KR20050120014A - Reference and display method of electron dictionary using voice - Google Patents

Abstract

The present invention is an electronic dictionary built on a local computer or a network server capable of wired / wireless communication by spelling, syllable, or word-by-word voice to search and read various types of information about the word and output it as text and / or synthesized sound. Can be used to communicate with a local computer or a server on a network, and it is easier and more accurate to enter a word to retrieve information constructed in an electronic dictionary using a terminal such as a mobile phone or a PDA with minimal keys. Can be.

Description

{Reference and display method of electron dictionary using voice}

The present invention relates to a method of word search and result display of an electronic dictionary through voice recognition. More particularly, the present invention relates to an electronic computer built in a local computer or a network server capable of wired and wireless communication by voice inputted in spelling, syllable, or word units. The present invention relates to a method of searching a word and displaying a result of an electronic dictionary through voice recognition for searching and reading a variety of information on a corresponding word using a dictionary and outputting (displaying) the text and / or a compound sound.

Conventional book dictionaries such as Korean dictionary, English dictionary, Japanese dictionary, etc. take a method of searching words in the alphabetical order. Such a book dictionary has a hassle that a user must find a word that is bulky and spelled in order.

Advances in computer and communication technology have led to electronic dictionaries, where information on each word is databased on personal computers, laptops, and web servers. This electronic dictionary can be used anywhere and anytime. If you only have a laptop or a terminal capable of Internet communication, you can search various information about the word more quickly by simply inputting the word you want to search.

However, a terminal such as a personal computer or a laptop has keys for alphabets that can input various languages such as Korean, English, Japanese, etc., so that words can be easily input, but mobile and mobile phones, PDAs, etc. The terminal is provided with only a few keys, which is not suitable for electronic dictionaries that require manual entry of words.

The search engine for foreign words, idioms, sentences and sentences on the Internet of Korean Patent Application Publication No. 2000-30825 (published on June 05, 2000) is not only simple word information but also phonetic symbols, words, There is disclosed a technology for searching information on idioms and hearing the results by voice, but this does not deviate from the hassle of having to input a word to search by key manipulation.

In addition, an electronic dictionary with a speech database containing a number of words as a target word appeared, which has the advantage of being able to recognize and search a large number of words, but the recognition rate for the pronunciation (word) to be searched because there are too many similar pronunciations. This drops significantly.

The present invention for overcoming the limitations of the prior art as described above, the user voice using a terminal such as a mobile phone or PDA connected to a local computer or a server on a network capable of wired and wireless communication without resorting to a simple key input method. It aims to be able to easily input words to search from the electronic dictionary regardless of time or space.

In addition, an object of the present invention is to input a voice for a word to be searched from the electronic dictionary separately by spelling, syllable or word, so that the voice can be recognized more accurately.

The present invention for achieving the above object, the first step of selecting one or more output modes of the voice recognition mode of the spelling, syllable, word mode and the character and / or synthesized sound mode for the information output for the word And a second step of recognizing the words inputted in the spelling, syllable, and word state for each spelling, syllable, and word, temporarily storing them in a buffer, and combining one word with one word by the spelling, syllable, and word. When the input of the voice for constituting the voice is completed, the third step of searching for information on the corresponding word from the electronic dictionary database and the information on the searched word are output as letters and / or synthesized sounds according to the output mode selected in the first step. It characterized in that it comprises a fourth step.

Also confirmed in the present invention, in the second step, it outputs the homonym (同音異議語) of the word after completion of input of the pronunciation for the words or one word, a combination of one or more spelling or syllable as a letter or synthesized Preferably, the step is further provided.

In the present invention, in the third step, while inputting the spelling, syllables, or words for the words constituting the single word to be searched, a predetermined time has elapsed or no end of the voice input is input. When a specific key is manipulated or a voice set arbitrarily to notify the completion of the voice input is input, it is preferable to determine that the voice input for the word to be searched is finished.

In addition, in the present invention, it is preferable that the input and recognition of the voice for the word search, the search of the word and the output of information on the word are made in a local computer or a network server connected to communicate with the word.

Hereinafter, the present invention will be described in detail.

1 and 2, a configuration and process for searching for information on a specific word from an electronic dictionary database mounted on a local computer or the like through voice recognition will be described.

Referring to Fig. 1, in the present invention, a microphone for collecting a normal voice, a speaker for outputting voice, a monitor for displaying text, and the like are connected to a control unit 1 of a local computer. ) Controls the various processes for searching the electronic dictionary and displaying the result by recognizing the voice inputted in the spelling, syllable or word unit. In addition, the controller 1 has a buffer 3 for temporarily storing data about a voice recognized in units of spells, syllables, or words in a process of processing a voice.

The voice recognition unit 2 connected to the control unit 1 recognizes a spell, syllable or word that is input through a microphone, and transmits data on the spelling, syllable or word to the control unit 1. This voice recognition unit 2 is divided to selectively recognize the input voice according to spelling, syllable or word mode, and each mode is a shared or optimized individual HMM (Hidden Markov Model) parameter. , Grammar information, spelling or syllable or word mode word dictionary, spelling or syllable word mode word list, etc. are driven in real time by the voice recognition engine.

The control unit 1 includes an electronic dictionary database 4 capable of searching for information on various recognition result words recognized by the speech recognition unit 2, and information on words retrieved from the electronic dictionary database 4. A speech synthesizer 5 is connected to form a synthesized sound for information of words for display on a monitor or output through a speaker.

2, a process of searching for a word from an electronic dictionary and displaying the result of the word through an electronic recognition on a local computer having the configuration as shown in FIG.

First, according to an external input, the control unit 11 selects a voice recognition mode of one of spelling, syllable, or word modes provided in the voice recognition unit 2 for dictionary search and result display through voice recognition. Select one or more output modes of the character and / or synthesized sound mode for outputting information on (S11).

When the selected speech recognition mode is the spelling mode (S12), the controller 1 drives the speech recognition unit 2 to recognize the user's utterance in the spelling unit. The HMM parameter, the corresponding grammar information, the spelling mode word dictionary, the spelling mode word list, and the like are loaded (S13).

When the user's voice is input from the outside (S15) while waiting for the user's voice in the spelling mode (S14), the controller 1 dynamically loads the voice recognition engine of the voice recognition unit 2 and the database of the spelling mode in real time. loading) to perform voice recognition (S16). That is, the controller 1 outputs the recognized spelling as a character (monitor) or a synthesized sound (speaker) so that the spelling recognized by the speech recognition unit 2 can be confirmed as the same as the spelling made by the user (S17). do.

In addition, when it is difficult to recognize the spelling spoken by the speech recognition unit 2, that is, when the speech is too small or too much noise is difficult to recognize the specific spelling, the control unit 1 reproduces the speech for the spelling. The user is asked to input (S18).

Subsequently, after the pronunciation of one spell is input, the controller 1 checks whether the input of the utterance for all the spells constituting one word is finished (S19) and then spells one word. When the input of the voice for is terminated, information about the corresponding word is searched from the electronic dictionary database 4 (S21).

However, when all the spells for constituting a single word have not been entered, the controller 1 temporarily stores the data of the spells entered so far in the buffer 3 (S20) and again until the last spell is input (S20). Repeat steps S14 to S19 to recognize all the spellings that make up one word.

At this time, the control unit 1 has a waiting time for a predetermined time (3, 4 seconds) after the last voice is input, or a spell for forming a single word when a specific key is input or a specific voice is input. It can be recognized that the input of the utterance for is finished. However, even if 3 or 4 seconds have elapsed after the utterance of the last spell is input, the controller 1 does not form a single word if the spelled words have not been composed until now. Prompt for spelling.

Preferably, after input of a pronunciation of a word having a combination of one or more spells is completed, one or more words (synonyms) having the same pronunciation as the word are output as a letter or a compound sound so that the user can select the desired word. You will be able to recognize and search words more accurately.

The control unit 1, which retrieves information about the word composed of one or more spells from the electronic dictionary database 4, checks whether the output mode is selected as a character mode or a character and compound sound mode in step S11 (S22). ). If only the character mode is selected, the control unit 1 displays the meaning of the word through the monitor in letters (S24), and if both the character and the synthesis sound mode are selected, the control unit 1 displays the meaning of the word in letters. Simultaneously, the voice synthesizing unit 5 is driven to output the meaning of the word as synthesized sound (S23).

In addition, when the speech recognition mode selected in step S11 is the syllable mode S26, the controller 1 drives the speech recognition unit 2 to recognize the user's utterance in syllable units. Individual HMM parameters, corresponding grammar information, syllable mode word dictionary, syllable mode word list, etc. are loaded (S27).

When the user's voice is input from the outside (S29) while waiting for the user's voice in syllable mode (S28), the controller 1 dynamically loads the voice recognition engine of the voice recognition unit 2 and the database of the syllable mode in real time. loading) to perform voice recognition (S30). That is, the controller 1 outputs the recognized syllables as text (monitor) or synthesized sound (speaker) so that the syllable recognized by the speech recognition unit 2 is the same as the syllables uttered by the user (S31). do.

In addition, when the syllable spoken by the speech recognition unit 2 is not recognized, that is, when the speech is too small or excessively loud, it is difficult to recognize the syllable as a specific syllable, the controller 1 may re-utter the speech for the syllable. The user is asked to input (S32).

Subsequently, after the pronunciation of one syllable is input, the controller 1 checks whether input of utterances for all syllables constituting one word is completed (S33) and then syllables for composing one word. When the input of the voice for is terminated, information about the corresponding word is searched from the electronic dictionary database 4 (S35).

However, when all the syllables for constituting a single word are not input, the controller 1 temporarily stores data about the syllables input so far in the buffer 3 (S34) and again until the last syllable is input (S34). S28 to S33) are repeatedly performed to recognize all the syllables that constitute one word.

At this time, the control unit 1 has a waiting time for a predetermined time (about 3 or 4 seconds) after the utterance for the last syllable is input, or when a specific key is input or a specific voice is inputted to form a single word. The speech input for the syllable may be recognized as being terminated. However, even if 3 or 4 seconds have elapsed after the utterance of the last syllable is input, the controller 1 does not form a single word with the syllables input until now until a single word is composed. It may require input of syllables.

Preferably, after a pronunciation input for a word composed of one or more syllable combinations is completed, one or more words having the same pronunciation as the word may be output as letters or a compound sound so that the user can select the word that the user wants more accurately. You will be able to recognize and search.

The controller 1, which retrieves information about the word composed of one or more syllables from the electronic dictionary database 4, checks whether the output mode is selected as the text mode or the text and the synthesized sound mode in step S11 (S36). ). If only the character mode is selected, the controller 1 displays the meaning of the word through the monitor in letters (S38). When both the character and the synthesized sound mode are selected, the controller 1 displays the meaning of the word in letters. Simultaneously, the voice synthesizing unit 5 is driven to output the meaning of the word as synthesized sound (S37).

Finally, when the voice recognition mode selected in step S11 is the word mode S40, the controller 1 drives the voice recognition unit 2 to recognize the user's speech as a single word unit. The shared or optimized individual HMM parameters, corresponding grammar information, word mode word dictionary, word mode word list, etc. are loaded (S41).

When the user's voice is input from the outside (S43) while waiting for the user's voice in the word mode (S42), the controller 1 dynamically loads the voice recognition engine of the voice recognition unit 2 and the database of the word mode in real time. loading) to perform voice recognition (S44). That is, the controller 1 outputs the recognized word as a character (monitor) or a synthesized sound (speaker) so as to confirm whether the word recognized by the speech recognition unit 2 is the same as the word spoken by the user (S45). do.

In addition, when the speech recognized by the speech recognizer 2 is not recognized, that is, when the speech is too small or the noise is excessively large and it is difficult to recognize the specific word, the control unit 1 again reproduces the speech for the word. The user is asked to input (S46).

Subsequently, when a time of about 3 or 4 seconds has elapsed after the pronunciation of one word is input, the controller 1 searches the electronic dictionary database 4 for information on the word (S47). .

In this case, the controller 1 may recognize that the input of the utterance of one word is ended when a waiting time of a certain time elapses, a specific key is input, or a specific voice is input after the utterance of a specific word is input. have. However, even if 3 or 4 seconds have elapsed after the utterance of the last word is input, the control unit 1 does not form a single word using the words entered so far until a single word is composed. May require input of words.

The control unit 1, which retrieves information about one word from the electronic dictionary database 4, checks whether the output mode is selected as a character mode or a character and synthesized sound mode in step S11 (S48). When only the character mode is selected, the controller 1 displays the meaning of the word through the monitor as a character (S50). When both the character and the synthesized sound mode are selected, the controller 1 displays the meaning of the word as a character. Simultaneously, the voice synthesizing unit 5 is driven to output the meaning of the word as synthesized sound (S49).

Up to now, the electronic dictionary database has been described in a case where the local computer is installed. Hereinafter, a network server (or a remote server) connected to communicate with a terminal such as a mobile phone or a PDA through voice recognition will be described with reference to FIGS. 3 and 4. The configuration and process for retrieving information about a specific word from an electronic dictionary database mounted in the system will be described.

Referring to FIG. 3, in the present invention, a terminal such as a mobile phone, a regular phone, or a local computer capable of internet communication is connected to a control unit 11 on a network server 10 through a wireless network, a PSTN, or an Internet network. The controller 11 controls a variety of processes for retrieving an electronic dictionary by recognizing a voice input by spelling, syllable or word unit, and displaying the result. In addition, the control unit 11 has a buffer 13 for temporarily storing data about a voice recognized in units of spells, syllables, or words in the process of processing a voice.

In particular, unlike a local computer or the like, anyone can connect using a communication terminal, so it would be desirable to have a separate authentication unit 16 so that only the members registered as members can search the electronic dictionary and check the results. Of course, it is natural to have a separate member information database for storing member information such as ID and password of each member. Of course, a mobile phone connected to a wireless network or a general phone connected to a PSTN may be identified as a member through CID authentication without using an ID or password.

The voice recognition unit 12 connected to the control unit 11 located on the network server 10 is a voice input through a mobile phone, a PDA, a general phone, a local computer, etc. (hereinafter, referred to as a "terminal"). Recognizes the spelling, syllable, or word and transmits data on the spelling, syllable, or word to the controller 11, and the voice recognition unit 12 transmits the input voice to the spelling, syllable, or word mode. Each mode has its own HMM parameter, grammar information, spelling or syllable or word mode word dictionary, spelling or syllable word mode word list, etc. Is driven in real time.

Here, the voice recognition unit 12 on the network server 10 to recognize the speech in the form of spelling, syllables, or words, firstly, when the user speaks, the voice is generated in real time instead of the independent voice file. A part may be transmitted to the voice recognition unit 12 as a stream and recognized by the voice recognition engine. The second method is to detect one syllable when the user speaks one syllable or one word, and transmits it to the speech recognition unit 12 as a signed independent voice file such as WAV or PCM to be recognized by the speech recognition engine. to be. The last method is a method of extracting only a feature for speech recognition of one word or syllable when the user utters one word or one syllable and transmitting the file to the speech recognition unit 12 in a file form for recognition by the speech recognition engine.

In particular, the voice feature transmission method has a smaller file size than the voice file transmission method, and the voice streaming method is transmitting in the case of poor network conditions or poor wired / wireless networks. Loss of some packets can reduce speech recognition performance.

The control unit 11 on the network server 10 includes an electronic dictionary database 14 capable of searching for information on various words recognized through the voice recognition unit 12, and the voice recognition unit 12. The synthesized sound of the information of the word to display the information about the word searched from the electronic dictionary database 14 by the spelling, syllable or word unit, or the electronic dictionary database 14 in a letter form or as a synthesized sound. The speech synthesis unit 15 to be formed is connected.

Fig. 4 retrieves words from the electronic dictionary through voice recognition on a terminal such as a mobile phone, PDA, general telephone, local computer, etc. and a network server (or remote server) connected thereto with the configuration as shown in Fig. 3 and displays the result. Explain the process.

First, the control unit 11 on the network server 10 provides user information such as an ID and password to a user who accesses the network server 10 by using a terminal such as a mobile phone, a PDA, a general phone or a local computer (S57). After requesting the input (58), it searches for the inputted member information and the member information stored in the member information database (S59) to check whether the connected user is a legitimate member (S60). Of course, when the user is a mobile phone connected to a wireless network or a general phone connected to a PSTN, the user may be identified as a member through CID authentication without using an ID or password.

As a result of the verification, in the case of a legitimate member (confirming the user's CID), the control unit 11 on the network server 10 according to an external input may spell or be provided in the speech recognition unit 12 for preliminary search and result display through voice recognition. In addition to selecting one of the syllable and word modes, a voice recognition mode is selected, and at least one output mode among the text and / or synthesized sound modes for outputting information on a specific word is selected (S61).

When the selected voice recognition mode is the spelling mode (S62), the control unit 11 on the network server 10 drives the voice recognition unit 12 to recognize the user's utterance in spelling units, that is, spelling mode, that is, sharing or optimization. The individual HMM parameters, the corresponding grammar information, the word dictionary in the spelling mode, and the word list in the spelling mode are loaded (S63).

When the user's voice is input from the terminal through various communication networks (S65) while waiting for the user's voice in the spelling mode (S64), the control unit 11 on the network server 10 is connected to the voice recognition engine of the voice recognition unit 12. Speech recognition is performed by dynamically loading the database in the spelling mode in real time (S66). That is, the control unit 11 on the network server 10 outputs the recognized spelling as a character or a synthesized sound so that the spelling recognized by the speech recognition unit 12 can be confirmed as the spelling made by the user (S67). do.

In addition, when it is difficult to recognize the spelling spoken by the speech recognition unit 12, that is, when the speech is too small or too much noise is difficult to recognize the specific spelling, the control unit 11 on the network server 10 may respond to the corresponding spelling. The user is asked to input the utterance for spelling again (S68).

Subsequently, when a time of about 3 or 4 seconds has elapsed after the pronunciation of one spell is input, the control unit 11 on the network server 10 inputs the utterance of all the spells for forming one word. When the input of the voice for spelling for constituting a single word is terminated after checking whether the word is terminated (S69), the electronic dictionary database 14 searches for information on the word (S71).

However, when all the letters for constituting a single word are not entered, the control unit 11 on the network server 10 temporarily stores the data for the spellings input so far in the buffer 13 (S70) and again the last letter. Steps S64 to S69 are repeatedly executed until inputting is performed to recognize all the spellings that constitute one word.

At this time, the control unit 11 on the network server 10 has a wait time of a certain time after the last utterance is input, or a specific key is input or a specific voice is inputted to form a word. It can be recognized that the input of utterances for the voice is completed. However, even if 3 or 4 seconds have elapsed after the utterance of the last spell is input, the controller 1 does not form a single word if the spelled words have not been composed until now. Prompt for spelling.

Preferably, after the pronunciation input for a word composed of one or more combinations of letters is completed, one or more words having the same pronunciation as the word may be output as letters or a compound sound so that the user can select the desired word more accurately. You will be able to recognize and search.

The control unit 11 on the network server 10, which retrieves information about the word composed of one or more spells from the electronic dictionary database 14, changes the output mode to the character mode or the character and synthesized sound mode in step S61. Check whether it is selected (S72). If only the character mode is selected, the control unit 11 on the network server 10 displays the meaning of the word as a character through a monitor (S74). When both the character and the synthesized sound mode are selected, the controller 11 on the network server 10 is displayed. The controller 11 displays the meaning of the word as a letter and simultaneously drives the speech synthesis unit 15 to output the meaning of the word as a synthesized sound (S73).

In addition, when the speech recognition mode selected in step S61 is the syllable mode S76, the controller 11 on the network server 10 drives the speech recognition unit 12 to recognize the user's speech in syllable units. That is, the shared or optimized individual HMM parameters, the corresponding grammar information, the word dictionary in the syllable mode, the word list in the syllable mode, and the like are loaded (S77).

When the user's voice is input from the terminal through various communication networks (S79) while waiting for the user's voice in the syllable mode (S78), the control unit 11 on the network server 10 is connected to the voice recognition engine of the voice recognition unit 12. Speech recognition is performed by dynamically loading the syllable mode database in real time (S80). That is, the control unit 11 on the network server 10 outputs the recognized syllables as text or synthesized sound so that the syllables recognized by the speech recognition unit 12 are the same as the syllables uttered by the user (S81). do.

In addition, when it is difficult to recognize the syllables spoken by the voice recognition unit 12, that is, when the voice is too small or excessively contains noise, it is difficult to recognize the specific syllables. The user is asked to input the utterance for the syllable again (S82).

Subsequently, when a time of about 3 or 4 seconds has elapsed after the pronunciation of one syllable is input, the control unit 11 on the network server 10 inputs the utterance input for all syllables to compose one word. When the input of the voice for the syllable for constituting a single word is finished after checking whether the word is completed (S83), the electronic dictionary database 14 searches for information on the word (S85).

However, when all syllables for composing one word are not input, the control unit 11 on the network server 10 temporarily stores data about the syllables input so far in the buffer 13 (S84) and again the last syllable. Steps S78 to S83 are repeatedly performed until this input is performed to recognize all syllables for forming one word.

At this time, the control unit 11 on the network server 10 has a waiting time of a certain time after the voice is input for the last syllable, a specific key is input, or a specific voice is input to the syllable for forming a single word. It can be recognized that the input of utterances for the voice is completed. However, even if 3 or 4 seconds have elapsed after the utterance of the last syllable is input, the controller 1 does not form a single word with the syllables input until now until a single word is composed. It may require input of syllables.

Preferably, after a pronunciation input for a word composed of one or more syllable combinations is completed, one or more words having the same pronunciation as the word may be output as letters or a compound sound so that the user can select the word that the user wants more accurately. You will be able to recognize and search.

The control unit 11 on the network server 10, which retrieves information about the word composed of one or more syllables from the electronic dictionary database 14, changes the output mode to the character mode or the character and synthesized sound mode in step S61. Check whether it is selected (S86). If only the character mode is selected, the control unit 11 on the network server 10 displays the meaning of the word in letters (S88). When both the character and the synthesized sound mode are selected, the control unit 11 on the network server 10 is selected. ) Displays the meaning of the word as a letter and simultaneously drives the speech synthesis unit 15 to output the meaning of the word as a synthesized sound (S87).

Finally, when the speech recognition mode selected in step S61 is the word mode S90, the controller 11 on the network server 10 may recognize the speech recognition unit 12 in order to recognize the user's speech as one word unit. In operation S91, a word mode, that is, an individual HMM parameter shared or optimized, corresponding grammar information, a word dictionary of a word mode, a word list of a word mode, and the like are loaded.

When the user's voice is input from the terminal through various communication networks (S93) while waiting for the user's voice in the word mode (S92), the control unit 11 on the network server 10 is connected to the voice recognition engine of the voice recognition unit 12. Speech recognition is performed by dynamically loading the database of the word mode in real time (S94). That is, the control unit 11 on the network server 10 outputs the recognized syllables as characters or synthesized sounds so that the words recognized by the voice recognition unit 12 are identical to the words spoken by the user (S95). do.

In addition, when it is difficult to recognize a word spoken by the voice recognition unit 12, that is, when the voice is too small or excessively much noise, it is difficult to recognize a specific word, the control unit 11 on the network server 10 The user is asked to input the utterance for the word again (S96).

Subsequently, when the pronunciation of one word is input and the time of about 3 or 4 seconds has elapsed, the control unit 11 on the network server 10 sends information about the word to the electronic dictionary database 14. Search (S97).

At this time, the control unit 11 on the network server 10 inputs a utterance for a single word when a waiting time of a certain time elapses, a specific key is input, or a specific voice is input after a utterance is input for a specific word. It can be recognized as finished. However, even if 3 or 4 seconds have elapsed after the utterance of the last word is input, the control unit 1 does not form a single word using the words entered so far until a single word is composed. May require input of words.

The control unit 11 on the network server 10 that retrieves information about one word from the electronic dictionary database 14 checks whether the output mode is selected as a text mode or a text and synthesized sound mode in step S61. (S98). If only the character mode is selected, the control unit 11 on the network server 10 displays the meaning of the word in letters (S90). When both the character and the synthesized sound mode are selected, the control unit 11 on the network server 10 is selected. ) Displays the meaning of the word as a letter and simultaneously drives the speech synthesis unit 15 to output the meaning of the word as a synthesized sound (S99).

Therefore, according to the word search and result display method of the electronic dictionary through the voice recognition according to the present invention,

1) The present invention is to search from the electronic dictionary by user voice even using a terminal such as a mobile phone or a PDA connected to a local computer or a network server (remote server) capable of wired or wireless communication without relying on a simple key input method. There is an effect that can easily enter a word regardless of time or space.

2) In another aspect, the present invention has an effect of improving the speaker independent speech recognition rate by separately inputting a voice for a word to be searched from an electronic dictionary by spelling, syllable or word.

Although the present invention has been described in detail only with respect to the specific embodiments described, it will be apparent to those skilled in the art that various changes and modifications are possible within the technical scope of the present invention, and such modifications and modifications belong to the appended claims. .

1 is a block diagram of a main part of a local computer having a function of searching for a word in an electronic dictionary and displaying the result through voice recognition according to the present invention.

FIG. 2 is a flowchart illustrating a method of searching a word from an electronic dictionary and displaying the result of the word through a voice recognition on a local computer having the configuration as shown in FIG.

3 is a block diagram of a terminal such as a mobile phone or a PDA having a function of searching a word in an electronic dictionary and displaying the result through voice recognition according to the present invention, and a main part of a network server connected thereto.

FIG. 4 is a flowchart illustrating a method of searching for a word from an electronic dictionary and displaying the result through voice recognition on a terminal such as a mobile phone or PDA having a configuration as shown in FIG. 3 and a network server connected thereto.

-Explanation of symbols for the main parts of the drawings-

1, 11: control unit 2, 12: speech recognition unit

3, 13: buffer 4, 14: electronic dictionary database

5, 15: speech synthesizer 10: network server

16: certification unit

Claims (4)

Selecting at least one output mode among a voice recognition mode of one of spelling, syllable, and word modes, and a text and / or synthesized sound mode for outputting information about a word; A second step of recognizing the audio inputted in the spelling, syllable, and word states for each spelling, syllable, and word, temporarily storing the voice in a buffer, and combining the words; A third step of retrieving information on the word from the electronic dictionary database when the input of the voice for constituting a single word is terminated by the spelling, the syllable, and the word; And A fourth step of outputting information on the searched word as text and / or synthesized sound according to the output mode selected in the first step; Word search and result display method of the electronic dictionary through the voice recognition, characterized in that it comprises a. The method of claim 1, In the second step, after the input of a word having a combination of one or more letters or syllables or a pronunciation of one word is completed, the step of outputting and confirming a homonym of the word as a letter or a compound sound is further provided. Word search and result display method of the electronic dictionary through the voice recognition, characterized in that. The method of claim 1, In the third step, while inputting the spelling, syllable, or the voice for the word constituting the single word to be searched, a specific time has elapsed without inputting any voice signal or a specific key for notifying the end of the voice input is operated or When the voice is set arbitrarily to inform the completion of the voice input, the word search and result display method of the electronic dictionary through the voice recognition, characterized in that it is determined that the input of the voice for the word to be searched is finished. The method of claim 1, The word search of the electronic dictionary through voice recognition, characterized in that the input and recognition of the voice for the word search, the search of the word and the output of information on the word is performed on the local computer or a network server connected to communicate with the word. How to display the results.