KR950000543B1 - Korean-character code generator - Google Patents

Abstract

This generator employs Hangul code system which is adaptable to many functional characteristics such as information exchange, character display, information processing. Hangul character recognizer (1) recognizes input code as Hangul code and outputs it. Hangul character phase structure detector (2) detects phase structure referencing Hangul consonant/vowel code table (4). Hangul character geometric code generator (3) finally outputs character code.

Description

Hangul Character Code Generator

1 is a device configuration diagram of the present invention.

2 is a Hangul alphabet code table of the present invention.

3 is a diagram illustrating the internal information of a Hangul alphabet code.

4 is a data flow diagram in the apparatus of the present invention.

5 is a block diagram of the present invention applied to the DIS 10646.

6 is an area allocation diagram of the DIS 10616 in BMP in FIG.

The present invention relates to an apparatus for generating Korean character information as a code sequence of an internal expression of a computer and to a device for generating Korean character codes. More particularly, the structure of the Korean character system based on the structural characteristics analysis of Korean characters and input using the code system The present invention relates to an apparatus for generating Korean character codes with high reliability, generality and compatibility by recognizing Korean characters as Hangul characters and outputting the information type code representations in character form through a geometry code generator.

Korean character information can be represented by computer code as N byte type, combination type, complete type, etc., but these methods were developed only for functional demands at the time of appearance. .

In other words, the N-byte type aims to display Korean characters on large computers, and the combination type aims to enhance Korean characters on personal computers and improve the quality of display characters. Because of the importance, the application of each method to the various technical demands of the character code was limited.

An object of the present invention is to provide a high-efficiency Korean code system and code generation device that can be applied homogeneously to the functional characteristics of Korean character codes, that is, for information exchange, character display, and information processing. .

1 is a device configuration diagram of the present invention.

This device is a one-chip hardware consisting of a Hangul character recognizer 1, Hangul character phase structure detector 2, Hangul character geometric code generator 3, Hangul alphabet code table 4 as shown. The Hangul Character Recognizer 1 recognizes and outputs the Hangul character code as the Hangul Character Code. The output of the Hangul Character Recognizer 1 becomes the input of the Hangul Character Phase Structure Detector 2. The Hangul character phase structure detector 2 operates by detecting the phase structure of the character by referring to the Hangul alphabet code table 4 from the input character code, and the result is the input of the Hangul character geometry code generator 3.

The Hangul character geometry code generator 3 inputs the output from the Hangul character phase structure detector 2 and finally outputs the character code of erasing by referring to the information of the Hangul alphabet code table 4.

2 shows a code structure of the Hangul Jamo code table 4 and a method of expressing and using the Jamo information of the present invention. The size of the code table is configured as 256 byte cells and the cells are grouped again to form 8 groups. Construct space.

These eight group spaces correspond to the topological structure type of the Hangul characters, and each cell of each spatial group is assigned a Hangul character synthesis letter corresponding to the topological structure type of the character.

The Hangul Character Recognizer 1 in FIG. 1 detects whether a character is suitable for Hangul syllables by continuously inputting a Hangul Jamo (consonant, vowel), and if it is an input string that can be composed of Hangul characters. It is possible to generate an output signal 1 indicating that this is possible, and to send the character synthesis possible alphabet among the input sequences to the Hangul character phase structure detector 2 of FIG. The patterns of the characters that can be composed of Hangul characters in the Hangul alphabet input are as follows.

① Type 1 (consonants + vertical vowels: a, yes, etc.)

② Type 2 (consonants + horizontal vowels: high, yo, etc.)

③ Type 3 (consonants + vertical horizontal double vowels: right, back, etc.)

④ Type 4 (consonants + vertical horizontal + consonants: angles, falls, etc.)

⑤ Type 5 (consonants + horizontal vowels + consonants: music, green, etc.)

⑥ Type 6 (consonants + vertical horizontal double vowels + consonants: etc.)

The Hangul Character Recognizer 1 is a device that recognizes the above character pattern from the input letter, and its configuration is expressed as an automaton of a mealy type machine as follows.

In addition, the opomaton above is represented by a state table as follows.

The initial state of the Hangul Character Recognizer 1 is always at S ₀ , and in this state, the consonants , school boy, ㅆ) this is when to produce the letter corresponding to the Hangul syllable Contributions ohms to state S _1, and then wait for the output signal that the character configuration is not possible to zero input.

If the input is a vowel after waiting for the next input in state S ₁ , it moves to the next state S ₂ with the input letter stored in the input buffer and returns to S ₁ if the input is a consonant.

State S ₂ in the input consonant when character recognition replaces the detection signal in one state S ₃ to / 0, if it is input while waiting for the next input bar staying intact S ₂ state is larger likely to be diphthong waits for the next input. In the state of S ₃ , if the input is a consonant, the character recognition signal output is shifted to the next state S ₄ with the output of / 0. If the input is a vowel, one of Korean characters (consonant + vowel) can be synthesized in the input string from S ₁ to S ₂ , so the character recognition signal is set to / 1 to input S ₀ and S ₁ from the input buffer After sending to the phase structure detector 2 of FIG. 1, the state is transferred to S ₄ while leaving the remaining input letters, i.e., the input letters at S ₂ and S _3, in the input buffer. In state S ₂ , if the input letter is consonant, Korean characters can be composed as the input letter from S ₀ to S ₃ , so the character recognition signal is set to / 1, and the input sequence from S ₀ to S _{3 is} converted to After sending, transfer the next state to S ₁ . At this time, the input buffer contains an input from S ₄ , that is, one consonant. In S ₄ the input is a collection one case because the input series in the middle input line of the S ₀ to S ₄ to S ₂ in S ₃ enables the open-circuit Hangul character configuration recognize the input series of the S ₀ to S ₂ characters in Along with the detection signal / 1, it is sent to phase detector 2 and the next state is transferred to S ₂ The input buffer it is possible with 2 of the type that is consonant, vowel in S ₃ and S ₄ remain.

In the state table of the state transition diagram above, the input Del represents a Deliminater and indicates an input key value such as a comma. If the input is Del while S ₂ is input to the vowels and consonants in the S ₀ to S _1, because it is in the S ₂ space, making it possible that the Hangul consonants and vowels constructed and the character recognition detection signal / 1, S ₀ The input sequence of and S ₁ is sent to the phase structure detector 2. If S ₃ is Del, consonants + vowels + consonants are composed of letters, and if S ₄ is Del, consonants + vowels + consonants are composed of letters automatically, and the phase sequence detector 2 passes the input sequence up to that state. will be.

As a specific example, the configuration and function of the Hangul character recognizer 1, when '1' is input from the initial state S ₀ , 'b' is a consonant, so it is impossible to compose a syllable with a consonant Hanja. and it moves the next input standby state to S _1. If 'ㅏ' is input in S ₁ , 'ㅏ' is a vowel, so if you combine 'n' in S ₀ and 'ㅏ' in S ₁ , you can construct a Korean syllable as 'I'. However, until that state, we cannot determine whether there is an input to 'I' or 'day', so we need to look ahead, so the character recognition signal is set to 0 and the next state is displayed to see the next input. Move to S ₀ and wait for the next input.

If the input from S ₂ is the consonant 'ㄹ', the input sequence from S ₀ to S ₂ can recognize 'day' and generate it as Hangul characters, but it is still unknown whether it is 'day' or 'ㄴ ㅏㄹ' for 'country'. Nonderterministic state Subsequently, the character recognition signal cannot be set to 1, and it must go to the next state again and wait for input. The following states as S ₃ comes in two at S ₃ consonants 'b' can not discern whether this consonant to whether the 'Fly' or 'beat' in the state back to the S ₄ following conditions in order to determine the next input It will move and wait for the next input. If the vowel 'ㅏ' is input in S ₃ , the character recognition signal is set to 1 and S ₀ and S because 'b ㅏ' can be composed of 'B' in the input sequence 'B ㅏ ㄹ ㅏ'. 'send to the phase detector structure 2, at the input of S ₂ and S _{3' 1} input alphabet "b""of the trestle d" and "trestle" is to move the next type state while stored in the input buffer to S _2.

If the attempt to enter a "clear""Wh trestle d b b heh 'input sequence _{S 0 → S 1 → S 2} → S 3 → and state transitions to S ₄ in, when in S ₄ b is input from the S ₀ The input sequence 'ㅁ ㅏ ㄹ ㄱ' up to S ₄ and the character recognition signal "1" are transferred to the phase structure detector 2, and the next input is transferred to S ₁ without leaving the 'buffer' in the input buffer. I will wait.

Of course, if you enter the same break key of space or comma in S ₂ , S ₃ , and S ₄ , then, in real time, `` me '' and `` sunny '' will be recognized as the Hangul character generation unit and sent to Hangul character phase structure detector 2. It is processed according to the procedure. Of course, the input code can use the combination type as it is, the compatibility of the input code is maintained.

Also, the Hangul character phase structure detector 2 extracts the topological structure of the Hangul characters by referring to the Hangul alphabet code table 4 using the output from the Hangul character recognizer 1 as an input.

Here, the structure refers to the relationship existing between individual elements in a set. In the structure of Hangul characters, there are order structure, topology structure, and algebraic structure. As a concept, for example, (downstream) = (a) is topologically the same, except that the geometry of one-dimensional and two-dimensional arrays is different. The number of characters that can be generated by synthesizing the Hangul alphabet is 11172 characters. If a common structure is found in these sets, the information topology can be extracted by expressing them in the meaning of the classified characters.

When extracting a topological structure from a letter of two-dimensional notation, the type of vowel included in one letter and the presence or absence of the final consonant become semantic information for determining the topological structure. That is, the vowels are classified into horizontal vowels (ㅗ, ㅛ, TT, ㅠ, ㅡ ........) and vertical vowels (ㅏ, ㅑ, ㅓ, ㅕ, ...., etc.) and neutron consonants Since the geometric data of each letter is different again, the letters are classified into small and large again, and the types of topologies are 6 types.The representative types are the first type (A), the second type (I), and the second type. Type 3, type 4 is (Han), type 5 is (Writ), and type 6 is (Qué). Here, the letters in parentheses are merely type predictive characters.

The Korean character phase structure detector 2 performs a function of extracting and outputting the phase information. The function operation is as follows.

The output code of Hangul Character Recognizer 1 is only the character boundary information corresponding to a substance, depending on whether it can be synthesized as Hangul characters. It will be the same as when writing. Therefore, when the output code of the Hangul Character Recognizer 1 is collected in two dimensions, that is, the shape of the Hangul alphabet in the character is determined according to the character type, first, which phase structure the target character belongs to? The character phase structure detector 2 detects the information. The information used to detect the phase structure is obtained according to the type of vowels (vertical, horizontal, and double) included in one character.

For example, if (a), the vowel is a vertical vowel (ㅏ) and thus belongs to the phase structure of the first region, so that a bit string (000) representing the first region is set and output.

The Hangul character geometry code generator 3 inputs the output information of the Hangul character phase structure detector 2 and outputs the final character code using the Hangul alphabet code table 4.

The structure of the Korean geometric code generator 3 is as follows. The code information processed and output by the Hangul character phase structure detector 2 is b ₅ b of FIG. 2 in the one-byte information of the letter of the Jamo constituting the character according to the Hangul character type (one of type 1 to type 6). _{After 6} b ₄ digit bits are set to the character phase code, the remaining bit information of b ₀ b ₁ b ₂ b ₃ b ₄ retains the code information when it is input to the character recognizer. The input data consisting of the alphabet codes is input to the Korean geometric code generator 3.

In the Hangul geometric code generator 3, bits ₀ b ₁ b ₂ b ₃ b ₄ of the one letter of each Hangul alphabet correspond to each alphabet code of each group defined in the Hangul alphabet code table of FIG. Arithmetic processing is performed. Bit operations use logical operators or mask bits between each bit.

This operation is described graphically as follows.

The bit information of b ₅ b ₆ b ₇ is determined according to the type of characters in the phase structure detector, and the 5-bit information from b ₀ to b ₄ is the letter of each group (military type) defined in the alphabet code table. A bit operation is performed so that a value from b ₀ to b ₇ is set among the code, i.e., bit values in the code table of FIG. For example, the code information of the phase structure detector of 'I' is treated as ???? 000 (b), ???? 000 (ㅏ) because 'I' is the first type of character, and then the geometry The bit information of ???? (b) and ????? (ㅏ) input to the code generator is 01000 (b) and 11000 as defined in 'b' and 'ㅏ' of the first group of character code tables. (I) After each Baht operation is set, the final output is 0000000 (geometric data) 000 (phase data), 11000 (geometric data), 000 (phase data) and 0100000011000000 (b). This is the address of this Hangul fonter table. It is then stamped on a display or printer by a font generator.

2 shows the details of the Korean alphabet code table 4, which is the core of the function of the Hangul character phase structure detector 2 and the Hangul character geometric code generator 3 in the present invention, and the design principle of the table structure and the code expression of the Hangul character information. Is as follows.

In the internal structure of this table, the physical code area is composed of 2 ⁸ = 258 byte cell memory space and address space. This address space corresponds to the actual code value, which in turn contains information representing the area space and the area space, corresponding to the topology space and the geometry space, respectively.

In Fig. 2, vertically in columns 1 and 2, type 1, type 2 in columns 3 and 4, type 3 in columns 5 and 6, type 4 in columns 7 and 8, and type 5 in columns 9 and 10 And 6, type 6 are defined in columns 11 and 12, and 4 and 6 type consonants are defined in rows 13 and 14, and 5 type type consonants are defined and assigned in columns 15 and 16, respectively. The areas of the grouping result are named first, second, third, fourth, fifth, six, seven, eight groups, respectively.

The letter in each group has control information of the display function. That is, all consonants of the third group's allocators in the first group and allocators of the sixth group in the fourth group are designated as non-spacing characters.

FIG. 3 shows an output code format corresponding to one letter in the Hangul alphabet code table 4 of the Hangul character phase detector 2 of FIG. Bits b0 to b4 are numerical information (Hangul character geometry code) for each Jamo classification, and bit information from b5 to b7 is phase structure information (Hangul character phase structure code), and Korean characters are identified by the letter codes from b0 to b7. Is identified.

Looking at the operation function of the Hangul character geometry code generator 3, the output code of the Hangul character phase structure detector 2 contains information for determining which character's topological structure belongs to the input character. It should detect the geometric data and generate a proper code. Since the output code of Hangul Character Phase Structure Detector 2 is connected to the phase group of Hangul Code Table 4, the Hangul Character Phase Among the output codes of the structure detector 2, the mapping function from b0 to b4 is mapped to the physical data of the Korean alphabet code table 4 to obtain a predetermined output code. This mapping function is a simple bit operation. It is calculated as (bit operation), so you get the final output.

FIG. 4 is a data flow diagram showing the actual functional operation of the apparatus configuration shown in FIG. When the input code of Jamo corresponding to the device of Figure 1 (ㅎㅜㄴ, ㅁ ㅣ ㄴ, ㅁ ㅣ ㄴ, ㅈ ㅓㅇ, ㅇ ㅡ ㅁ) has been entered, the Hangul Character Recognizer 1 has the function operation (ㅎㅜㄴ, ㅁ ㅣ B, ㅁ ㅣ ㄴ, ㅈ ㅓㅇ, ㅇ ㅡ ㅁ) will output a code of information with a text boundary as a result of the function. If (ㅎㅎㄴ) is applied to the Hangul Character Phase Structure Detector 2 of FIG. 1, as in the data flow diagram of FIG. 3, the (ㅎㅎㄴ) has a consonant consonant and its vowels are horizontal vowels. The bit pattern of (111) is set in the upper 3 bits of the final consonant, and then output.

The Hangul character geometric code generator 3 of FIG. 1 inputs this bit pattern as an input to the code table corresponding phase region of the Hangul alphabet code table 4, and then converts the 5 bits of the surplus non-self identification code into the Hangul alphabet code table 4 A predetermined final character output code is output by using a mapping function that converts to a Jamoexpression of. If the input code is enclosed in the complete KSC 5636, the character code is generated as a three-byte long character code with a bit pattern of (01001001 10101001 01000111). An example of the whole is (ㅎㅎㄴ) = (01111001 11001001 00100111) = (01001001 10101001 01000111) and the bit order is the rightmost bit of the byte.

The present invention is based on the premise that the Korean input code uses the seven-character Hangul date code (KS C5636). Therefore, the Hangul alphabet input code to the Hangul character recognizer 1 of FIG. 1 is generated by a key in of the keyboard, and the generated code is a 7 unit Hangul date code defined in KS C5636.

This code data is input to the Hangul Character Recognizer 1, and the Hangul Character Recognizer 1 recognizes and detects only the Jamo sequences that can be combined into one character in the Jamo input sequence.The output code of the Hangul Character Recognizer is recognized as a character. The code information of the number of letters only keeps the code at the time of input, that is, the KS C 5636 code value. When the input / output of continuous input / output of Jamo input system is processed by Hangul character recognizer, it passes through the character input sequence information having a character boundary. When such a character sequence is input to the Hangul character phase detector, the 7 unit input code is the Hangul Jamo code of FIG. You will be replaced with the 8-unit code system defined in the table. In this code conversion, even if the KS C5636 code is a 7-unit code, it does not use information of the MSB, but the 8-bit coding length is not different. However, since the bit operation of the bit information of b b b as the bit value of FIG. 2 is performed by the character phase detector, the Hangul Jamo code table 4 of FIG. 2 is started by the character phase detector 2. The output code from Hangul character phase detector 2 has the effect of using all the information up to b, i.e., MSB, in one byte, and this 8-unit code is input to the Hangul geometry code generator 3 to match the remaining 5 bits of information. This final code matches the address of the font table for displaying Korean characters.

The above process is represented graphically as follows.

Therefore, the Hangul Jamocode Table of FIG. 2 is commonly used in the process of Hangul character phase detector 2 and the process of Hangul geometric code generator 3. As a specific example, the process connection process between each component will be described as follows.

If 'B ㅏ ㄹ ㅏ (00100000, 01000110, 10010010, 01000110: Code value of KS C 5636)' is inputted into Hangul character recognizer, it is recognized as 'Nara (0010000001000110, 1001001001000110)' as two characters The code information is converted into code information having a boundary of 2 bytes each. Since the code does not include display information for collecting and displaying the code, it is input to the phase detector, which is the first process to add the display information. After the process of the phase detector is finished, the code is converted to 'country (0010000001000000, 10010000001000000)' as the first type and then output. This code information has only the topological information of type 1 and does not have any information that 'b' or 'ㅏ' should have any geometric data. Therefore, it must be processed by the Korean geometric code generator. When the Hangeul geometric code generation process is performed, the code is converted into 'Nara (0100000011001000, 1010000011001000)', and this code is followed by the address of the Hangul alphabet font, and the Hangul alphabet font stored in the address is displayed or printed. will be.

5 shows one embodiment as an information exchange code of the present invention. The configuration of FIG. 5 is composed of the apparatus 5 of the present invention shown in FIG. 1, its output code C1, the basic multilingnal plan (BMP: Basic Multilingnal Plan) of DIS 10646, and the output code C2 of BMP.

If the memory space and address space of the Hangul alphabet table of the present invention are mapped to the BMP of FIG. 5 as it is, the efficiency is high in code expression and usage.

6 shows a case where the code table of the present invention is arranged in the DIS 10646 BMP.

The BMP of DIS 10646 is represented by two-byte Chinese character modalities with a logical and physical space size of 256 × 256 bytes. Therefore, when the Hangul Jamo code table of the present invention is arranged on one line of BMP, the one-byte Jamo representation is represented by two bytes, but one-byte Hangul alphabet is represented as one-byte output code regardless of BMP size. The control method is as follows.

As in the 6th method of information exchange code method, when 1 row of A area of BMP is allocated and defined as a set of Korean alphabets, the initial value of the designation table is set to the code of the first byte. It can be used as a control code to link to the phase area in a row, and then the remaining 5 bits can be used as the output code using the address space of the second byte corresponding to the Hangul alphabet area as it is. However, communication will be possible with one byte per Hangul alphabet.

As described above, the members of the Hangul character generator of the present invention harmonize well with the structural characteristics of the Hangul character structure, and the characteristics of the present invention as code representation power and information exchange code of a character are as follows.

In the communication environment of the character information exchange, the 7-bit system is the basic, and since the 8-bit system is extended, the character code system that maintains up-down compatibility in the 7, 8-bit system is most efficient and desirable. The code representation of the device is compatible with the 7 and 8 bit systems. In addition, no limitation is imposed on the combination type, completion type or other input code or characteristics of the input device such as double type, triple type keyboard type or output code.

On the other hand, the character code extension method includes the international standard of ISO-2022 and DIS 10646, which is currently in development. The difference between the two methods lies in the character assignment area and method of each country in the world for the code area.

The allocation of the character code area of ISO-2022 is divided into 97 (7 bits) and 96 (8 bits) in the single byte system in accordance with the characteristics of each country character type and number in the 256 × 256 code area. There are 94x94 and 96x96 (2 byte systems), but 94x94 is assigned to the D area of ISO-2022 as the Hangul kanji area.

Since the size of the code area of the present invention is 256 bytes, if the area of the Korean code table of the present invention is taken as 96 × 3 in the C area of the ISO-2022 code, the address of the Hangul code table becomes an absolute address, and each code corresponding thereto is The relative address of the area is the replacement code according to the standard of ISO-2022. Of course, the present invention code system, which is a single-band system, is calculated to increase to 2 bytes per kanji by the 2-byte system of ISO-2022.

On the other hand, when DIS 10646 is applied, if one line of area A is allocated among 256 × 256 of BMP of DIS and the code of the present invention is mapped as it is, the size of one line can be allocated. Therefore, the ISO-2022 method is also present. It is an efficient way to control code tables.

As described above, the present invention can be applied to ISO-2022 and DIS 10646 without changing device functions and code expressions.

Compared to the prior art, the apparatus of the present invention is characterized by the size of the output code, code expression power, information expression used in code generation, and functional operation of the device configuration. In the present invention, the Hangul character code expression is composed of two bytes or three bytes of variable length, but the Hangul characters are characters + character, character + character, character + character. Since the structural correspondence between the character element and the code element is maintained, the processing efficiency is high in the code application technology.

Combination type code generation system uses combinatorial code as internal representation of character combination, but it is not structural system technology compared to this method, and it is excellent in terms of function separation and extensibility of device configuration.

As a result, the code system according to the present invention can guarantee compatibility with existing code systems, convenience of character display function, applicability to international standard information communication system, and correspondence of processing technology according to new character code technique. .

Claims (2)

Hangul Character Recognizer that inputs Hangul characters and recognizes them as Hangul character codes by inputting Hangul characters and outputs the information type code representation in character form. , The Korean character phase structure detector 2 which extracts the phase structure of the character by referring to the Korean alphabet code table 4 from the output code representation, the Korean alphabet code table 4 which simultaneously represents the phase information and the geometric information, and the desired character code Hangul character code generator, characterized in that consisting of the geometric character generator 3 to output. The Hangul Jamo code table 4, which simultaneously represents the geometric information of the topological information, has a spacing and a non-spacing information representation as the identification, discrimination information, and display control functions of the Hangul alphabet. Hangul character code generator.