CN1641635B - Computer executable word/phrase matching and querying system and method thereof - Google Patents

Abstract

The invention discloses a system and method for matching and inquiring words/phrases which can be executed by a computer, wherein the system comprises a database, a database and a query module, wherein the database provides the words/phrases and corresponding explanations for the words/phrases; a character recognition module, which generates a character/word to be queried after recognition according to the characters on the picture; a judging module, which determines a corresponding rule according to the letter case of the word/phrase to be inquired; a query module for querying whether there is a completely matched word or phrase in the database according to the corresponding rule; a spelling correction and fuzzy query module, which finds out a most similar word in the database when there is no completely matched word; and an output module. The method is to judge whether the word or phrase to be inquired belongs to the situation of full capitalization of the whole sentence and whether the word or phrase is in capitalization of the first letter, and to match a proper word or phrase in the database according to a preset corresponding rule so as to correctly match the word or phrase and correctly display the translation result.

Description

Computer-executable word/word matching and query system and method

technical field

The present invention relates to a system and method related to data processing, which can be applied to electronic dictionaries, especially a system and method for word/word matching and query by judging the case of words/words.

Background technique

Generally speaking, there will be many unfamiliar foreign language new characters in the picture displayed by the computer, some of these new words are single characters, and some are phrases. For these new words, you can consult the electronic dictionary (translation software) so that you can understand their meanings and make up for the obstacles caused by insufficient language skills. In addition to providing word/phrase translation queries, electronic dictionaries also provide many additional functions, such as full-text translation, web page translation or real-time translation, etc. These additional functions make translation faster and more convenient. Taking full-text translation as an example, its feature is that it can translate a large number of words at one time, and organize them into meaningful sentences, avoiding the disadvantage of slowly searching word by word. Taking real-time translation as an example, its feature is the convenience of operation, as long as the cursor touches the word/phrase to be queried on the screen, a window will be displayed and the translation of the word/word to be queried will be displayed, and the user does not need to You can get the translation of the word/phrase by typing the word.

For example, when a user opens an electronic file in a foreign language to read, in the past, when encountering an unfamiliar foreign language word, he must look through a foreign language dictionary (dictionary) to obtain the meaning of this foreign language word. ) dictionary is quite inconvenient and takes a long time to look up new words, so the faster way to look up is to use the electronic dictionary in the computer as an aid. It is very convenient and fast to use the electronic dictionary to look up new words in a foreign language. Just input the word/phrase to be queried to get the explanation of the word/phrase. Although the processing speed of the electronic dictionary query word/word is quite fast, it still takes some time to input the word/word. This is a fly in the ointment, so the faster and more convenient query method is to use the additional functions of the electronic dictionary , such as full-text translation, web page translation, or real-time translation.

Let's take real-time translation as an example to illustrate how it works and what problems exist. As shown in Figure 1, it is a flow chart of real-time translation operation. When the real-time translation function of the electronic dictionary is activated, a word/word to be queried will be retrieved according to the position of the cursor (step 100); Whether there is a fully matching word/phrase in the database (step 110); If a fully matching word/phrase can be found in the database, then display the explanation (step 130) of this word/phrase to be queried; Can not find the word/word that can fully match in the database, then carry out spell correction and fuzzy query (step 120), find out a word/word most similar with the word/word to be queried in the database, then display this recent Interpretation (step 130) of similar word/phrase. Although the use of the real-time translation function does not require the operation of keyboard input, and only needs to use the mouse to complete the purpose of querying new words, but in some cases, the correct explanation of the word/phrase to be queried may not be displayed correctly. For example, when the word/phrase to be queried appears at the beginning of a sentence or in the title line, it may appear in the form of capitalizing the first letter or the entire sentence, and some words/phrases have both uppercase and lowercase forms. different interpretations. When querying such words/phrases in this situation, it will cause the word/phrase to be queried to be matched with the database to get an incorrect translation result, and thus what is displayed is not the query result that the user actually wants.

The following is a specific example as an illustration. We take the frequently used word as as an example. The uppercase AS has two interpretations, namely "Anglo-Saxon", that is, "Anglo-Saxon" and "antisubmarine", that is, "anti-submarine". of"; its lowercase as has multiple interpretations, for example, like..., as, as...etc., its uppercase and lowercase translations have completely different interpretations. Therefore, when the entire sentence is all capitalized or the first letter is capitalized , for example, in the case of AS IT IS... or As it is..., etc., the lowercase as should be used as the translation result, but the word/word matching result will be matched with the uppercase AS words/phrases, so the translation is incorrect. Take AS IT IS as an example, because a word/phrase that exactly matches AS can be found in the database, that is, uppercase AS. Taking As it is as an example, although it is impossible to completely match a word/phrase of As, after spelling correction and fuzzy query results, the matched word/word is AS instead of as. The above two situations will result in an incorrect translation result, which is obviously not the query result that the user wants to see. Similarly, this type of situation may also occur in other additional functions of the electronic dictionary, such as full-text translation function, web page translation, etc.

Contents of the invention

In view of the aforementioned problems related to the additional functions of electronic dictionaries, the present invention proposes a computer-executable word/word matching and query system and method thereof, the main purpose of which is to determine whether the word/word to be queried belongs to The entire sentence is all capitalized and whether the first letter is capitalized, and then appropriate corresponding rules are adopted to solve the situation where the word/phrase to be queried is incorrectly matched with the database in the past, so as to achieve the correct matching of the word/phrase and the correct display of its translation results Purpose. For example, when the word/phrase to be queried belongs to the entire sentence in uppercase, the corresponding rules for the word/word to be queried should first be queried in its lowercase form, and when the lowercase form cannot completely match the database, other rules should be adopted. corresponding rules. Similarly, when the word/phrase to be queried is capitalized, the corresponding rules adopted for the word/phrase to be queried should also be queried in its lowercase form first.

In order to achieve the above object, the present invention provides a computer-executable word/word matching and query system, wherein the system includes: a database, which is built on a computer-accessible storage medium, storing words/words An explanation corresponding thereto; a text recognition module, which obtains the text input from an input device, and identifies the text to generate a word/word to be queried; a judgment module, according to the letter case of the word/word to be queried, Determine a corresponding rule; a query module, which searches out a completely matched word/word from the database according to the corresponding rule, and a corresponding translation of the completely matched word/word; a spelling correction and fuzzy query module, When the query module does not find the exact matching word/phrase, find a most similar word/phrase and the translation of the most similar word/phrase in the database; and a video output device for outputting The translation of the exact word/phrase and the translation of the closest word/phrase.

The present invention also provides a computer-executable word/word matching and query method, which is suitable for performing the following steps with a computer: obtaining a word/word to be queried inputted from an input device; The judgment of the upper and lower case of the letter of a word determines a corresponding rule; according to the corresponding rule, a completely matching word/word is found in a database, and the corresponding translation of the completely matching word/word is found; according to the corresponding rule, the database When no fully matching word/word is found in the database, a spelling correction and fuzzy query module is used to perform a spelling correction and fuzzy query function, and a most similar word/word is found in the database, and the most similar the translation of the word/phrase; and outputting the translation of the exact matching word/phrase or the translation of the most similar word/phrase with a video output device.

That is to say, the computer-executable word/word matching and query system and method thereof provided by the present invention must first judge whether the word/word to be queried belongs to the state of full capitalization and initial capitalization of the entire sentence, and then according to A predetermined query rule matches a suitable word/phrase in the database, and uses its corresponding interpretation as the output translation result. Concrete and feasible implementation modes of the present invention are described as follows with regard to coordinating accompanying drawings:

Description of drawings

Figure 1 is a flowchart of real-time translation operation;

Fig. 2 is the hardware architecture diagram of the system and method thereof of computer-executable word matching and query proposed by the present invention;

Fig. 3 is the system framework diagram of the system and method thereof of computer-executable word/word matching and query proposed by the present invention;

FIG. 4 is a detailed flow chart of the computer-executable word/word matching and query system and method thereof proposed by the present invention.

Wherein, the reference signs are explained as follows:

21 Input device 22 Central processing unit

23 Video output device

31 Database 32 Text Recognition Module

33 Judgment module 34 Query module

35 spelling correction and fuzzy query module 36 output module

Step 100 extracts the word/phrase to be queried

Step 110 inquires whether the word/phrase to be queried matches completely with the database

Step 120 performs spelling correction and fuzzy query functions

Step 130 outputs translation result

Step 400 extracts the word/phrase to be queried

Step 410 judges whether the whole sentence text is all capitalized

Step 420 query in lower case

Step 430 judges whether the lowercase form matches the database completely

Step 440 output result

Step 450 performs spelling correction and fuzzy query functions

Step 460 queries with the original word/word form

Step 470 judges whether the original word/word form matches completely with the database

Step 480 judges whether the first letter is capitalized

Detailed ways

As shown in FIG. 2 , it is a hardware architecture diagram of a computer-executable word/word matching and query system and its method. It is a computer system composed of an input device 21 , a central processing unit 22 and a video output device 23 . Its input device 21 can be the device that can operate the cursor that produces on the screen such as keyboard, mouse, digital board, light pen, and its video signal output device 23 also can be the display interface such as screen, liquid crystal display screen, touch screen, in order to provide use device for visual reception.

As shown in FIG. 3 , it is a system architecture diagram of a computer-executable word/word matching and query system and its method. It includes a database 31 , a character recognition module 32 , a judgment module 33 , a query module 34 , a spelling correction and fuzzy query module 35 and an output module 36 required to maintain system operation. Each module mentioned above is explained as follows:

The database 31 is a character (dictionary) database, which includes: a word/word field, which provides abundant words and words for query; an explanation field, which provides corresponding explanations of words and words.

The character recognition module 32 is used to capture the character/phrase on the screen to generate a single character or words that can be queried, that is, to generate a character/word to be queried after recognition.

The judging module 33 judges whether the word/phrase to be queried belongs to the whole sentence in all capital letters, and then determines the corresponding corresponding rules. If the word/phrase to be queried belongs to the whole sentence in all uppercase, then the lowercase form of the word/phrase to be queried is queried first; otherwise, the original form of the word/phrase to be queried is queried first. When the original form of the word/phrase to be queried cannot be completely matched with the database, then judge whether the word/phrase to be queried is capitalized, and then adopt the corresponding corresponding rules. If the word/phrase to be queried is capitalized, then Query in lowercase form of the word/phrase to be queried, otherwise perform spelling correction and fuzzy query functions.

The query module 34 performs matching with the database 31 according to the corresponding rules of the judging module 33, and generates a corresponding processing strategy according to the matching result. When the corresponding rules of the word/word to be queried completely match the database, its processing strategy is to output the explanation of the word/word to be queried; when the corresponding rules of the word/word to be queried cannot match the database, its processing strategy is Two, one is to judge whether the word/phrase to be queried is capitalized, and then query according to its corresponding rules; the other is to perform spelling correction and fuzzy query functions when it has been judged whether the word/phrase to be queried is capitalized .

Spelling correction and fuzzy query module 35, the function of this module is to find a word/word most similar to the word/word to be queried in the database 31, and output its corresponding explanation as the query result. And carry out spelling correction and fuzzy query module 35 have two opportunities, and one is when the corresponding rule of word/word to be queried and database 31 can't match, and another is that judging module 33 judges that word/word to be queried is not first letter when capitalized.

The output module 36 is used to output the translation result, which is to generate a window and use this window to display the translation result. The output information includes the matched words and their corresponding explanations, that is, the displayed explanations can be exact matching results or fuzzy query results.

As shown in FIG. 4 , it is a detailed flowchart of the computer-executable word/word matching and query system and its method of the present invention, which is also a preferred embodiment of the present invention. Its process is to extract a word/word to be queried (step 400); judge whether the word/word to be queried is the whole sentence text all capitals (step 410), if the word/word to be queried belongs to the whole sentence text all capitals, Then query (step 420) with the lowercase form of the word/word to be inquired about; Then judge whether there is the word/word (step 430) of complete match in the database, if there is the word/word of complete match, then output its translation result (step 430) 440); If there is no word/word of complete match, then carry out spell correction and fuzzy query function, match one and the word/word (step 450) that is most similar to the word/word to be queried, then output its translation result (step 440) ; If the word/word to be inquired is not the whole sentence text full capitalization, then with the original word/word query (step 460) of the word/word to be inquired about, judge whether there is the word/word (step 470) of complete match in the database, If there is a complete matching word/word, then output its translation result (step 440); if there is no complete matching word/word, then judge whether the word/word to be queried is capitalized (step 480), if the word to be queried If the /word is capitalized, then query with the lowercase form of the word/word to be queried (step 420), otherwise perform spelling correction and fuzzy query functions (step 450).

The above description is only one of the preferred embodiments of the present invention, and is not used to limit the implementation scope of the present invention; that is, all equivalent changes and modifications made according to the patent scope of the present invention are covered by the patent scope of the present invention.

Claims (8)

1. the executable character/word of computing machine mates and the system of inquiring about, and wherein this system includes:

One database, this database are to build on the accessible storage medium of a computing machine, store the character/word explanation corresponding with it;

One text-recognition module obtains the literal from input media input, and this literal is carried out identification to produce a character/word to be checked;

One judge module is write according to the alphabet size of this character/word to be checked, determines a corresponding rule;

One enquiry module, this enquiry module finds out a character/word that mates fully according to this rule of correspondence in this database, and the translation of the character/word that should mate fully mutually;

One spelling correction and fuzzy query module, when this enquiry module does not find this character/word that mates fully, in this database, find out one recently like character/word, and this recently like the translation of character/word; And

One video signal output unit, in order to the translation of exporting this character/word that mates fully with should be recently like the translation of character/word.

2. executable character/word coupling of computing machine as claimed in claim 1 and the system of inquiring about, wherein this judge module more comprises one whole literal and capitalizes judge module entirely, in order to judge whether this character/word to be checked belongs to whole sentence literal and capitalize entirely.

3. executable character/word coupling of computing machine as claimed in claim 1 and the system of inquiring about, whether wherein this judge module more comprises an initial caps judge module, be initial caps in order to judge this character/word to be checked.

4. the system of executable character/word coupling of computing machine as claimed in claim 1 and inquiry, wherein this rule of correspondence reaches group's combination of inquiring about with spelling correction and fuzzy query mode for the inquiry of small letter form, primitive form inquiry with this character/word to be checked.

5. the executable character/word of computing machine mates and the method for inquiring about, and is applicable to a computing machine and carries out the following step:

Obtain a character/word to be checked of importing from an input media;

According to the judgement that this character/word alphabet size to be checked is write, determine a corresponding rule;

In a database, find out a character/word that mates fully according to this rule of correspondence judgement, and the translation of the character/word that should mate fully mutually;

When in this database, not finding the character/word that mates fully according to this rule of correspondence, then carry out a spelling correction and fuzzy query function with a spelling correction and fuzzy query module, and in this database, find out one recently like character/word, and this recently like the translation of character/word; And

With a video signal output unit export this character/word that mates fully translation or should be recently like the translation of character/word.

6. executable character/word coupling of computing machine as claimed in claim 5 and the method for inquiring about, wherein this alphabet size is write judgement, for judging whether this character/word to be checked belongs to whole sentence literal and capitalize entirely.

7. executable character/word coupling of computing machine as claimed in claim 5 and the method for inquiring about, wherein this alphabet size is write judgement, whether belongs to initial caps for judging this character/word to be checked.

8. the method for executable character/word coupling of computing machine as claimed in claim 5 and inquiry, wherein this rule of correspondence is for reaching group's combination of inquiring about with spelling correction and fuzzy query mode with the small letter form inquiry of this character/word to be checked, the inquiry of former character/word.

Images