JP2001043224A - Language processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent the reduction of a recognition rate or increase in processing time caused by defining a large amount vocabulary or expression as object by performing language processing to an inputted component while using selected data for processing. SOLUTION: A user inputs the contents of a category to speak to a processing content determining part 1, selects the category contents by pressing a Japanese/English select button on a screen and inputs the selected category content information to a data selecting part 2 for external output and a data selecting part 3 for internal processing. Then, corresponding to the contents of respective determined category classes, the data selecting part 2 for external output selects respective relevant background and object images out of storage parts 5 and 6. Namely, when the background image is the domain of Japanese in the category contents stored for each language, the background image corresponding to Japanese and the object image corresponding to the domain are selected and the respective images are composited by a data compositing part 7 and outputted on a display panel 8.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a language processing apparatus for inputting speech or characters of a sentence, performing language processing on the input sentence, and outputting a processing result.
The present invention relates to a sentence analysis device, a language conversion device, and a dialogue device.

[0002]

2. Description of the Related Art A conventional language interpreter with image display means will be described below as an example of an interpreter. In recent years, an interpreter system that translates a speech input sentence into another language and outputs the translated speech has been put into practical use as computer software, and a device that is easy for a user to use using a display panel as a data display unit is being studied. .

FIG. 3 shows an example of these interpreters. Two windows previously stored in the image data storage unit are displayed on the display panel. When a voice is input, a sentence uttered by the voice recognition unit is recognized as a word string, and the character string of the recognition result is synthesized with the display window 1 of the screen by the external output data synthesis unit and output to the display panel.
Next, the recognized result sentence is input to the language translation unit, and the translation result character string is synthesized with the display window 2 and output to the display panel. In the current technology, a spoken sentence that is uttered cannot always be correctly recognized one by one, and a partially incorrect word string is output as a recognition result.

[0004] In this case, a device has been devised in which the user corrects an erroneous character on the display panel and translates the corrected input again to obtain a correct translation result required by the user. (Interpreting software CROSSROAD by NEC Corporation).

[0005]

However, in the above configuration, since the contents which the user may speak are not limited in advance, it is necessary to recognize many vocabularies expected to be spoken. Must. For this reason, a dictionary and a language model for large vocabulary are required, which makes it difficult to maintain a high recognition rate and increases processing speed. Moreover, as long as the vocabulary size is limited, it is actually difficult to recognize an arbitrary sentence, but in the above configuration, the user does not take any means to indicate the recognizable expressions and vocabulary, so There is a problem that it is difficult to speak.

The present invention has been made in consideration of the above problems, and presents category contents that can be used using an image display means to a user in a manner that allows each content to be easily understood before interpreting. By processing only expressions and vocabulary equivalent to, it is possible to prevent a decrease in recognition rate and an increase in processing time due to targeting a large number of vocabulary and expressions, and the user naturally determines the utterance content. It is an object of the present invention to provide a language processing device that can input data without hesitation.

[0007]

In order to solve the above-mentioned problems, a language processing apparatus according to claim 1 includes: a category content determining unit that determines the content of various categories of an input sentence; Display data to be displayed, processing data to be used when language processing the input sentence,
A data selection unit for selecting display data and processing data according to the content determined by the category content determination unit, a data display unit for displaying the selected display data, and inputting a sentence using voice or characters as a medium By providing a sentence input unit and a language processing unit for performing language processing on the input sentence using the selected processing data, the recognition rate is reduced by targeting a large amount of vocabulary and expressions. And a language processing apparatus capable of preventing an increase in processing time and allowing the user to naturally understand the utterance content and input without hesitation.

Further, the language processing means in claim 2 is
2. The language processing apparatus with data display means according to claim 1, wherein the display data is image data, thereby providing a more user-friendly language processing apparatus that allows a user to intuitively and naturally grasp speech content.

Further, the language processing device according to claim 3 is:
3. The language processing device with data display means according to claim 2, wherein the image data is data decomposable into a background and an object, so that the user can instantaneously and intuitively grasp the contents of a plurality of categories. User-friendly language processing device.

The language processing device according to claim 4 is
4. The language processing device with data display means according to claim 3, wherein when the category type is a language, the processing data and the background of the image data are used instead of each language, and the object of the image data is commonly used for the language. By using this, a user-friendly language processing device that allows a user to instantaneously and intuitively grasp the contents of a plurality of categories is provided.

Further, the language processing device according to claim 5 is
2. The language processing apparatus with data display means according to claim 1, wherein the processing data is composed of a word or a word string dictionary and language rules, thereby reducing a recognition rate by targeting a large amount of vocabulary and expressions. Provided is a language processing unit that prevents an increase in processing time, and that allows a user to naturally grasp utterance contents and input without hesitation.

[0012] Further, the language processing means in claim 6 comprises:
6. The language processing apparatus with data display means according to claim 5, wherein when the category type is a domain, the language rule is used commonly for the domain, and a word or a word string dictionary is used instead of each domain category content. Accordingly, it is possible to provide a language processing apparatus capable of preventing a reduction in recognition rate and an increase in processing time due to a large amount of vocabulary and expressions as well as processing required expressions.

Further, the language processing apparatus according to claim 7 is
6. The language processing apparatus with data display means according to claim 5, wherein the display data comprises image data and all or a part of a word or a word string dictionary.
By displaying the image data and the dictionary in combination in accordance with the category content, it is possible to prevent a decrease in recognition rate and an increase in processing time due to targeting a large amount of vocabulary and expressions.
In addition, the present invention provides a language processing device that allows a user to naturally understand the utterance content and input without hesitation.

[0014]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of a language processing apparatus according to claims 1 to 6 of the present invention. In the first embodiment, a bidirectional Japanese-English interpreter, which is one of the language processing devices, will be described. Japanese-English bidirectional interpreter recognizes and translates Japanese utterances and outputs English text or voice, and also processes Japanese utterances and outputs Japanese text or voice. Device. In this embodiment,
The contents of the domain will be described using a device for translating a conversation voice that will be used in a restaurant, a pharmacy, or a museum.

In the interpreting apparatus of the embodiment, there are languages and domains as category types that can be selected by the user. First, the user inputs the content of the category to be spoken to the processing content determining unit 1. For input method, for example, there are Japanese and English selection buttons and each domain selection button on the screen,
The user makes a selection by pressing a button. Next, the selected category content information is output to the external output data selection unit 2,
The data is input to the internal processing data selection unit 3. The external output data storage unit 4 stores the background image and the object image separately, and the external output data selection unit 2 stores the background image and the object image corresponding to the content of the determined category type. Each object image is selected from each of the storage units 5 and 6. For example, here, the background image is stored for each language and the object image is stored for each domain. If the category content is Japanese and the pharmacy domain is, the background image corresponding to Japanese and the pharmacy are stored. Select the corresponding object image. Then, the external output data synthesizing unit 7 synthesizes the two images, and
Output to the display panel 8.

On the other hand, the internal processing data storage section 9 stores a vocabulary dictionary and language rules as language rules necessary for speech recognition and language translation processing. The corresponding vocabulary dictionary and language rule are selected from the respective storage units 10 and 11 according to the content of each category type. For example, here, the language rules for speech recognition and language translation are stored for each language, and the vocabulary dictionary is stored for each language and domain. For example, if the category content is Japanese and the drug domain, then the Japanese language rules for recognition and the translation rules for translating from Japanese to English, and the vocabulary dictionary required for the Japanese pharmacy domain Are stored in the processing buffer.

Next, Japanese speech is input from a speech input terminal, and speech recognition is performed using the selected recognition language rule and vocabulary dictionary. The recognition result is input to the language translator and translated into English, and the translation result text is displayed on the display panel.

In the Japanese-English bidirectional interpreting apparatus of the embodiment described above, the user selects a language and a domain to be uttered before inputting a voice, and corresponds to a background image and a domain corresponding to the selected language. By synthesizing and displaying the object image and presenting it to the user, it is possible to instantaneously and intuitively grasp the category contents to be uttered by the user. In addition, the interpretation process is performed only on the vocabulary corresponding to the selected language and domain, so that it is possible to prevent a decrease in recognition rate and an increase in processing time due to the large number of vocabulary and expressions. Becomes

Hereinafter, a second embodiment of the present invention will be described with reference to the drawings.

FIG. 2 is a block diagram of a language processing apparatus according to claim 7 of the present invention. Similar to the first embodiment, a Japanese-English bidirectional interpreter, which is one of the language processors, will be described.

The interpreting apparatus of the embodiment selects a category type that can be selected by the user, inputs the selected category content to the external output data selection section 2 and the internal processing data selection section 3, and outputs image data and image data, respectively. The process of selecting a vocabulary dictionary is the same as in the first embodiment. Next, the external output data synthesizing unit 7 combines the above two images with a character string representing the selected vocabulary and outputs the result to the display panel 8. The interpreting process is the same as in the first embodiment.

The bidirectional Japanese-English bilingual interpreter of the embodiment
In addition to the effects of the embodiment, since the vocabulary that the user can utter can be confirmed at the character level, it is possible to instantaneously and intuitively grasp the category contents to be uttered.

[0024]

As is apparent from the above description, by using the method of the present invention according to claims 1 to 7, the recognition rate is reduced and the processing time is increased by targeting a large number of words and expressions. In addition, it is possible to provide a user-friendly language processing unit that allows the user to naturally understand the utterance content and input without hesitation.

[Brief description of the drawings]

FIG. 1 is a block diagram according to a first embodiment of the present invention;

FIG. 2 is a block diagram according to a second embodiment of the present invention.

FIG. 3 is a block diagram of a conventional method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Processing content determination part 2 External output data selection part 3 Internal output data selection part 4 Image data storage part 5 Background image data storage part 6 Object image data storage part 7,15 External output data synthesis part 8,16 Display panel 9 language processing data storage unit 10 vocabulary dictionary storage unit 11 language rule storage unit 12 buffer 13 speech recognition unit 14 language translation unit

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Kenji Matsui 1006 Kazuma Kadoma, Kadoma City, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. F term (reference) 5B091 AA03 BA03 CB12 CC03 CD01 DA02 DA06

Claims (7)

[Claims] 1. A category content determination unit for determining various contents of a language or style of an input sentence, a task or a domain (these are referred to as category types), and a plurality of display data displayed to prompt a user to input. And a plurality of language processing data to be used in language processing of an input sentence, and data for selecting appropriate data from a plurality of display data and processing data according to the content determined by the category content determining unit. Selection, a data display unit for displaying the selected display data, a sentence input terminal for inputting a sentence using voice or characters as a medium, and a language using the selected processing data for the input sentence A language processing device comprising: a language processing unit that performs processing. 2. The language processing apparatus according to claim 1, wherein the display data is image data. 3. The language processing apparatus according to claim 2, wherein the image data is data decomposable into a background and an object. 4. When the category type is language, the background of the processing data and the image data is used for each language instead,
4. The language processing apparatus according to claim 3, wherein the object of the image data is commonly used for a language. 5. The language processing apparatus according to claim 1, wherein the processing data comprises a word or word string dictionary and language rules. 6. The language according to claim 5, wherein when the category type is a domain, the language rule is used commonly for the domain, and a word or a word string dictionary is used instead of each domain category content. Processing equipment. 7. The display data comprises image data and all or a part of a word or word string dictionary, and the data display means displays the image data and the dictionary in combination in accordance with the category contents. The language processing device according to claim 5, wherein