention.

FIG. 23 is a flowchart showing a decoding process of the embodiment of FIG.

FIG. 24 is an explanatory diagram of a conventional image data compression device.

FIG. 25 is an explanatory diagram of a conventional image data restoration device.

FIG. 26 is an explanatory diagram showing conversion and inverse conversion between two-dimensional image data and a pattern run length code.

FIG. 27 is an explanatory diagram showing a tree structure of a dictionary used in universal encoding.

FIG. 28 is an explanatory diagram of a universal code.

FIG. 29 is a flowchart showing a conventional LZA encoding algorithm.

FIG. 30 is a flowchart showing a conventional LZW decoding algorithm.

FIG. 31 is an explanatory diagram concretely showing LZW encoding by taking abc3 characters as an example.

32 is an explanatory diagram of a dictionary used in the encoding of FIG. 31.

FIG. 33 is an explanatory diagram concretely showing LZW decoding by taking abc3 characters as an example.

[Explanation of symbols]

10: Intermediate Coding Means (Pattern Run Length Encoder) 12: Universal Coding Means (Universal Encoder) 14, 22, 24, 32, 34: Dictionary Selection Means (Dictionary Selection Unit) 16-1: Blank Dictionary 16-2: Character dictionary 18: Universal decoding means (universal decoder) 20: Intermediate decoding means (pattern run length decoder) 28: Dictionary mechanism 30-1 to 30-2: Dictionary

Claims (9)

[Claims] 1. Image data composed of a one-dimensional bit stream is divided into line data of a predetermined length and converted into two-dimensional data, and the two-dimensional data is divided into patterns in a dividing direction for every predetermined number of lines. Intermediate coding means (10) for converting into intermediate data consisting of a set of run lengths indicating the number of consecutive identical patterns in the line direction, and the intermediate coding means (1)
In the image data compression apparatus provided with a universal encoding means (12) for encoding the intermediate data converted in 0) into a universal code using the dictionary, a blank dictionary (16) is used as the dictionary for the universal encoding. -1) and a non-blank dictionary (16-2) are provided, and it is judged from the two dictionaries whether or not a plurality of lines to be converted into the intermediate data are blank based on the pitch information of the character given in advance. An image data compression apparatus, characterized in that a dictionary selection means (14) for selecting one is provided. 2. A two-dimensional intermediate consisting of a set of a pattern in a dividing direction and a run length indicating the number of consecutive identical patterns in a line direction for every predetermined number of lines from code data universally coded using a dictionary. Universal decoding means (18) for decoding data, and intermediate decoding means (20) for inversely converting the two-dimensional intermediate data decoded by the universal decoding means (18) into image data composed of a one-dimensional bit stream. ) And an image data decompression device including a blank dictionary (16-1) and a non-blank dictionary (16-2) as dictionaries to be used for the universal decoding, and a predetermined pitch of characters is provided. A dictionary selecting means (14) for selecting one of the two dictionaries by judging whether or not a plurality of lines to be converted into the intermediate data are blank based on information is provided. An image data compression device characterized by: 3. Image data composed of a one-dimensional bit stream is divided into line data of a predetermined length and converted into two-dimensional data, and the two-dimensional data is divided into patterns in a dividing direction for every predetermined number of lines. Intermediate coding means (10) for converting into intermediate data consisting of a set of run lengths indicating the number of consecutive identical patterns in the line direction, and the intermediate coding means (1)
In the image data compression apparatus provided with the universal encoding means (12) for encoding the intermediate data converted in 0) into a universal code using the dictionary, a blank dictionary ( 16-1) and a non-blank dictionary (16-2) are provided, the appearance frequency of the pattern is determined for each of a plurality of lines that have been converted to the intermediate data, and it is determined whether the pattern is blank or not.
An image data compression apparatus comprising a dictionary selection means (22) for selecting one of the two dictionaries and encoding the selected dictionary number. 4. A two-dimensional intermediate consisting of a set of a pattern in the dividing direction and a run length indicating the number of consecutive identical patterns in the line direction for every predetermined number of lines from the code data universally encoded using a dictionary. Universal decoding means (18) for decoding data, and intermediate decoding means (20) for inversely converting the two-dimensional intermediate data decoded by the universal decoding means (18) into image data composed of a one-dimensional bit stream. ) And a blank dictionary (16-1) and a non-blank dictionary (16-2) are provided as dictionaries used for the universal decoding, and the dictionaries restored from the code data are provided. An image data compression apparatus, characterized by further comprising dictionary selection means (24) for selecting one of the two dictionaries according to a number. 5. Image data composed of a one-dimensional bit stream is divided into line data of a predetermined length and converted into two-dimensional data, and the two-dimensional data is divided into patterns in a dividing direction for every predetermined number of lines. Intermediate coding means (10) for converting into intermediate data consisting of a set of run lengths indicating the number of consecutive identical patterns in the line direction, and the intermediate coding means (1)
In the image data compression apparatus provided with the universal encoding means (12) for encoding the intermediate data converted in 0) into the universal code using the dictionary, the dictionary used in the universal encoding has a halftone dot cycle. A plurality of corresponding dictionaries (32-1 to 32-n) are provided,
Dictionary selecting means (26) for selecting, from the plurality of dictionaries (32-1 to 32-n), one dictionary corresponding to a plurality of lines which have been converted into the intermediate data according to halftone dot period information given in advance. An image data compression device, characterized by being provided with. 6. A two-dimensional intermediate consisting of a set of a pattern in a dividing direction and a run length indicating the number of consecutive identical patterns in a line direction for every predetermined number of lines from code data universally coded using a dictionary. Universal decoding means (18) for decoding data, and intermediate decoding means (20) for inversely converting the two-dimensional intermediate data decoded by the universal decoding means (18) into image data composed of a one-dimensional bit stream. ), And a plurality of dictionaries (32-1 to 32-n) corresponding to halftone dot periods are provided as dictionaries used for the universal decoding.
Halftone dot dictionary selecting means (26) for selecting one dictionary corresponding to a plurality of lines to be converted into the intermediate data from the plurality of dictionaries (32-1 to 32-n) according to halftone dot period information given in advance. ) Is provided, the image data compression device. 7. Image data composed of a one-dimensional bit stream is divided into line data of a predetermined length and converted into two-dimensional data, and the two-dimensional data is divided into patterns in a dividing direction for every predetermined number of lines. Intermediate coding means (10) for converting into intermediate data consisting of a set of run lengths indicating the number of consecutive identical patterns in the line direction, and the intermediate coding means (1)
In an image data compression apparatus equipped with a universal encoding means (12) for encoding the intermediate data converted in 0) into a universal code using a dictionary, a dictionary used for the universal encoding is A plurality of corresponding dictionaries (32-1 to 32-n) are provided,
The distribution of the appearance frequency of the pattern is obtained for each of the plurality of lines that are the conversion targets to the intermediate data, and the plurality of dictionaries (32-1 to 32-n) are obtained from the difference with the appearance frequency obtained in the past.
Dictionary selecting means (3) for selecting one of the selected dictionary or a newly created dictionary and encoding with the dictionary number of the selected dictionary.
2. An image data compression device characterized by being provided with 2). 8. A two-dimensional intermediate consisting of a set of a pattern in the dividing direction and a run length indicating the continuous number of the same pattern in the line direction for every predetermined number of lines from code data universally encoded using a dictionary. Universal decoding means (18) for decoding the data, and intermediate decoding means (20) for inversely converting the two-dimensional intermediate data decoded by the universal decoding means (18) into image data composed of a one-dimensional bit stream. ), And a plurality of dictionaries (32-1 to 32-n) corresponding to halftone dot periods are provided as dictionaries used for the universal decoding.
Dictionary selection means (34) for selecting one of the plurality of dictionaries (32-1 to 32-n) according to the dictionary number restored from the code data or selecting a new created dictionary and performing decoding. An image data compression device characterized by being provided. 9. The image data compression or decompression device according to claim 7 or 8, wherein the least common multiple of the halftone dot period and the number of lines to be converted is obtained, and the least common multiple is converted into a plurality of lines to be converted. An image data compression or decompression device characterized in that a divided value is used as the number of divisions of a dictionary.