JP2825960B2 - Data compression method and decompression method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] One of a plurality of dictionaries is designated by an index based on a dependency relationship (history) with the last character of an immediately preceding encoded character string,
Specify the input character string with the reference number of the substring that matches the maximum length among the already encoded substrings registered in the specified dictionary and LZ
Regarding the data compression method and decompression method that encodes to W code and decompresses character strings from LZW code, the purpose is to simplify initial registration of multiple dictionaries and improve encoding efficiency. A dictionary group consisting of a predetermined number of dictionaries less than the number of dictionaries is provided, and each dictionary is initially registered with a reference number assigned to each character type in units of one character.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data compression method and a decompression method using an LZW code known as an improvement of an incremental decomposition type, which is a kind of universal code.

In recent years, various types of data such as character codes, vector information, and images have been handled by computers, and the amount of data handled has rapidly increased. When dealing with large amounts of data, you can reduce the storage capacity by compressing the amount of data by eliminating redundant parts of the data,
It can be transmitted quickly.

As described above, universal coding has been proposed as a method capable of compressing various data by one method.

Here, the field of the present invention is not limited to character code compression, and can be applied to various types of data. In the following, one word unit of data is called a character, following the name used in information theory, Data in which arbitrary words are connected is called a character string.

As a typical method of the universal code, there is a Ziv-Lempel code (for details, for example, Munakata "Ziv-Lempel Data Compression Method", Information Processing, Vol. 26,
No. 1, 1985).

 For Ziv-Lempel codes, two algorithms, a universal type and an incremental parsing type, have been proposed.

Furthermore, as an improvement of the universal algorithm, LZ
There is an SS code (TCBell, “Better OPM / L Text Compres
sion ", IEEE Trans.on Commun., Vol.COM-34, No.12, Dec.
1986).

In addition, as an improvement of the incremental decomposition algorithm, LZW
(Lempel-Ziv-Welch) code (TAWelch, "A Te
chnique for High-Performance Data Compression ", Co
mputer, June 1984).

Among these codes, the LZW code is used for file compression of a storage device or the like because of its high speed processing and the simplicity of the algorithm.

[Prior Art] A flowchart of a conventional LZW code encoding algorithm is shown in FIG. 5, and a flowchart of a decoding algorithm is shown in FIG.

First, LZW encoding has a rewritable dictionary, divides the input character string into different character strings, assigns numbers to the character strings in the order in which they appear, registers them in the dictionary, and registers the currently input character string. The sequence is represented and encoded only by the reference number of the longest matching character string registered in the dictionary.

FIG. 7 shows a specific example of LZW encoding, and FIG. 9 shows LZW encoding.
FIG. 8 shows a specific example of decoding, and FIG. 8 shows the contents of a dictionary used in encoding and decoding. FIGS. 7, 8, and 9 illustrate a case where a character string composed of a combination of three characters of abc is compressed and decompressed for simplicity of description.

First, the LZW decoding process in FIG. 5 will be described as follows.

Step S1 (hereinafter, “step” is omitted): A character string consisting of one character for all characters is registered as an initial value before encoding starts. The encoding of S1 is performed by searching the dictionary using the first character K input to obtain a reference number ω,
This is a prefix string.

S2: Read the next character K of the input data.

S3: Check whether character input is completed.

S4: A search is performed to determine whether or not (ωK) obtained by adding the character K read in S2 to the initial character string ω obtained in S1 is in the dictionary.

S5: If the character string (ωK) is in the dictionary in S4, the character string (ωK) is replaced with the reference number ω in S5, and the process returns to S2 again until the character string (ωK) cannot be searched from the dictionary. Continue searching for.

S6: If the character string (ωK) is not found in the dictionary in S4, the process proceeds to S6, where the reference number ω of the character K obtained in S1 is a code word code.
(Ω), add a new reference number to the character string (ωK), register it in the dictionary, replace the input character K of S2 with the reference number ω, and increment the dictionary address n to S2 Return and read the next character K.

This will be specifically described with reference to FIGS.

First, the input data input of FIG. 7 is read from left to right.
When the first character a is entered, there is no matching character string in addition to the character a in the dictionary, so OUTPUT CODE 1 (reference number ω)
Is output as a code word. Then, the character a is set to the initial character string ω.

Next, assuming that the second character b is input, since the extended character string ωK = ab obtained by adding the input character to the initial character string ω is not in the dictionary, the OUTPUT CODE 2 of the character b is output as a code word. . Then, a reference number 4 is added to the extended character string ωK = ab and registered in the dictionary. The actual dictionary registration is registered as a character string 1b as shown on the right side of FIG. And the letter b
Becomes the initial character string ω.

Subsequently, if the third character a is input, the extended character string ωK = ba = 2a obtained by adding the initial character string ω to the character a is not in the dictionary, so that OUTPIT CODE 2 of the character b is output as a code word. After that, the extended character string ωK = ba is represented by 2a, and is added to the reference number 5 and registered in the dictionary. Then, the character "a" becomes the initial character string ω.

For the fourth input character b, the expanded character string ωK = ab is
Since the code word 4 of 1b is already registered in the dictionary, the character string ωK is set as a new initial character string ω, and the fifth character c is input to create an extended character string ωK = 4c = abc. Since this extended character string ωK = abc is not registered in the dictionary, OUTPUT CODE 4 of the character string ab = 1b is output as a code word, and the extended character string ωK = abc is written in the dictionary as code word 6 in the form of 4c. Register, and so on.

Next, the decoding process will be described with reference to FIG. In this decoding, similarly to the decoding, a character string consisting of one character for every character is registered in the dictionary as an initial value before decoding starts.

S1: The first code CODE is read and the reference number ω is decoded.
The current reference number ω is OLDω, and since the first code corresponds to one of the reference numbers of one character already registered in the dictionary, a character D (K) that matches the input reference number ω is searched for and the character K is output. I do. Note that the output character K is set in FINchar for later exception processing.

S2: Read the next code CODE.

S3: It is checked whether or not there is a new code, that is, whether or not code input has been completed.

S4: The reference number ω is decoded from the read code CODE, and INω
Set as

S5: It is checked whether or not the code CODE input in S4 is registered in the dictionary (ω ≧ n).

S6: Normally, the input codeword has been registered in the dictionary in the previous process, so the process proceeds to S6, where the character string D (ω′K) corresponding to the reference number ω is read from the dictionary.

S7: The character string K is temporarily stacked, the reference number ω ′ is set as a new ω, and the process returns to S6 again. The procedure of S6 is recursively repeated until the reference number ω reaches one character.

LILO (Last In Fast Out) for characters stacked in S8: S7
Output in pop-up format. At the same time, the reference number OLDω used last time and the first letter K of the character string restored this time
Is replaced with a string (OLDω, K) with a new reference number n
And register it in the dictionary.

This LZW decoding processing will be specifically described with reference to FIG.

First, the first input code is 1, and one character a, b, c is already registered in the dictionary as reference numbers 1, 2, and 3 as shown in Table 1. The output is replaced with the character string a having the same reference number. Similarly, the next code 2 is replaced with the character b and output. At this time, a new reference number 4 is added to (1b), which is a combination of the previously processed code and the first character b decoded this time, and registered in the dictionary.

The third code 4 replaces 1b with ab by searching the dictionary and outputs a character string ab. At the same time, a character string 2a (= ba), which is a combination of the previously processed code 2 and the first character a of the currently decoded character string, is added to the new reference number 5 and registered in the dictionary.

 Hereinafter, similarly, this processing is repeated.

 In the decoding of FIG. 9, there is the following exception processing.

This exception processing occurs when the sixth input code 8 is decoded. The code 8 is not defined in the dictionary at the time of decoding and cannot be decoded. In this case, a character string 5b is obtained by adding the first character b of the previously processed character string ba to the previously processed code 5, and further replaced with 2ab and bab and output. Then, a reference number 8 is added to a 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 S5 and S9 in the decoding processing flow of FIG. 6, and finally, in S8, a character string is output and registered in a dictionary in which a reference number is added to a new character example.

Note that the encoding / decoding processing in FIGS. 4 and 5 is performed while creating the same dictionary.

[Problems to be Solved by the Invention] As described above, in the conventional LZW code, when the input character string is divided into different character strings and encoded, each character string currently being encoded is different from the previous character string. Has the form of encoding as appearing independently.

This method is not problematic for encoding a memoryless source. However, many data, such as actual sentences, are regarded as storage information sources, and the history of occurrence of character strings cannot be used sufficiently with conventional LZW codes. There was a disadvantage that redundancy remained.

In order to solve such a drawback, the present inventors reduced the redundancy between character strings by taking the dependency relationship between the encoded character string and the last character of the immediately preceding character string, that is, the history into the dictionary, and reducing the compression between the character strings. It proposes a data compression and decompression scheme that increases the rate.

Specifically, as shown in FIG. 10, the dictionary 10 is divided into a plurality of dictionaries 10-1, 10-2, 10-3, and 10-4 and indexed. A specific dictionary 10-2 is selected using the last character b as an index. In each of the dictionaries 10-1 to 10-4, only a character string that follows the index character is registered.

According to this method, conventionally, a character string in a dictionary is designated by a reference number viewed from the whole, but since it can be designated by a reference number of only a series leading to an index, a small reference number can be used, and an LZW code can be used. Can be expressed short to improve coding efficiency.

However, in this method, the method of setting the initial value is problematic as follows.

In the LZW code, when data is handled in byte units, the code words are registered in a dictionary in advance as initial values of all 256 character types in order to simplify the code words and represent only the reference numbers. When this initial registration is applied to the method in Fig. 10, 256 dictionaries are indexed using all 256 character types, and the initial registration is performed for each dictionary. It is necessary to keep. Actually, since many of the 64K characters are not used, in the method of registering the initial value as it is, the reference number is increased by the unused characters, and the coding efficiency is reduced.

Therefore, if the initial value is not registered in the dictionary in advance, but is registered in the dictionary when a character string consisting of one character newly appears,
The inefficiency of registering the initial value can be solved. However, if this method is adopted, not all codewords can be represented by reference numbers. A character string in which the codeword is composed of one character encodes raw data, and a character string of two or more characters is represented by a reference number code. And the algorithm becomes complicated.

The present invention has been made in view of such a problem, and designates one of a plurality of dictionaries using an index based on a dependency relationship with the last character of the immediately preceding character string that has been encoded. It is an object of the present invention to provide a data compression method and a decompression method in which the initial registration of a plurality of dictionaries is simplified by grouping the subordination relations to improve the coding efficiency.

[Means for Solving the Problems] FIG. 1 is an explanatory view of the principle of the present invention.

First, the present invention relates to a data compression method of encoding an input character string into an LZW code by designating the input character string by a reference number of a subsequence having the maximum length matching among already encoded subsequences registered in the dictionary 10.

In the present invention as such a data compression method,
The dictionary 10 is composed of a dictionary group consisting of a predetermined number of dictionaries 10-1 to 10-N smaller than the number of all character types to be processed, and each of the dictionaries 10-1 to 10-N has one character unit. With a reference number for the initial registration.

Then, at the time of encoding the input character string, the encoding is performed by designating a specific dictionary 10-i in the dictionary group according to the index information indicating the dependency relationship (history) with the previously encoded character string,
At the same time, if there is no input character string in the command dictionary 10-i, a character string obtained by adding the next character to the reference number of the previous encoded character string is registered with a new reference number. And

Here, at the time of encoding the input character string, a specific dictionary 10-1 in the dictionary group is designated according to the index information obtained from a part of the last character code of the character string that has just been encoded. More specifically, a specific dictionary 10-i in the dictionary group is specified in accordance with index information indicated by the upper bits of the last character code of the character string that has just been encoded.

On the other hand, when encoding an input character string, a specific dictionary 10-1 in the dictionary group is designated in accordance with index information obtained by referring to a look-up table using the last character code of a character string that has just been encoded. May be. Specifically, a specific dictionary 10-i in the dictionary group is designated according to the index information obtained by referring to the look-up table using the upper bits of the last character code of the character string that has just been encoded.

Also, the present invention provides an original character string from a code word encoded by designating an input character string by a reference number of a subsequence that matches the maximum length among already encoded subsequences registered in the dictionary 10. For the data restoration method to be restored, the dictionary 10 has a predetermined number of dictionaries 10-1 to 10-10 smaller than the number of all character types to be processed.
-N, and all the character types are initially registered in each of the dictionaries 10-1 to 10-N with reference numbers assigned in units of one character. Then, at the time of restoring the input codeword, a particular dictionary 10-1 in the dictionary group is designated and restored according to the index information indicating the dependency relationship with the previously restored character string. A character string obtained by adding the first character of the restored character string to the reference number of the restored character string is registered with a new reference number. Here, the specification of the specific dictionary 10-i used at the time of restoration is the same as that of the encoding.

[Operation] According to the data compression method and the decompression method of the present invention having such a configuration, the following operation is obtained.

First, the history indicating the subordinate relationship with the last character of the immediately preceding character string has 256 states as it is, but there is a bias in the appearance of characters, and there is a state where it does not appear in 256 cases. Therefore, the present invention reduces the number of dictionaries by merging and reducing the history of the final character to reduce the number of significant states, for example, 8 to 16 states.

Since the number of states in the history is small, the number of registrations as the initial value for each dictionary of all 256 character types is the number of histories, that is, the number of dictionaries x
The number is 256, so that no large waste is produced.

As a method of compiling the histories, for example, if the upper 4 bits of the last character of the encoded character string immediately before are taken, the histories are compiled into 16 states. As a method of compiling the histories, it is desirable to use a state in which the histories appear evenly in order to use the dictionary effectively. However, it is not necessary to use the bits of the raw data in the character, and a look-up table (LUT) that maps the last character of the immediately preceding encoded string to the state of the history is prepared according to the rough nature of the data. Thus, the history state of the immediately preceding character, that is, the index of the dictionary may be designated.

[Embodiment] Fig. 2 is an embodiment configuration diagram showing one embodiment of the present invention.

In FIG. 2, reference numeral 12 denotes a CPU as control means,
The program memory 14 and the data memory 26 are connected to the PU 12.

The program memory 14 has control software 16 and LZW
Maximum match length search software 18 that performs maximum match length search using codes, encoding software 2 that converts input character strings to LZW codes
0, decoding software 22 for restoring the code converted to the LZW code by the encoding software 20 to the original character string, and dictionary initial value creation software 24 for initially registering all character types to be processed, for example, 256 character types Is provided.

On the other hand, the data memory 26 has a data buffer 28 for storing a character string to be encoded or a code string to be decoded, and a data buffer 28 for sequentially creating and encoding LZW codes. The dictionary 10 is used while being used.

The dictionary 10 is, for example, classified by index information indicating a subordinate relationship consisting of the upper 4 bits of the final character code of the encoded character string.
It is composed of dictionaries 10-1 to 10-16. The index specification of the dictionary by the upper 4 bits of the final character code of the encoded character determination may be directly specified, but in the following description, the dictionary index is read by referring to the look-up table (LUT). An example is given below.

The outline of data compression and decompression of the present invention in the embodiment shown in FIG. 3 is as follows.

The CPU 12 activates dictionary initial value creation software 24 under the control of the control software 16, and performs dictionary initial value creation processing. More specifically, the dictionary initial value creation software 24 assigns a reference number to each of the 256 character types and registers them in each of the 16 dictionaries 10-1 to 10-6 constituting the dictionary.

The data buffer 28 of the data memory 26 temporarily stores data to be encoded for a plurality of characters of a fixed length from the outside, and transfers the characters one by one according to a request of the encoding software 20. Each time a character in the data buffer 28 becomes empty, a plurality of characters are fetched from outside similarly.

Next, the encoding algorithm of the present invention will be described with reference to the flowchart of FIG.

 In FIG. 3, first, the following processing is performed in S1.

All kinds of character strings consisting of one character are previously registered as initial values in N dictionaries Di (i = 1,..., N) selected as the last character of the immediately preceding character string. In the present invention, the total number N of dictionaries is as small as N = 16 for all 256 character types.

The total number of reference numbers each dictionary Di managed by n _1, upon initialization, n ₁ = sets (character type +1) to the dictionary number N pieces of n _1.

The history from the immediately preceding character string, that is, the upper 4 bits of the last character code of the immediately preceding character string is defined as PK, and PK = PK =
Set 0.

The first character is input K and converted into a reference number (initial character string) ω.

LU that associates with the history state from the last character K1 of the previous character string
Set T. However, since there is no previous character string at first,
K1 indicating the last character of the immediately preceding character string is set to K1 = 0, and the LUT is set so that the index PK obtained from the LUT when K1 = 0 is PK = 0.

When the processing of S1 is completed, encoding is performed according to the procedures of S4 to S7. The procedure of S4 to S7 is basically the same as the conventional one shown in FIG.

The difference is that in the conventional LZW encoding, there was only one dictionary, but in the present invention, the first character is S1 and the subsequent characters are the last character K1 of the encoded character string shown in S6. LUT
A specific dictionary D _PK is selected from a plurality of dictionaries based on the history state LUT (K1) = PK obtained by referring to the dictionary, and the maximum match length is checked by comparing with a character string registered in the selected dictionary D _PK for the string, the point to be registered in the dictionary D _PK selects the string ωK that extended character the maximum match length is different.

After the registration in the dictionary D _PK in S8, the counter n _PK for managing the reference number of the dictionary D _PK is incremented by one as n _PK = n _PK +1. Also, as described above, in order to select the dictionary of the next character string, a new history state PK
Is required.

Next, the decoding algorithm of the present invention will be described with reference to FIG.

Decoding is the reverse operation of encoding. First, the initialization of the dictionary shown in S100 is the same as the encoding. The steps of S1 to S8 are
This is basically the same as the conventional system shown in FIG.

The difference of the decoding of the present invention is that the input code CODE
After decoding the reference number ω in S4, a dictionary D _PK is selected using the history state PK obtained from the last character of the immediately preceding character string, and a character string corresponding to the reference number ω is selected from the selected dictionary D _PK. Ask.

The registration of a new character string in the dictionary is the same as in the case of LZW encoding, but is performed one tempo later than the time of encoding. Immediately, at the time of encoding, when the encoding of the target character string is completed, the character string ωK (the target character string + the next one character) extended by one character is registered in the dictionary. When the character string ω is extended by one character, the character string is registered in the dictionary together with the first character of the next character string. Therefore, the registration is performed when the restoration of the next character string is completed.

Specifically, as shown in S7, the reference number OL of the immediately preceding character string
The set of Dω and the first character K1 of the restored character string is registered in the dictionary D _PK1 selected in the history state PK1 from the last character of the immediately preceding character string. Therefore, the current history state PK is moved to PK1 in order to extend the restored character string and register it next, and a new history state is obtained from the last character K2 of the restored character string.

In the above-described embodiment, a case is described in which the dictionary is configured with 16 characters according to the history state for all 256 character types. The number of dictionaries may be used.

Also, the number of character types is appropriately determined as needed.

[Effects of the Invention] As described above, according to the present invention, initial values of a plurality of dictionaries can be registered simply and without waste according to the history state of a character string, and previously appeared for a character string being encoded. Since the history of character strings can also be adopted, redundancy between character strings is reduced, a higher compression ratio than conventional LZW codes can be obtained, and codes can be represented only by reference numbers, so that they can be executed with a simple algorithm.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention; FIG. 2 is a block diagram of an embodiment of the present invention; FIG. 3 is a flowchart of an encoding algorithm of the present invention; FIG. 5 is a flowchart of a conventional LZW encoding algorithm; FIG. 6 is a flowchart of a conventional LZW decoding algorithm; FIG. 7 is an explanatory diagram of a specific example of conventional LZW encoding; FIG. FIG. 9 is an explanatory diagram of a specific example of conventional LZW decoding; FIG. 10 is an explanatory diagram of encoding that has already been proposed by the inventors of the present invention in which subsequence decomposition and capture of a history between character strings are performed. is there. [Explanation of Codes] 10, 10-1 to 10-N; Dictionary 12: CPU 14: Program Memory 16: Control Software 18: Maximum Match Length Search Software 20: Encoding Software 22: Decoding Software 24: Create Dictionary Initial Value Software 26: Data memory 28: Data buffer

────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Hirotaka Chiba 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Inside Fujitsu Limited (56) References JP-A-60-116228 (JP, A) JP-A-3-262331 (JP, A) JP-A-3-270417 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) H03M 7/40

Claims (10)

(57) [Claims] 1. A data compression method for encoding an input character string by designating it by a reference number of a subsequence having a maximum length matching among already encoded subsequences registered in a dictionary (10). The dictionary (10) is composed of a dictionary group including a predetermined number of dictionaries (10-1 to 10-N) smaller than the number of all character types to be processed, and each dictionary (10-1 to 10-N) At least all character types are initially registered with reference numbers in units of one character, and at the time of encoding an input character string, are stored in the dictionary group according to index information indicating a subordination relationship with a previously encoded character string. When the specified dictionary (10-i) is specified and encoded, and there is no input character string in the specified dictionary (10-i),
A data compression method characterized in that a character string obtained by adding the next character to the reference number of a previously encoded character string is added with a new reference number and registered. 2. The data compression method according to claim 1, wherein at the time of encoding the input character string, the dictionary group is encoded according to index information obtained from a part of the last character code of the character string which has been encoded immediately before. A data compression method characterized by designating a specific dictionary (10-i). 3. The data compression method according to claim 2, wherein at the time of encoding the input character string, the dictionary group is encoded according to the index information indicated by the high-order bit of the last character code of the character string that has just been encoded. A data compression method characterized by designating a specific dictionary (10-i). 4. The data compression method according to claim 1, wherein at the time of encoding the input character string, index information obtained by referring to the look-up table using the last character code of the character string that has just been encoded. A data compression method characterized by designating a specific dictionary (10-i) in the dictionary group according to the following. 5. The data compression method according to claim 4, wherein at the time of encoding of the input character string, the dictionary group is encoded according to index information created from the most significant bit of the last character code of the character string that has just been encoded. A data compression method characterized by designating a specific dictionary (10-i) in the following. 6. A method according to claim 1, wherein an input character string is designated by a reference number of a subsequence having a maximum matching length among already encoded subsequences registered in a dictionary, and an original character is obtained from a codeword. In the data restoring method for restoring a column, the dictionary (10) is composed of a dictionary group composed of dictionaries (10-1 to 10-N) with a predetermined number smaller than the number of all character types to be processed. For each dictionary (10-1 to 10-N), at least all character types are initially registered with reference numbers in units of one character, and when the input codeword is restored, the subordination relation with the previously restored character string is determined. The specified dictionary (10-i) in the dictionary group is designated and restored according to the index information shown, and each time the restoration is performed, the reference number of the previously reduced character string is replaced with the first one of the currently restored character string. A data restoration method characterized by registering a character string with added characters with a new reference number . 7. A data restoration method according to claim 6, wherein at the time of restoring an input codeword, the data in the dictionary group is restored in accordance with index information obtained from a part of the last character code of the character string just restored. A data restoration method characterized by designating a specific dictionary (10-i). 8. The data restoring method according to claim 7, wherein at the time of restoring the input code word, the input code word is restored in the dictionary group according to the index information indicated by the upper bit of the last character code of the character string just restored. A data restoration method characterized by designating a specific dictionary (10-i). 9. The data restoring method according to claim 6, wherein at the time of restoring an input codeword, said input codeword is referred to in accordance with index information obtained by referring to a look-up table using the last character code of a character string just restored. A data restoration method characterized by designating a specific dictionary (10-i) in a dictionary group. 10. A data restoration method according to claim 9, wherein at the time of restoring an input codeword, an index obtained by referring to a look-up table using upper bits of the last character code of a character string just restored. A data restoration method characterized by designating a specific dictionary (10-i) in the dictionary group according to information.