JP3083550B2 - Data compression and decompression method - Google Patents

Description

Detailed Description of the Invention [Table of Contents] Outline Industrial field of application Conventional technology (FIGS. 13 to 18) Problems to be Solved by the Invention Means for Solving the Problems (FIG. 1) Action Embodiment (A) Description of the first embodiment (FIGS. 2 to 5) (b) Description of the second embodiment (FIGS. 6 to 8) (c) Description of the third embodiment (FIG. (FIGS. 9 to 11) (d) Description of a fourth embodiment (FIG. 12) (e) Description of another embodiment [Overview] Data compression and decompression method for compressing data by universal coding For the purpose of preventing the registration of data between unrelated data sequences and improving the compression ratio, the encoded data is divided into different sub-sequences, and the sub-sequences are registered in a dictionary and input. Searches the data for a subsequence registered in the dictionary, and substitutes the input data for the subsequence registered in the dictionary. In the data compression method of designating and encoding the data having the maximum length matching by the reference number and registering the data in the dictionary, a break of the input data is detected from the input data. The input data is to be newly registered in the dictionary.

[Industrial applications]

The present invention relates to a data compression and decompression method for compressing data by universal coding.

In recent years, various types of data such as character codes, vector information, and images have been handled by computers.
The amount of data handled is also rapidly increasing. When dealing with a large amount of data, by compressing the amount of data by omitting redundant portions in the data, it becomes possible to reduce the storage capacity or to transmit data quickly.

Universal coding has been proposed as a method that can compress various data in one system. Here, the field of the present invention can be applied not only to character code compression but also to various types of data. In the following, following the name used in information theory, one word unit of data is called a character, and Are connected to an arbitrary word.

As a typical method of universal code, Ziv-Lempel
There are (Jib-Lempel) codes (for details, for example, Munakata “Ziv-Lempel Data Compression Method”, Information Processing, Vol. 26, N
o.1, 1985).

For the Ziv-Lempel code, two algorithms of a universal type and an incremental decomposition type (Incremental parsing) have been proposed. Furthermore, as an improvement of the universal algorithm, there is LZSS code (TCBell, “Better OPM
/1.Text Compression ", IEEE Trans.on Commun., Vol.COM
-34, No. 12, Dec. 1986). As an improvement of the incremental decomposition type algorithm, there is an LZW (Lempel-Ziv-Welch) code (TAWelch, "A Technique for High-Performa").
nce Data Compression ", Computer, June 1984).

In such a data compression method using universal coding, an improvement in compression ratio is required.

[Conventional technology]

 FIG. 13 is an explanatory diagram of the prior art.

Originally, the universal code is an information preservation type data compression method. Since the statistical properties of the information source are not assumed in advance at the time of data compression, it can be applied to various types of data (character codes, object codes, etc.). Can be.

In a document image, there is a similarity in the linearity and the degree of bending of a character line of a character. In a halftone dot image, an enormous number of change points appear because the entire image is halftone dispersed, but the connection relationship of the outlines is similar due to the halftone dot periodicity and the same halftone dot shape. Redundancy having this similarity can be reduced by universal coding, and effective compression can be performed.

As described above, since universal coding is effective for image data, as shown in FIG. 13, preprocessing for converting image data into two-dimensional information (intermediate data) by MMR method or the like is performed. We have already proposed a method to further improve the compression efficiency of universal coding for images by performing this process of performing universal coding using LZW codes.

In the proposed method, the input image is converted into relative address data of changed pixels, and from the converted data, a statistical property such as continuity of converted pixels in a scanning line direction is learned by a universal coding method while learning. To achieve efficient compression of images of various properties.

That is, the MMR method (modified from the standard method) uses the detection of the mode of the changed pixel as preprocessing, and universal coding (Ziv-Lempel code) is applied to it, and the continuity of the changed pixel, the linearity of the contour line, and the curvature are applied. The global nature of the condition is represented by a universal code index.

 This universal encoding will be described using an LZW code as an example.

FIG. 14 is a flowchart of the LZW encoding process, and FIG. 15 is a flowchart of the LZW decoding process.

LZW coding has a rewritable dictionary, divides the input character code and data into different character strings, assigns numbers to these character examples in the order they appear, and registers them in the dictionary.
The currently input character string is represented by the number (index) of the longest matching character string registered in the dictionary and encoded.

 The encoding process will be described with reference to the flowchart of FIG.

First, in step S1 (hereinafter "step" is omitted), a character string consisting of one character for all characters is registered in advance as an initial value, and then encoding is started. In the encoding of S1, the dictionary is searched by the first character K input to obtain a reference number ω, which is used as a prefix string.

Next, at S2, the next character of the input data is read, and at S3, it is checked whether the character input is completed.
A search is made to see if the dictionary has a character (ωK) obtained by adding the character K read in S2 to the initial character string ω obtained in S5 or ω in S5.

If the character string (ωK) is not in the dictionary in S4, the process proceeds to S6 and S1
The reference number ω of the character K obtained in step is output as a 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 of S2 is further referred to as the reference number ω And the dictionary address n is incremented, and the process returns to S2 to read the next character K.

On the other hand, if the character string (ωK) is found in the dictionary in S4, the character string (ωK) is replaced with the reference number ω in S5, and the process returns to S2 to search for the maximum matching length until the character string (ωK) cannot be searched from the dictionary. Continue.

The decoding process of FIG. 15 performs the reverse operation of the encoding of FIG.

In the decoding of FIG. 15, similarly to the encoding, the decoding is started after a character string consisting of one character for every character is registered in the dictionary as an initial value in advance.

First, the first code (reference number) is read in S1, and the current code is read.
CODE is OLDcode, and since the first code corresponds to one of the reference numbers of one character already registered in the dictionary, a 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 FINchar for exception processing in S8 described later.

Then, the process proceeds to S2, where the next code is read and set as CODE in INCODE.

In S3, it is checked whether there is a new code, that is, whether or not the code input has been completed, and the process proceeds to S4, where the code C input in S3
Check whether ODE is defined (registered) in the dictionary.

Normally, since the input code word is registered in the dictionary in the previous processing, the process proceeds to S5, where the character string c corresponding to the code CODE is obtained.
ode (ωK) is read from the dictionary, the character string K is temporarily stacked in S6, the reference number code (ω) is set as a new CODE, and the process returns to S5 again.
Is repeated until one character is reached. Finally, the process proceeds to S7, where the characters stacked in S6 are popped up and output in LIFO (Last In Fast Out) format. At the same time, in S7, a new reference number is added to the character string represented as a set (ω, K) of the code ω used last time and the first character K of the character string restored this time, and registered in the dictionary.

A pre-processing method of image data for improving the merit of such universal coding of an image will be described with reference to FIGS. 16 and 17, taking an example of a previously proposed method as an example.

Fig. 16 is an explanatory diagram of two-dimensional encoding of image data already proposed,
FIG. 17 is an explanatory diagram of the encoding example.

As shown in FIG. 16 (B), 1-byte horizontal mode H, vertical mode V, and bus mode P are assigned as mode information.

Next, as horizontal mode run length information, 1 or 2 bytes are used for white / black run length according to the length.
As the vertical mode shift length information, one or two bytes are used for the change point shift length according to the length.

The encoding is performed using the horizontal run length of the first line, and the second line and subsequent lines are encoded using the vertical deviation from the previous line at the change point and the pass mode, similarly to the MMR method.

For example, for the image of FIG. 16 (A), two-dimensional encoded data as shown in FIG. 17 is obtained.

That is, in the first line, the horizontal mode H and the white run length RL1 and in the horizontal mode H2 and the black run length RL2, and in the second line, the change point (white → black) of the first line Vertical deviation V1-ZL21 from the point of change (black →
On the third line, the vertical shift lengths V2-ZL31 and V2-ZL32 of the change points are shown in the vertical shift lengths V2-ZL22 in white).
The horizontal mode H3 and white run length RL3 are encoded, and the horizontal mode H4 and black run length RL4 are encoded.

FIG. 18 shows an example in which such an image is two-dimensionally encoded and universally encoded.

In the figure, the input symbol is the result of two-dimensional encoding,
The output code is the universal coding result.

For example, the input symbol “10000000” is index “1
28, and (registered data + next data) is registered in the dictionary.

[Problems to be solved by the invention]

By the way, in universal encoding, regardless of the data sequence, the maximum length matching registered data plus the next one symbol is registered in the dictionary.

For this reason, the registration numbers 258,260,264,266,268,27 in FIG.
Combinations of completely unrelated data series, such as 2,274,277, are also registered.

As a result, in the related art, since data between irrelevant data series is also registered and registered, a large number of data having an extremely low appearance frequency are registered, and the compression ratio due to a delay in learning effect and an increase in index (registration number) is registered. There is a problem that it causes a decrease in

Accordingly, an object of the present invention is to provide a data compression and decompression method that can prevent registration of data between irrelevant data sequences and improve the compression ratio.

[Means for solving the problem]

 FIG. 1 is a diagram illustrating the principle of the present invention.

In the present invention, as shown in FIG. 1, the encoded data is divided into different sub-sequences, the sub-sequences are registered in a dictionary, and the sub-sequences registered in the dictionary are searched by input data. In the data compression method for registering the input data by designating the input data by the reference number of the subsequence registered in the dictionary that matches the maximum length and detecting the delimiter of the input data from the input data, Then, with the boundary as a boundary, input data after the boundary is to be newly registered in the dictionary.

The present invention also provides a data restoration method for comparing encoded data with a reference number of a subsequence registered in a dictionary and restoring the encoded data to a subsequence with a matching reference number. A data break is detected from the restored partial sequence, and a new partial sequence is registered in the dictionary at the boundary.

[Action]

According to the present invention, since a break of a data sequence is detected and input data after the break is registered as a new data sequence at the break, the registration of combined data having different data sequences is prevented, and The present invention eliminates infrequently unnecessary registrations that are considered to be wasteful, speeds up useful registrations, and promotes the learning effect. Eliminates unnecessary registrations, suppresses an increase in the index (the number of bits), and improves the compression ratio.

〔Example〕

(A) Description of the first embodiment FIG. 2 is a processing flowchart of the first embodiment of the present invention.

As described in FIGS. 16 and 17,
Convert to two-dimensional information and preprocess.

Next, a break (delimiter) of the converted data is recognized (detected), and the above-mentioned universal encoding is performed while the registration in the dictionary is selected, thereby performing the main processing.

 This processing is performed as follows.

The same character (data) sequence as the input sentence next (data) sequence is searched in the dictionary (dictionary search step).

Next, the index of the character (data) series that has been searched and matched in the dictionary is encoded (index encoding step).

Recognize a break in a character (data) string and register it in a dictionary (dictionary registration step).

FIG. 3 is a flowchart of the encoding process according to the first embodiment of the present invention, FIG. 4 is an explanatory diagram of the operation, and FIG. 5 is a flowchart of the decoding process.

It is assumed that a processor (not shown) creates and executes a dictionary in a memory for both encoding and decoding.

 The encoding process will be described with reference to FIG.

The same steps as those shown in FIG. 14 are denoted by the same symbols.

S1) To initialize the dictionary, the first 256 characters (words) are registered, and the start address N of the dictionary is set to 256.

Next, type the first letter K, and substitutes the address in the dictionary to ω _1.

S2) Input the next character K and proceed to S4.

S4) a combination of a string ω ₁ and the letter K determine whether there is in the dictionary.

S5) If ω ₁ K exists in the dictionary, substitute ω ₁ K for the registered address.

S3) Next, it is checked whether the data is completed.
If not, return to S2.

Thus, the longest matching character string in the dictionary is searched.

If S7) data end, and outputs an address omega ₁ as a code, ends.

S8) At step S4, if ω ₁ K does not exist in the dictionary, it is determined whether or not ω ₁ and K are data breaks.

For example, if K is a horizontal mode code (10000000) and a pass mode code (00000001), or if K is a shift length code in the data configuration, it is determined that the data is a break.

S6) conventionally, if not break the data and outputs the address omega ₁ as a code, to register the omega ₁ K in the dictionary address N, increments the start address N to (N + 1), a dictionary address of K omega Substituted into ₁ and proceed to step S3.

S9) On the other hand, if the break in the data, and outputs the address omega ₁ as a code, without the registration in the dictionary, and assigns the dictionary address of K to omega _1, the process proceeds to step S3.

Thus, the step S8, S9 provided, if the data break, abort the dictionary registration, the data after the cut to K as omega ₁ is the subject of a new dictionary registration.

The compression (encoding) process will be specifically described with reference to the example of FIG.

Here, the input symbols will be described using the same symbols as in FIG. 18 described above, and the arrows indicate the separation positions.

First, since the first line is a combination of the horizontal mode and the black and white run length, the first line is separated by the horizontal mode (10000000) and the 2-byte white run length (RL W128).

Here, as for the black and white run length, as shown in FIG. 16 (B), a one-byte code at the beginning “0” or a two-byte code at the beginning “1” is assigned, and thus the beginning of the run length is “1” here. Therefore, it can be recognized as a 2-byte code.

The next horizontal mode and the 2-byte black run length are also collected as one data series, and serve as a delimiter.

Therefore, on the first line, the horizontal mode is output code "128", the first byte of the white run length is output as output code "128", and these are collectively registered as a registration number "256" in the dictionary. .

Next, the second byte of the white run length is output as an output code “1”, and the contents of the previous registered dictionary (10000000)
Is registered as a registration number "257" in the dictionary.

Next, since the second horizontal mode has a break before it, it is not registered in combination with the white run length, and the second horizontal mode and thereafter are newly registered and searched.

That is, the contents of the first byte of the horizontal mode and the black run length are searched first, and since this exists in the dictionary as the registration number "256", it is output as the output code "256". Then, the content of the second byte of the black run length is output as the output code “72”, and the previous dictionary content “25” is output.
The output code “72” is registered in the dictionary as the registration number “258”.

Similarly, since the second and subsequent lines are basically shifted, boundaries between the shifted lengths are found, and breaks are detected.

As shown in FIG. 16 (B), the first two bits are a 1-byte or 2-byte code of “00”, “01”, “10”, or “11”. Mode (0000
0000) and the path mode (00000001) are not assigned, and can be distinguished from these.

Therefore, similarly, the gap of the deviation can be detected as shown in FIG. 4, and a new dictionary search and registration are performed at the gap.

 Hereinafter, the same is performed for the third and subsequent lines.

In this way, a break is detected from the code format of the input symbol.

Therefore, in this example, dictionary search registration is performed for each section, and it is not necessary to register a combination of irrelevant data series, the learning effect is improved, and the increase in the number of bits of the index (registration number) is minimized. Can be suppressed.

 Next, the restoration process will be described with reference to FIG.

Note that the same steps as those shown in FIG. 15 are denoted by the same symbols.

S1) As in the case of the encoding, the first 256 characters are registered in the dictionary in advance, and decoding is started after the head address n of the dictionary is set to “256”.

First, the first code (reference number) is read and the current CO
Since DE is OLDcode and the first code corresponds to one of the reference numbers of one character already registered in the dictionary, a 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 FINchar for later exception processing.

S2) Next, proceed to S2, read the next code, and set INCO to CODE.
Set as de.

S4) Next, proceed to S4, and check whether or not the code CODE input in S2 is defined (registered) in the dictionary.

S5) Usually, since the input code word has been registered in the dictionary in the previous processing, the process proceeds to S5, where the character string code (ωK) corresponding to the code CODE is read from the dictionary.

S6) Character string K is temporarily stacked in S6, and reference number co
de (ω) is returned to S5 as a new CODE, and S5 and S6
Is repeated recursively until the reference number ω reaches one character.

S9) Finally, proceed to S9, output the character K, and set K to FINchar
After setting to LIFO (Last In)
Pop-up and output in Fast Out) format.

S10) Next, the restored character is parsed in the same manner as in the above-described encoding step S8 to check whether there is a data break. If there is a data break, step S12
If not, the process proceeds to step S11.

S11) If it is not a data break, the code OLD used last time
code and the first letter K of the restored string this time (OLDc
ode, K) is registered in the dictionary with a new reference number, the reference number (address) n is incremented to n + 1, INcode is set to OLDcode, and the process proceeds to step S3.

S12) If there is a break in the data, do not register the dictionary.
Is set to OLDcode, and the process proceeds to step S3.

S3) Check whether the data is completed.
Return to step 2 and end the decoding if done.

S8) If the code is not registered in S4 (it occurs when referring to the immediately preceding reference number in encoding), FINchar is output in S8, OLDcode is set to CODE, and code (OL
After returning Dcode, FINchar) to INcode, go to S5.

That is, steps S10, S11, S12 are different from the conventional example of FIG.
Is added, and if it is not a data break, a new dictionary is registered as before, and if it is a data break, the new dictionary is not registered, and the registered contents are separated by the data break.

(B) Description of the second embodiment FIG. 6 is a flowchart of an encoding process according to a second embodiment of the present invention.
FIG. 7 is an explanatory diagram of the operation, and FIG. 8 is a flowchart of the decoding process.

Note that the same steps as those described in FIGS. 3 to 5 are denoted by the same reference numerals, and description thereof will be omitted.

In this example, while the first embodiment performs dictionary registration using a combination of the previous output code and the current character (data) between the delimiters, the dictionary registration registers the first character (data) of the delimiter. From the current character to the current character, and the compression ratio is increased by the continuous registration.

 The encoding process will be described with reference to FIG.

S1) It is basically the same as step S1 in FIG.

However, type the first letter K, it, string to be used when ₁ and the coded character string ω to be used when registering ω
Substitute for ₂ .

Here, ω ₁ = ω ₂ .

S2, S4, S8) These are the same as steps S2, S4, S8 in FIG.

S6) S9 if not break the data is not a data delimiter th code corresponding to the code string omega ₂ code
(Ω ₂ ) and register the combination of ω ₁ K in the dictionary,
After incrementing the registered address, the address at which ω ₁ K is registered is set as a new ω ₁ (this enables continuous registration), and K is set as a new ω ₂ , and the process proceeds to S3.

S7) If it is a data break, the code code (ω ₂ )
, And K as new ω ₁ and ω ₂ ,
Proceed to S3.

 In this way, search and registration are separated at data breaks.

S5) If ω ₁ K is found in the dictionary in S4, substitute the address where ω ₁ K is registered into a new ω ₁ and the address where ω ₂ K is registered into a new ω ₂ , and then go to S3. move on.

S3, S7) It is the same as S3, S7 in FIG.

 The encoding process will be specifically described with reference to the example of FIG.

Here, the difference from the example of FIG. 4 is that the search and registration between data breaks are the same, but the registered contents all start from the head of the break. With the number “257”, the first horizontal mode to the second byte of the run length are registered.

This is because ω ₁ K is set to ω ₁ in step S6.

 Next, the decoding process will be described with reference to FIG.

S1, S2, S4, S5, S6, S8, S9, S10) Steps S1, S2,
Same as S4, S5, S6, S8, S9, S10.

S11) In S11, a character string represented as a set (OLDcode, K) of the code OLDcode used last time and the first character K of the character string restored this time is registered in the dictionary with a new reference number, and the reference number n
Is incremented to n + 1, and code (OLDcode, K) is OLDc
Proceed to S3 as ode.

Since OLDcode is updated to (OLDcode, K), continuous registration is possible during restoration.

S12, S3) These are the same as steps S12, S3 in FIG.

(C) Description of Third Embodiment FIG. 9 is a flowchart of an encoding process according to a third embodiment of the present invention.
FIG. 10 is an explanatory diagram of the operation, and FIG. 11 is a flowchart of the decoding process.

The same steps as those described in FIGS. 6 and 8 are denoted by the same reference numerals, and description thereof will be omitted.

In this example, there is no difference in that the data break is at the head, but in addition to the second embodiment, the continuous registration is extended beyond the break, the continuous registration length is lengthened well, and the compression ratio is increased. Is to be further improved.

 The encoding process will be described with reference to FIG.

S1, S2, S3, S4, S5, S6, S8) Steps S1, S2, S3, S in FIG.
Same as 4, S5, S6, S8.

S7) If a data break is detected in S8, the code co
When de (ω ₂ ) is output, ω ₁ K is registered in the dictionary.

 That is, even if there is a break, continuous registration is performed.

Then, the address N is incremented to N + 1, and K is substituted into ω ₁ and ω ₂ , and the process proceeds to step S3.

Since the K and ω _1, the extension is from the separator to the first character.

 The encoding process will be specifically described with reference to the example of FIG.

Here, the difference from the example of FIG. 7 is that continuous registration is not terminated even if there is a break.

For example, with registration number "258",
The continuous registration is not aborted, but is extended to the second horizontal mode, which is the first byte of the delimiter.

At the same time, continuous registration is performed with the horizontal mode at the top, as indicated by the registration number "259".

This is because ω ₁ K is registered in the dictionary in steps S6 and S7.

 Next, the decoding process will be described with reference to FIG.

S1, S2, S3, S4, S5, S6, S8, S9, S10, S11) S1, S2, S in FIG.
3, the same as S4, S5, S6, S8, S9, S10, S11.

S12) When a data break is detected in S10, a character string that represents a pair (OLDcode, K) of the code OLDcode used last time and the first character K of the character string restored this time is assigned a new reference number, as in S11. To register in the dictionary, increment the reference number n to n + 1, set INcode to OLDcode, and proceed to S3.

As a result, even if there is a break, continuous registration is possible, and since INcode is set to OLDcode, continuous registration is terminated at the first character after the break.

(D) Description of Fourth Embodiment FIG. 12 is a processing flowchart of the embodiment of FIG. 4 of the present invention.

In this embodiment, a break (break) code is inserted at a break position when image data is converted into two-dimensional information as compared with the embodiment of FIG. 2. In this processing, a break code is detected. , And performs the encoding described in the first to third embodiments.

By doing so, there is an advantage in that the parsing does not have to be performed for the detection of the break, and the encoding and decoding times can be reduced.

(E) Description of Other Embodiments In addition to the above-described embodiments, the present invention can be modified as follows.

Although image data is targeted, other data may be used.

Although image data is converted into two-dimensional information and encoded, original data may be encoded.

Although the present invention has been described with reference to the embodiments, the present invention can be variously modified in accordance with the gist of the present invention, and these are not excluded from the present invention.

〔The invention's effect〕

As described above, the present invention has the following effects.

Detects a break in the data series,
Since the input data after the delimiter is registered as a new data series, it is possible to prevent the registration of combination data with different data series in the dictionary, eliminate unnecessary and infrequently registered entries, and speed up useful registration. To promote learning effects.

It is possible to eliminate useless registration, suppress an increase in the number of index bits, and improve the compression ratio.

[Brief description of the drawings]

 FIG. 1 is a principle diagram of the present invention, FIG. 2 is a processing flowchart of the first embodiment of the present invention, FIG. 3 is a flowchart of an encoding processing of the first embodiment of the present invention, and FIG. FIG. 5 is an explanatory diagram of the operation of the first embodiment of the present invention, FIG. 5 is a flowchart of a decoding process of the first embodiment of the present invention, FIG. FIG. 8 is an explanatory diagram of the operation of the second embodiment of the present invention, FIG. 8 is a flowchart of a decoding process of the second embodiment of the present invention, FIG. 9 is a flowchart of an encoding process of the third embodiment of the present invention, FIG. FIG. 11 is an explanatory diagram of the operation of the third embodiment of the present invention. FIG. 11 is a flowchart of a decoding process according to the third embodiment of the present invention. FIG. 12 is a flowchart of a process of the fourth embodiment of the present invention. FIG. 18 to FIG. 18 are explanatory diagrams of the prior art.

──────────────────────────────────────────────────続 き Continuation of front page (72) Inventor Yasuhiko Nakano 1015 Uedanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture Inside Fujitsu Limited (56) References JP-A-60-116228 (JP, A) JP-A-61-242122 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H03M 7/40

Claims (2)

(57) [Claims] 1. A method according to claim 1, wherein the encoded data is divided into different sub-sequences, the sub-sequences are registered in a dictionary, a sub-sequence registered in the dictionary is searched by input data, and the input data is stored in the dictionary. In a data compression method for specifying and encoding a registered subsequence having a maximum length matching by a reference number and registering the dictionary in the dictionary, a delimiter of the input data is detected from the input data, And input data after the delimiter to be newly registered in the dictionary. 2. The data restoration method according to claim 1, wherein the encoded data is compared with a reference number of a subsequence registered in a dictionary, and the encoded data is restored to a subsequence having a matching reference number. A data restoration method characterized by detecting a data segment from a set subsequence and registering a new subsequence in the dictionary at the boundary.