JPH03110675A - Word dictionary retrieving device - Google Patents

Abstract

PURPOSE:To execute the collation with a word dictionary at a high speed by providing a first - an m-th candidate shift registers of an (n) character portion, respectively for executing a forward feed of one character each synchronizing with a shift clock, and a first character - an n-th character comparater. CONSTITUTION:The device is provided with an input device 5 for inputting a character-string in which (m) kinds of candidates exist, (s) kinds of work dictionary memories 1 having data width of an (n) character portion, a controller 7 for generating a clock, whenever (m) kinds of candidates to one character are inputted by the input device, and an address counter 4 being common to the word dictionary memory. Also, the device is provided with a first, a second,... an (m)-th candidate shift registers 2 of an (n) character portion, respectively, a switch 8 for selecting the word dictionary memory, a first character, a second character,... an (n)-th character comparator 3, and a deciding circuit 6 for deciding that a word in the word dictionary memory appears in an inputted character-string in the case a coincidence signal is detected from all the compar ing circuits. In such a way, even if each character has plural candiates, the collation with a word dictionary can be executed at a high speed.

Description

Detailed Description of the Invention (Industrial Field of Application) The present invention is a word dictionary that searches for a portion of the input string in which a word that exists in the word dictionary appears by comparing an input string with a word dictionary. The present invention relates to a dictionary search device.

However, in the input character string, each character may have multiple candidates or may have only one candidate. In addition, search keys that can be used in word dictionary searches are often the word notation (character strings containing kanji and kana) and the readings (kana character strings).

When each character in the input character string is unique without multiple candidates, the word dictionary search device is the part of the kana-kanji converter that performs a word dictionary search for the kana character string entered from the keyboard, machine translation equipment, and sentence-to-speech converters. It is used in the parts of devices and grammar proofing devices that perform word dictionary searches for character strings containing kanji and kana created using word processors. In the former case, the search key is read as Na.
In the latter case, the search key is a notation.

When there are multiple candidates for each character in the input character string, the word dictionary search device uses a voice recognition device, character recognition device, etc. to select the most likely character from the multiple candidate characters in the recognition result. It is used in word dictionary search parts, etc. In the case of voice recognition, the search key is the pronunciation (syllable), and in the case of character recognition, the search key is the notation.

(Prior art) The word dictionary search methods conventionally used in kana-kanji conversion devices, sentence-to-speech conversion devices, etc. are basically based on the documents:
[Kana-Kanji Conversion by Computer] (Japanese translation, Ebara, NHK
Technical Research, Vol. 25, No. 5, pp. 23-60, 1973
It follows the method shown in 2010). That is,
The process of cutting out a substring from the input string and searching the word dictionary for a word that matches the notation (in kana-kanji conversion, "kana notation" or "reading") in the substring. (hereinafter, this will be referred to as the first prior art).

In the first conventional technique, for example, when performing a word dictionary search for a character string containing kanji and kana, such as "analyzing a sentence," in order to search for words starting from each character position such as the first or second character, "Analyze a text""Analyze a chapter""Analyze""Analyze""Analyze"
Partial strings such as ``suru'' and ``ru'' are extracted separately, and for each partial string, a partial string with the end removed is also generated, and each of them is searched from the word dictionary. In other words! When searching for words that start with the first letter, search for substrings such as ``parse the sentence'', ``analyze the sentence'', ``analyze the sentence'', ``solve the sentence'', ``sent wo J'kyuouJ``bun''. , repeat the search one after another, 2
When searching for words starting with the letter ``parse chapter'' ``parse chapter'' ``parse chapter'' ``solve chapter''
Searches are repeated one after another for substrings such as ``Chapter'' and ``阜''.

As a result, a word that matches the notation of the underlined substring is found.

In this first conventional technique, as mentioned above, word dictionary searches are repeated for multiple substrings.
By devising the structure of the word dictionary, the predetermined search time for a single substring can be shortened.1
For example, you can sort the words in a word dictionary in advance and perform a binary search, or you can divide the words into characters such as the first and second characters, group the common parts together, and perform character-by-character matching using a tree structure. etc. Regarding the structure of such a dictionary and improvements to the search method, please refer to the literature: “The Ar
of Computer Programming
3: Sorting and Searching”
(mouth, F, n1Jth, AddiSOn-14es
leV.

(1973).

However, this first conventional technique is intended to be implemented as a sequential program on a computer, and the efforts to reduce the required time are within the framework of sequential processing. Therefore, the process of comparing a certain partial character string of an input character string with a character string representing a certain word in a word dictionary is performed character by character sequentially.

For example, the substring ``sentence'' and ``sentence'' in the word dictionary are
To compare the word ``sentence'', ``sentence'' and 1st sentence'' are matched, and when they match, ``chapter'' and 1st chapter'' are compared, and then,
1.
The matching process is repeated character by character.

In addition, in the first conventional technique, when there are m candidates for each character of the input character string, m' combinations of candidates are prepared in advance for a partial character string of length. need to be generated.

Then, for each of them, a partial character string with the end truncated as described above is generated and searched in the word dictionary.

On the other hand, when comparing a character string of a certain length with an input character string, it is considered that instead of repeatedly matching each character, it is possible to simultaneously match L sets of corresponding character positions ( (hereinafter referred to as the second conventional technique) At that time, if the input string is stored in a shift register, the position of the collation can be shifted instead of cutting out the partial string from the input string again. This can be achieved by shifting all the characters in the register one character at a time. This second prior art is described in JP-A No. 63-261421 "Character string processing device" and JP-A No. 63-261422 "Character string matching device". In addition, in this second conventional technique, there are no multiple candidates for the input character string, and the number of character strings to be searched is limited to one, and not as many as in a word dictionary. Publication “Verification method”
and literature: [Dictionary matching algorithm for high-speed language processing in spoken Japanese input systems] (Hamaguchi and Suzuki, Transactions of the Institute of Electronics, Information and Communication Engineers, Vol. J70-D, No. 8,
1589-1596, 1987), a third prior art is shown.

This third conventional technique assumes that an input character string has a plurality of character candidates. First, set the type of character to M (for example, in the JIS character code table, hiragana is M
= 83), one M-bit memory is prepared for each character position such as the first character and second character, and each bit corresponding to multiple candidate characters is set to 1. , for each word in the word dictionary, read the bit contents (1 or 0) of the corresponding character from the corresponding M-bit memory at the same time for each character position such as the first or second character of the word notation. If 1 is read from the M-bit memory of , it means that a word in the word dictionary has appeared in the input character string.

In the first to third conventional techniques described above, only one type of search key is provided when searching a word dictionary.

However, in a document processing system, when searching a word dictionary,
It may be necessary to set multiple search keys'0
For example, literature: [Japanese writing support system COM
"E'r - Focusing on an integrated method for text analysis applications" (Fukushima/Inuyama, Information Processing Society of Japan/Text Processing and Human Interface Study Group, 20-2, 1988), Kana-Kanji conversion, sentence-to-speech conversion, text A system is described that integrates applications such as calibration. In the above literature, there is a dictionary that uses the pronunciation of the word as a search key for kana-kanji conversion to obtain the notation of the word, and a dictionary that uses the pronunciation of the word as the search key for sentence-to-speech conversion and grammar proofing to obtain the pronunciation of the word. , two types of independent word dictionaries are used (hereinafter, this will be referred to as the fourth conventional technique), and the method of the first conventional technique is applied to each of them. are doing. Therefore, the pronunciation and notation of a word are 2.
This means that it is registered twice in both dictionaries.

In contrast, we have made it possible to search a single word dictionary regardless of the pronunciation or spelling of the word, as described in the fifth section below.
- This is the sixth conventional technology.

In the fifth prior art, the first method uses word notation as a search key.
A word dictionary based on conventional technology and a kanji reading table are used. The kanji reading table is a table that registers the reading of a single kanji character, and the notation of the word in the word dictionary is expanded into the kanji reading table.
Search while generating readings of words. Unexamined Japanese Patent Publication 1986-
This method is described in Japanese Patent Publication No. 212876 ``Kana-Kanji Mutual Conversion Device'' and Japanese Patent Application Laid-Open No. 62-224859 ``Japanese Processing System''.

In the sixth prior art, a pointer is used to perform a word dictionary search for two types of search keys to which the method of the first prior art can be applied. In other words, the duplicate content of the fourth prior art is replaced with a pointer. Patent Publication 1986-4225 r Electronic Dictionary”, No. 4
In the prior art, one of the two word dictionaries is entirely replaced with a set of pointers. Japanese Patent Publication No. 61-3074
In No. 12 "Word Dictionary Configuration System", the search keys of two word dictionaries in the fourth prior art are configured to refer to each other using pointers.

(Problems to be Solved by the Invention) First, the problems of the prior art related to the search key of l[I'M will be described.

The first prior art has the drawback that, as described above, word dictionary searches for a large number of partial character strings must be repeated. This drawback is particularly noticeable when there are multiple candidates for each character in the input string. If there are m candidates for each character in an input string of length K, let the length of the longest substring be the length of the longest substring (usually, L is the length of the longest word in the word dictionary). , it becomes necessary to repeat the search for a maximum of (m' x L x K) substrings.

Normally, conditions are determined in advance to avoid searching for unnecessary substrings, so (m'XLXK) is the worst case, but the number of searches is still quite large.

A second drawback of the first conventional technique is that a certain partial character string is compared with a character string representing a certain word in a word dictionary in order, character by character, so the comparison takes time.

The second prior art addresses the second drawback of the first prior art. However, the second conventional technique cannot be applied when there are multiple candidates in the input character string, and the number of character strings to be searched is limited to one, and comparisons are made with many character strings like in a word dictionary. Considering this, the difference in the length of each word in the word dictionary becomes a problem and cannot be applied.

The third prior art addresses two drawbacks of the first prior art. However, in order to store the input string, it is necessary to prepare a memory with the number of bits for each type of character and the length of the string, so if there are many types of characters, the memory size becomes quite large. There is a problem with it. In the case of kana-kanji conversion and voice recognition, the input character string is limited to about 100 types of hiragana (or phonetic characters), so this is not much of a problem, but text-to-speech conversion and machine translation that target character strings containing kanji and kana In such cases, there are 3,000 to 4,000 types of characters, including kanji, and the memory becomes large, which becomes a problem.

In addition, in the third conventional technology, only the word starting from the beginning of the input character string is searched, so when searching for a word starting from the second character, a word starting from the third character, etc., the input There is also the problem that it is necessary to register the character string in memory again.

Next, problems with the prior art when a plurality of search keys are provided will be described.

In the first to third conventional techniques, only one type of search key is considered. Therefore, if a plurality of search keys are to be provided, a word dictionary must be prepared for each search key, as in the fourth prior art. In that case, information will be registered redundantly in a plurality of word dictionaries, resulting in a wasteful fly4 system.

Although the fifth and sixth conventional techniques eliminate this wasteful duplication, they have the following problems. First, in the fifth conventional technology, in the case of a search using the pronunciation of a word as a search key,
In addition to performing complex processing such as collating while expanding the kanji reading table and generating headings, headings other than the original pronunciation of the word are also generated, resulting in a large amount of processing and, as a result, processing time. In the 6th prior art, which increases to zero order, relationships are managed using pointers, so when adding or deleting items (words), it is necessary to check whether the relationships are broken or not. The effort required to maintain the word dictionary increases. Furthermore, although the apparent duplication of information is eliminated by replacing it with a pointer, the pointer area is taken up instead, which is wasteful in terms of actual dictionary contents, as in the fourth prior art. many.

SUMMARY OF THE INVENTION An object of the present invention is to provide a word dictionary search device that eliminates the drawbacks of the prior art as described above and is capable of high-speed matching with a word dictionary even if each character in an input character string has multiple candidates. ,
Moreover, it is an object of the present invention to provide a word dictionary search device that can maintain high speed even when a plurality of search keys are provided without adopting a structure that is wasteful in terms of capacity.

(Means for Solving the Problems) In the word dictionary search device of the first invention of the present application, there are m types of candidates from the first to the mth candidate (m is an integer where m≧1) for each character. An input device for inputting a character string to be displayed, and a header stored in each address with a data width of n characters (n is an integer with n≧1), and a predetermined residual symbol for the part less than n characters. A word dictionary memory of S types (s is an integer such that S≧1) that is divided according to the type of heading, and once every time m types of candidates for one character are input by the input device. a controller that generates a shift clock, a judgment clock and a counter clock a number of times according to the total number of words in the word dictionary memory; a reset/set synchronized with the shift clock; and a power run-up synchronized with the counter clock. an address counter common to the S types of word dictionary memories, and a shift counter corresponding to the first, second, . . . mth candidates of the character string inputted by the input device;
1st, 2nd, . . . m-th candidate shift registers for each n-character bone, which sequentially advance characters one by one in synchronization with the tsuku;
A switch for selecting one of the S types of word dictionary memories, and the first and second characters of the n-character bone data read from the word dictionary memory selected by the switch.
...The character at the corresponding position corresponding to the nth character is the first and second character.
. . . 1st character, 2nd character, . When a match signal is detected from all of the first character, second character,... nth character comparison circuits in synchronization with the determination clock, the word dictionary is detected in the character string input by the input device. and a determination circuit that determines that a word existing in the memory has appeared.

The word dictionary search device of the second invention of the present application comprises an input device for inputting a character string in which there are m types of candidates from the first to the mth candidate (m is an integer where m≧1) for each character. , for n characters (n is n≧1, and one heading is stored in each address with a data width of 11 numbers, and the part less than n characters is filled with predetermined residual symbols. A word dictionary memory of S types (s is an integer such that S≧1) divided according to
an additional information memory that stores additional information other than M headings; a shift clock that is activated once every time m types of candidates for one character are input by the input device; and a shift clock corresponding to the total number of words in the word dictionary memory. A common address for a controller that generates a count determination clock and a counter clock, the S type word dictionary memory that performs reset synchronized with the shift clock, and count alarm 1 synchronized with the counter clock, and the additional information memory. a counter, and first, second, . . . m-th candidates of the character string input by the input device, and performs original forwarding of each n character in synchronization with the shift clock.・Second... m-th candidate shift register,
A switch for selecting one of the S types of word dictionary memories, and the first and second characters of the data for n characters read from the word dictionary memory selected by the switch.
...The character at the corresponding position corresponding to the nth character is the first and second character.
. . . 1st character, 2nd character . . . nth character ratio soft circuit that outputs a match signal when a match j7 occurs with any character at the same position in the m-th candidate shift register or with the above-mentioned residual symbol. When a match signal is detected from all of the first character, second character,... nth character comparison circuit in synchronization with the determination clock, the character string input by the input device is detected. A determination circuit that determines that a word existing in the word dictionary memory has appeared; and a determination circuit that determines that a word existing in the word dictionary memory has appeared, and a word whose appearance is detected by the determination circuit is stored in the word dictionary memory other than the word dictionary memory selected by the switch and in the additional information memory. and a selector that selects and outputs.

(Example) The structure and operation of the present invention will be explained using the drawings.

FIG. 1 is a block diagram showing the configuration of a first embodiment of a word dictionary search device according to the first invention of the present application.

Below, each component will be explained first.

The input device 5 is a device for inputting a character string in which there are m types of candidates for each character, from the first candidate to the m-th candidate (m is an integer where m≧1), and is, for example, a voice recognition device. and character recognition devices.

The input device 5 simultaneously outputs m candidates for each character, and transmits the timing of their appearance to the controller 7 using an input clock 50.

Word dictionary memory 1 is for n characters (n is an integer where n≧1)
One heading is stored in each address having a data width of , and the portion less than n characters is filled with predetermined residual symbols. This word dictionary memory 1
There are S types of different headings (s is an integer such that S≧1).

Hereinafter, the S types of word dictionary memories 1 will be referred to as a first word dictionary memory, a second word dictionary memory, a 1st word dictionary memory, . . . , an S-th word dictionary memory. Incidentally, in FIG. 1, s=2. The first word dictionary memory uses the readings of words as headings, and the second word dictionary memory uses the spellings of words as headings. Another possibility is to use the English notation of the word as a heading. Assume that for S types of word dictionary memories, headings for the same word are stored at the same address. However, in this first embodiment, only the appearance of words is determined, so 1. It works even if the same address does not necessarily correspond to the same word. Furthermore, the arrangement of words in the word dictionary memory does not need to be sorted.

FIG. 5 is a diagram showing an example of the contents of the word dictionary memory 1. FIG. 5(a) is an example of the contents of the first word dictionary memory, and FIG. 5(b) is an example of the contents of the second word dictionary memory.

In both cases, n=6.

Note that Δ represents the residual symbol. 1 in Figure 5(a)
If the address a of the word dictionary memory is ``Nihon ΔΔ
n (=6) characters "Δ" are read out at the same time. For the same address, n (=6) characters "Japan ΔΔΔΔ" are simultaneously read out from the second word dictionary memory shown in FIG. 5(b). The word dictionary memory 1 can be realized using an IC memory or the like.Normally, Japanese character codes are expressed in 16 bits, so the data in Figure 5 contains 16 bits.
X6=96 bits. In current IC memories, the data read out at the same time is about 8 bits, so each word dictionary memory in FIG.
It can be realized by arranging two pieces in parallel.

The controller 7 generates one shift clock and a number of determination clocks and counter clocks corresponding to the total number of words in the word dictionary memory 1 each time m types of candidates for one character are inputted by the input device 5. do. FIG. 6 is an example of a time chart of input/output signals of the controller 7.

In the time chart of FIG. 6, each time the input tarlock 50 is input from the input device 5, the shift clock 7
0 is outputted once. In this embodiment, the input clock 50 is output as is as the shift clock 70. Next, the judgment clock 2 and the counter clock 71 are alternately set to N.
It is outputting times.

However, the counter clock 71 may be counted (N-1) times, where N is the total number of words in the word dictionary memory 1.

A person skilled in the art can easily realize a controller 7 that operates according to such a time chart.

The address counter 4 is a counter that performs reset in synchronization with the shift clock 70 and count-up in synchronization with the counter clock 71, and outputs the counter value as a common address value for the S types of word dictionary memories 1. Address counter 4 can be implemented using a conventional counter IC.

The first candidate shift register 2 sequentially shifts the first candidate (I is an integer satisfying 1≦l≦m) of the character string inputted by the input device 5 one character at a time in synchronization with the shift clock 70, and transfers n characters. This is a shift register that stores minutes. This shift register 2 is provided for each of the m candidates for each character of the input character string, and includes a first candidate shift register, a second candidate shift register,...
, there are m number of m-th candidate shift registers. Figure 7 shows
2 is a diagram showing a configuration example of each shift register 2. FIG. When 16 characters are expressed by d bits, the first candidate shift register 2 can be realized by (dXn) D flips 70 blocks synchronized with the shift clock 70, as shown in FIG. That is, d pieces arranged in parallel are connected in series. The d pieces connected in parallel correspond to one character, and the output is sent to the comparator circuit 3 for each character.

The switch 8 is a means for selecting one of the S types of word dictionary memories 1, and outputs a value corresponding to the word dictionary memory 0 selected as a selection signal 80.

For example, in the case of s=2, it is sufficient to prepare one bit of the selection signal line, and when the value is O, it corresponds to the first word dictionary memory, and when the value is 1, it corresponds to the second word dictionary memory.Generally, ,
The selection signal line requires logs bits. This switch 8 can be realized by a toggle switch, dip switch, or the like. Note that it is also possible to configure the switch 8 such that it changes over in accordance with the character code of the text input from the input device 5.

The j-th character comparison circuit 3 reads out the word dictionary memory 1 selected by the switch 8 and compares the first address (j is 1≦j This circuit outputs a match signal when a character (integer ≦n) matches one of the first characters of m shift registers 2 or a residual symbol. The comparison circuits 3 are provided in a number corresponding to each of the data width n characters of the word dictionary memory 1, and include a first character comparison circuit, a second character comparison circuit, etc. There are n number of n-th character comparison circuits. FIG. 8 is a diagram showing an example of the configuration of the j-th character comparison circuit 3. When the character 1 is represented by d bits and m shift registers 2 are provided, the j-th character comparison circuit 3 has (m+1) d-bit comparators and 1 It can be configured with 2 OR gates and 1 d-bit selector. The d-bit selector selects one of the j-th character data of the S type read from the five word dictionary memories according to the selection signal 80 (
(In FIG. 8, s=2) 0m d-bit comparators compare the character selected by the d pitch selector with the j-th character of each shift register 2.

On the other hand, the remaining d-bit comparator checks whether the character selected by the d-bit selector is a residual symbol. At the output of the final OR gate, when a match is detected in any of these (m+1> comparators), the match signal 30
is output. FIG. 9 is a diagram showing another example of the configuration of the j-th character comparison circuit 3. In the configuration shown in FIG. 9, for each J character read from each word dictionary memory, (m+
1) The d-bit comparators compare the ``th'' character of the m shift registers with the remaining symbols, and the results are selected and output using the 1-bit selector.

In synchronization with the determination clock 72, the determination circuit 6 determines whether a word exists in the word dictionary memory 1 in the character string input by the input device 5 when a matching signal is detected from all of the n comparison circuits 3. This is a circuit that determines that a word has appeared.

FIG. 10 is a diagram showing a configuration example of the determination circuit 6. The 0 determination circuit 6 includes one AND gate and one
This can be achieved with 70 pieces of D-flips.

Next, the operation of this embodiment will be explained using an example.

11 and 12 (a) to () are the first candidate shift registers when n=6 in the data in the word dictionary memory, number of candidates m=3, and the length of the input character string is 6.・This is a diagram showing an example of changes in the contents of the second candidate shift register and the third candidate register. FIG. This is an example of a case where Further, FIG. 12 shows an example in which a voice recognition device is used as the input device 5 and a kana character string is input. In both FIG. 11 and FIG. 12, one row of the 3×6 matrix corresponds to the contents of each shift register 2, and the row direction corresponds to the character position (1 to 6), and the column direction corresponds to the candidate level (1 to 6). 3) 0 represents 0. For example, in the example shown in Figure 11, the character string input to the shift register 2 is as follows: ``1 day'', ``white'', ``eye'', ``white'', ``eye'',
The second character candidates are ``hon jr book'' [dai], the third character is ``E'', ``story'', ``tsume'', and the fourth character is ``no''.
[me] [tsu], the fifth character is [jo] [mata] [bun]
, the sixth character is ``Chapter 1'', ``kusa'', and ``early'', and the shaded area indicates that no characters are stored.

For an input character string as shown in FIG. 11 (when the input device 5 is a character recognition device), a switch is set so that word dictionary search processing is performed using the word notation as the heading as shown in FIG. 5(b). Set 8. That is, the second word dictionary memory (word dictionary memory when the notation is used as a heading) is selected.

At that time, in Fig. 11, (a) → (b) → (c)
→(d)-? (e) → (f) → (g) → (h) → (i)
The change →(J) →(k) →(1) indicates a change that occurs each time the shift clock 70 is generated. In each of the states (a) to (k), the counter clock 71 and the determination clock 72 are generated eight times, which is the same as the total number of words N. The address counter 4 has a shift clock 7
Since it is reset by 0 and counted up N times by the counter clock 71, in each state, the word dictionary memory 1 sequentially reads from the first word to the last word.
The notation of all words (N pieces) is read out. And n pieces (
In this example, in the comparator circuit 3 (6m), all the headings in the word dictionary memory 1 selected by the switch 8 and the 11th
A comparison will be made with the contents of the shift register shown in the figure.

As a result, in each state, for example, the following words (headings) in the second word dictionary memory 1, the judgment time 1i1
16 detects the occurrence.

(a) Not applicable b) Not applicable c) Not applicable d) Not applicable e) Not applicable f) 1 day” “White” “Eyes” “Japan” “0 large” “
"Shiraki" rJapanese j (g) "tree""book" r large""final" (h)
・1 word” [story] “tsume” (1) “no” “me” “tsu” (J) “length” [mata] “bun” “bun” (k) “chapter 1” “kusa” “haya” this Among these, each comparison circuit 3 is used when comparing the contents of the shift register 2 in FIG. 7 in state (f) with the heading "Japan" (= "Japan ΔΔΔΔ") in the word dictionary memory 1 in FIG. Explain the operation.

The first character comparison circuit compares "1 day of Japan ΔΔΔΔ" and the first character.
A match signal 30 is generated by a match with "1st day of the first character of the candidate shift register." The second character comparison circuit generates a match signal 30 based on a match between "hon" of "Japan ΔΔΔΔ" and "one of the second characters of the second candidate shift register". 3
From the character comparison circuit to the 6th character comparison circuit, "Japan Δ
Detect “Δ” (residual symbol) of “ΔΔΔ” and generate match signal 3
Generates 0. As a result, the determination circuit 6 receives the match signals 30 from all the comparison circuits 3 and detects the appearance of a word.

Now, when a character string as shown in FIG. 12 is input (in case the input device 5 is a voice recognition device), a word dictionary search process is performed using the pronunciation of the word as a heading as shown in FIG. 5(a). Switch 8 is changed so as to perform the following steps. That is, the first word dictionary memory (word dictionary memory with readings as headings) is reselected, and in the same way, all characters in the first word dictionary memory 1 are selected while sequentially advancing one character at a time. The heading is compared with the contents of shift register 2.

As a result, the words (headings) in the first word dictionary memory detected in each state are, for example, as follows.

(a (b (C) (d (e (f) Not applicable Not applicable Not applicable Not applicable Not applicable ``ni'' ``iJ ``su'' ``nij ``nihon''``ihan''``iho'' ``ion ” “Iou” “Ha” “O” “Han” “Han” “Habo” “Hau” “On” “Ou” “Hangu” “Hanbu” “Piggyback” “Hangun” (g) “Hanbun” “Hanbun” “Oubou” (h) “N” “U” “Hagu” “Habo” “Habun” (l) “Gu” “Bu” “Bo” “ ``gun''``bun''``bon'' ``gunkaJ ``bunkaJ (J) run''``u''``other'' (k) ``kaJ '<J'ga'' Note that if the input string length is K , the sequential feeding in the shift register 2 must be performed at least (K+n-1) times. Therefore, after inputting the input character string of length K, the input device 5 further performs the sequential feeding (n-1) times. It is necessary to input a dummy character string.Or, the controller 7
detects the end of the input character string, and then shifts the shift clock 70 once to the counter clock 71 and the judgment clock 7.
2 may be repeated N times each time (n-1) times.

Also, during the first (n-1) sequential feeds (from (a) to (e) in Figures 11 and 12), the input character string has not reached the beginning of the shift register 2, so There is no point in checking against word dictionary memory 1, so
During this period, the controller 7 may continuously generate only the shift clock 70 without generating the counter clock 71 or the judgment clock 72.

As described above, in this first embodiment, when a character recognition device is used as the input device 5 to perform a word dictionary search using a notation as a heading, a voice recognition device is used to search a word dictionary using a reading as a heading. When performing a word dictionary search, it can be applied universally by simply flipping a switch.

FIG. 2 is a block diagram showing the configuration of the second embodiment. In the first embodiment, only the occurrence of a word is detected, but in the second embodiment, information regarding the detected word is output. The second embodiment has an additional information memory 9 and an output register 1 added to the first embodiment whose configuration is shown in FIG.
0 is added.

Here, the additional information memory 9 is a memory that stores additional information other than the headings of the S types of word dictionary memory 1, such as part of speech, accent, meaning classification, notation, reading length, etc. of the word. The additional information memory 9 stores the additional information at the same address as the word in the word dictionary memory 1. FIG. 13 is a diagram showing an example of correspondence between the contents of the word dictionary memory 1 and the additional information memory 9. However, address a, for which only the part of speech is registered as additional information, corresponds to a word whose notation is "Japan", whose reading is "nihon", and whose part of speech is a proper noun. Address counter 4 is S
Not only types of word dictionary memory 1 but also additional information memory 9
Also, give a common address.

The output register 10 outputs all the information for one word read from the S types of word dictionary memory 1 and the additional information memory 9 (the contents of the S types of headings and the contents of the additional information) and outputs them as they are from the determination circuit 6. It is latched at the timing of clock 60.

The output register 10 is a register implemented by a flip-flop.

If we consider a reading device that recognizes text containing kanji and kana characters, and also performs sentence analysis and speech synthesis, as a result of a word dictionary search process using the notation as a heading,
Although the pronunciation and part of speech are necessary, the notation does not need to be output because it is not used in the speech synthesis process that is performed after the word dictionary search process. Conversely, if we consider an audio word processor that creates sentences using voice input, as a result of a word dictionary search process using readings as headings, the spelling and parts of speech are necessary, but the readings are not output. Therefore, instead of outputting all word information as in the second embodiment, the third embodiment does not output the headings used during detection.
This is an example.

FIG. 3 is a block diagram showing the configuration of this third embodiment. This third embodiment is based on the second invention of the present application.

In the third embodiment, a selector 11 is added in contrast to the second embodiment whose configuration is shown in FIG. In addition,
An additional information memory 9, a selector 11, and an output register 10 are added to the first embodiment.

In FIG. 3, the selector 11 selects (s-1) types of contents other than the word dictionary memory selected by the switch 8 among the S types of contents of the word dictionary memory 1 (s types of headings) and the additional information memory. For example, select and output the contents of 9 and 6.
When the second word dictionary memory 1 is selected by the switch 8 and the heading "Japan" in the second word dictionary memory 1 is detected, the first word dictionary memory 1 is selected. The content "Nihon" and the content "proper noun" of the additional information memory 9 are output. In other words, "Japan" is not output.

In the first to third embodiments above, examples were shown for the general case where there are m candidates for each character in the input character string. In word dictionary searches for kanji-kana conversion for audio output, each character in the input string is unique (m=
1>, in the second invention of the present application whose configuration is shown in FIG. 3, the case where m=1 is shown in FIG. 4 as a fourth embodiment. FIG. 2 is a block diagram showing the configuration of an embodiment of the present invention.

In the fourth embodiment, only one shift register 2 is required, and the components and operations are m
= 1, so a detailed explanation will be omitted. During kana-kanji conversion, the input kana character string and the first word dictionary memory (word dictionary memory with readings as headings)
The spelling, part of speech, etc. of the detected words are output. When converting Kanji to Kana, the input character string containing Kanji and Kana is compared with the headings in the second word dictionary memory (word dictionary memory with spellings as headings), and the readings of the detected words are calculated. , part of speech, etc. are output.

Note that the number of bits to represent one character may differ depending on the heading. For example, 16 bits are usually required to represent one character written in kanji and kana, but one character in reading is can be expressed in 8 bits. One way to deal with such cases is to match the number of bits with a large number of bits.
Another solution is to add 8 extra bits and allocate 16 bits.Another way to deal with this is to use only valid delimiters using the number of bits in the common divisor as a unit.For example, use 8 bits as a unit. When performing sequential forwarding and comparison, and in the case of notation-related verification, only the even-numbered forwarding times need to be valid. Note that since the pronunciation is 8 bits, it is valid every time.

(Effects of the Invention) As explained above, according to the present invention, even if each character has multiple candidates for a character string consisting of many types of characters such as kanji, it can be quickly compared to a word dictionary. A word dictionary search device capable of comparison is obtained. In particular, regardless of the number of candidates for each character in the input string and the length of the word in the word dictionary, matching the input string with one word in the word dictionary can be performed within about two clocks. The effect is large.

Furthermore, the number of search keys is not limited to one type, and matching can be performed by switching between multiple types of search keys. Furthermore, the word dictionary capacity does not increase unnecessarily due to the use of multiple types of search keys. The speed of the search is maintained for all types of search keys.

Further, each component of the present invention can be realized by combining a small number of logic ICs. Therefore, an advantage of using LSI technology is that it can be realized as a very compact device. In addition, unlike the first conventional technology, it is not implemented as software on a general-purpose computer;
Because it can be realized as dedicated hardware/special LSI,
The clock frequency itself can be set much higher than that of general-purpose computers, and in this respect, it is also superior in speed.

Due to the above-mentioned high speed, the present invention does not limit the number of words to be matched in the word dictionary as in the first prior art described above, but even when matching is performed with all words in the III word dictionary, A word dictionary search device that is sufficiently faster than the conventional one can be obtained. As a result, there is an advantage that the word dictionary does not need to be sorted in the order of the heading code. Therefore, even if words are added or deleted, there is no need to reorganize the word dictionary, and maintenance of the word dictionary is extremely easy.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of the first embodiment of the present M invention, FIG. 2 is a block diagram showing the configuration of the second embodiment, and FIG. 3 is a block diagram showing the configuration of the third embodiment. 4 is a block diagram showing the configuration of the fourth embodiment, FIG. 5 is a diagram showing an example of the contents of the word dictionary memory 1, FIG. 6 is a time chart of input signals of the controller 7, and FIG. is a diagram showing an example of the configuration of the shift register 2, FIGS. 8 and 9 are diagrams showing an example of the configuration of the comparison circuit 3, FIG. 10 is a diagram showing an example of the configuration of the determination circuit 6, and FIGS. (1) is a diagram showing an example of the change in the contents of shift register 2 in response to the input of a character string containing kanji and kana; Illustration showing an example, No. 13
The figure shows the correspondence between the word dictionary memory 1 and the additional information memory 9. DESCRIPTION OF SYMBOLS 1... Word dictionary memory, 2... Shift register (first candidate shift register), 3... Comparison circuit (J-th character comparison circuit), 4... Address counter, 5... Input device, 6... Judgment circuit, 7... Controller, 8...
... Switch, 9... Additional information memory, 1o... Output register, 30... Match signal, 5o... Input tarlock, 60... Detection clock, 7o... Shift clock, 71... - Counter clock, 72... Judgment clock, 80... Selection signal.

Claims (2)

[Claims] (1) An input device that inputs a character string in which there are m types of candidates from the first to the mth candidate (m is an integer where m≧1) for each character, and One heading is stored in each address with a data width of (an integer), and the part less than n characters is filled with predetermined residual symbols. a word dictionary memory (an integer where s≧1), one shift clock every time m types of candidates for one character are input by the input device, and determination of the number of times according to the total number of words in the word dictionary memory. a controller that generates a clock and a counter clock; an address counter common to the s types of word dictionary memories that performs reset in synchronization with the shift clock and count-up in synchronization with the counter clock; 1st, 2nd, . . . , m-th candidates corresponding to the first, second, . A shift register, a switch for selecting one of the s types of word dictionary memories, and the first and second characters of data for n characters read from the word dictionary memory selected by the switch.
Character: When the character at the corresponding position corresponding to the n-th character matches the character at the same position in the first, second, m-th candidate shift register, or the residual symbol. 1st character, 2nd character...nth character comparison circuit which outputs a match signal, and 1st character, 2nd character...nth character comparison circuit synchronized with the judgment clock. and a determination circuit that determines that a word existing in the word dictionary memory has appeared in the character string input by the input device when matching signals are detected from all of the characters. Device. (2) An input device for inputting a character string in which there are m types of candidates from the first to the mth candidate (m is an integer where m≧1) for each character; One heading is stored in each address with a data width of (an integer), and the part less than n characters is filled with predetermined residual symbols. an additional information memory storing additional information other than the s types of headings regarding a word at each address; and each time m types of candidates for one character are inputted by the input device. a controller that generates a shift clock once per cycle, a determination clock and a counter clock a number of times according to the total number of words in the word dictionary memory; a reset synchronized with the shift clock; and a count-up synchronized with the counter clock. an address counter common to the s types of word dictionary memory and the additional information memory; and the shift clock corresponding to the first, second, . a first, second, . . . , m-th candidate shift register for each n character, which sequentially shifts characters one by one in synchronization with the above; a switch for selecting one of the s types of word dictionary memories; and a switch for selecting one of the s types of word dictionary memories; The first and second characters of the data for n characters read out from the word dictionary memory selected in...The characters corresponding to the nth character are the first and second characters. The first and second characters that output a match signal when they match the character at the same position in any of the m-th candidate shift registers or the remaining symbol...
・If a match signal is detected from the n-th character comparison circuit, and all of the first character, second character, etc. n-th character comparison circuit in synchronization with the determination clock, the signal is input by the input device. a determination circuit that determines that a word existing in the word dictionary memory has appeared in a character string; and a determination circuit that determines that a word existing in the word dictionary memory has appeared in the character string; A word dictionary search device comprising: a selector for selectively outputting information in the additional information memory.