JPH06131403A - Multimedium word dictionary and its retrieval device - Google Patents

Abstract

PURPOSE:To not only display a natural language but also display a picture and output sounds at the time of retrieval of a word by registering natural language information, picture information, and sound information related to one word in a word dictionary as items of this word. CONSTITUTION:A multimedium word dictionary where natural language information related words, picture information related to words, and sound information related to words are registered as items of words is used as the word dictionary. Since natural language information, picture information, and sound information are simultaneously acquired at the time of accessing the word dictionary by a keyword in this constitution, required information is selectively taken and is outputted from a display device. For example, when the word dictionary is accessed by keyword 'snow grouse', pictures of a snow grouse in summer and winter are displayed and its song is outputted so that the snow grouse can be intelligibly explained to persons. Further, the retrieval time is short because it is sufficient if a data base is retrieved only once.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multimedia word dictionary and a search device therefor, and in particular, explanations of words are explained not only in natural language but also in images and voices, and as a result of searching words, the contents are understood. The present invention provides a search device that configures a dictionary so that it is easy to perform, and that can easily search for necessary contents.

[0002]

2. Description of the Related Art In a conventional word dictionary, explanation of a word is described in natural language, and therefore, an item which is easier to understand with an image or a voice is also explained in the language. For example, if you look up "black-tailed gull" in the word dictionary, an explanation of what kind of bird it is is presented. In this case, the appearance of "black-tailed gull" is only described as "similar to a seagull", and the cry of "black-tailed gull" is also described as "similar to a cat".

[0003]

Therefore, since the conventional word dictionary is explained by using the natural language as described above, it is difficult for a human to intuitively understand.

In a database system capable of retrieving image and sound data, even if the object to be retrieved is information on one word, the information on that word is not managed in one dictionary. It was not possible to make direct inquiries about the natural language, images, and sounds resulting from the above.

For example, in the case of searching for "grouse", when the user wants to inquire about the property of "grouse" after searching for the image of "grouse", conventionally, a description about "grouse" is given via the keyword attached to the image. I had to retrieve the information in the sentence. Also, when you want to hear the call of "grouse", if you search the image for the voice "I want to hear this call", you can retrieve the voice data via the keyword "grouse" attached to the image. Was needed and the search was troublesome.

[0006]

In order to solve the above problems, the present invention, as shown in FIG. 1, uses a word dictionary for natural language information about words and image information (still images, moving images) about words. , A multimedia word dictionary in which voice information about a word is registered in the word dictionary.

[0007]

In the present invention, since the word dictionary is constructed as described above, when accessing the word dictionary with a keyword, for example, "grouse", natural language information, image information, and voice information are obtained at the same time as described above. Therefore, it is possible to easily select and obtain a necessary one of these and output it from the display or the speaker. For example, a summer image and a winter image of "grouse" can be displayed and a squealing voice can be output, so that it can be explained intuitively and easily to humans. Moreover, since the search for the database is performed only once, the search can be performed in a short time.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described with reference to FIGS. 2 and 3 are explanatory views of a multimedia word dictionary according to the present invention, FIG. 4 is a block diagram of a search device according to the present invention and an operation explanatory diagram thereof, FIG. 5 is an operation explanatory diagram of an input analysis unit, and FIG. Is an operation explanation diagram of the object search part,
FIG. 7 is an operation explanatory diagram of the object presenting unit, and FIGS. 8 and 9 are interactive retrieval state explanatory diagrams, respectively.

As shown in FIG. 2A, the word dictionary according to the present invention includes headwords (words to be searched, for example, "grouse") and respective items (items of speech, kanji, explanations, etc.). Each item is provided with a flag indicating what kind of data it is (text, still image, moving image, voice, etc.).

For example, the "grouse" shown in FIG. 2 (B).
In the case of, the headword is "grouse", the items are "part of speech", "kanji", "description", "summer grouse", "winter grouse", and "calling", and the data indicating the type is " “Text”, “still image”, and “voice” are described respectively.

In the case of "Melody Clock" shown in FIG. 3, "Melody Clock" is used as the entry word, and "Part of speech", "Product name", "Description", "Appearance", "Karakuri action", and “Melody sound” describes “text”, “still image”, “moving image”, and “voice” as data indicating the type.

Therefore, when the dictionary is searched with "grouse" as the search item, all the contents shown in FIG. 2B are extracted, and the text, image,
The voice or the like can be immediately selected and output.

Next, the retrieval device of the present invention is shown in FIG.
I will explain based on this. In the figure, 1 is a user input / output unit, 2 is a search unit, 3 is an input analysis unit, 4 is an object search unit, 5 is an object presentation unit, and 6 is a word dictionary.

The outline of the operation is shown in FIG. First, when an input sentence is input to the search unit 2 from the user input / output unit 1, the input analysis unit 3 analyzes the input, and if this is not a termination notification, the object search unit 4 searches the word dictionary 6 for an object. A search is performed and the search result is presented to the user input / output unit 1 via the object presentation unit 5. If the user inputs an end notification from the user input / output unit 1 after performing a predetermined search, this will be input to the input analysis unit 3
And the search process is terminated.

Each section shown in FIG. 4A will be described below. The user input / output unit 1 is used by the user to input the contents to be searched and output the object obtained as a result of the search, and includes a keyboard, a display unit, a speaker, and the like.

The search unit 2 decodes the item requested by the user via the user input / output unit 1, accesses the word dictionary based on this, obtains the required contents, and sends it to the user input / output unit 1. It is a thing.

The input analysis unit 3 processes the input sentence from the user input / output unit 1 and extracts what the user requests and what the user is trying to retrieve.
The following processing is performed as shown in.

First, a morphological analysis is performed to divide the input natural language sentence into words. Next, the result of the morphological analysis is semantically analyzed. It should be noted that these morphological analysis and semantic analysis are conventional techniques, and whether the user wants to search by semantic analysis, the instruction (for example, whether to display or question), and whether the user is asking about more detailed items of the target. Etc. information is obtained.

From the result of the semantic analysis, the type of the user's instruction (whether it is a display request or an answer request to the question) and the question content (the instruction type is an answer request to the question). In the case), the search target name, the item designation (which part of the search target about which information is to be obtained), etc. are extracted.

For example, when a sentence "Tell me about grouse" is input from the user input / output unit 1, the input analysis unit 3 performs the above-mentioned processing to analyze, and the kind of instruction = answer request, question content = none, Search target name = grouse and item designation = none.

In the case of the sentence "Where does the grouse live?", The type of instruction = answer request, the content of the question =
(Grouse X live), search target name = grouse, item designation = none. Here (live grouse X)
Indicates that the place X where the grouse lives is to be obtained.

Further, the sentence "I want to hear the timbre of a clock called melody clock" is as follows: instruction type = display request, question content = none, search target name = melody clock, item designation = tone.

The object retrieving unit 4 retrieves data based on the user's instruction type, question content, retrieval target name, and item instruction information extracted by the input analyzing unit 3. The following processing is performed as shown.

First, the search target name is searched from the entry word of the word dictionary 6. If the user's instruction is a display request, the presence or absence of the item instruction is checked next. If there is no item instruction, all the items retrieved above are passed to the object presentation unit 5.

However, if there is an item instruction, the content of the item instruction is compared with each item in the dictionary, and the data of the item closest to the item is selected and passed to the object presenting section 5. If the user's instruction is not a display request in the above, the data of the explanation item is taken out. Then, the explanation sentence is searched for a sentence including the word being queried, a flag that is an answer to the question is attached, and the sentence is passed to the object presenting unit 5.

For example, the result of the sentence "Tell me about grouse" analyzed by the input analysis unit 3 is:
As described above, since the type of instruction = answer request, question content = none, search target name = grouse, item instruction = none, and there is no item instruction, as described above, all items retrieved by “grouse” are presented as objects. It will be passed to Part 5.

Further, the result of analyzing "I want to hear the timbre of the clock called melody clock" by the input analysis unit 3 is instruction type = display request, question content = none, search target name = melody clock, item designated timbre. In this case, as described above, the "melody sound" is the closest to the "timbre" of the item instruction, so this is selected and passed to the object presenting section 5.

In addition, the result of analyzing "where do grouses live?" By the input analysis unit 3 is as follows: instruction type = answer request, question content = (grouse X live), search target name =
Grouse, item designation = None, which is not a display request as described above, but a response to a question is requested. Therefore, search for the item of grouse description and select from the sentence. A sentence including “live” is searched, a question answer flag is attached, and the object presenting unit 5 is passed.

The object presenting section 5 retrieves the results retrieved by the object retrieval section 4 according to the type of data.
The processing for displaying on the screen of the user input / output unit 1 and outputting sound is performed, and the following processing is performed as shown in FIG. 7.

If there is data to be sent and presented to the user input / output unit 1, the operation proceeds to the next step, but if there is no more data, the operation ends. If there is data to be presented, the type of the data to be presented is checked, and if it is an image, the image display processing is performed on the user input / output unit 1.

However, if the data type is not an image but the presented data is voice, the user input / output unit 1
The voice output processing is performed on the. However, if the type of data is not voice and the answer flag to the question is added in, the text sentence end is converted, for example, from "to be" to "to or is" key. And the user input / output unit 1
The display processing of the text is performed for and the process returns to the above. However, if the question answer flag is not attached, this text display processing is performed on the user input / output unit 1 and the process returns to the above.

In this way, the image is displayed on the user input / output unit 1 as to the image, and the voice data such as squeal is output from the speaker by the user input / output unit 1. The text is displayed on the user input / output unit 1 in this case. In this case, in the case of an answer to the question, the end of the sentence is converted into the Desmus tone and displayed.

As a result, for example, as shown in FIG. 8, if a question "Tell me about grouse" is asked from the user input / output unit 1, in this case, the item instruction =
None, the input analysis unit 3 analyzes and the object search unit 4
Passes all the items searched by "grouse" from the dictionary 6 to the object presenting unit 5, and based on this, the display as shown in FIG. 8 is performed on the display unit of the user input / output unit 1 and the crying is output from the speaker. .

Similarly, regarding the melody clock, FIG.
By sequentially asking the question Q as shown in (A), the response A is sequentially output.
Answering question A as shown in FIG.
Will be sequentially output. The interactive search can be performed as shown in FIGS.

Therefore, not only the word dictionary but also online shopping is very effective.

[0036]

According to the present invention, the word dictionary is registered with the natural language information, the image information, and the voice information concerning one word as the items of the word. Therefore, when searching for a word, not only the natural language display but also the natural language display is performed. Since the image cover, voice output, etc. can be performed, it is possible to construct a dictionary whose contents are very easy for the user to understand.

Moreover, since these items are registered as one word item, the data of these items can be obtained at the same time on the basis of the word search when searching these items. Therefore, natural language information, image information, and voice information can be obtained. Since it is not necessary to search individually, not only the search speed is improved, but also because each of these pieces of information is registered as one word item, maintenance of these data becomes easy.

[Brief description of drawings]

FIG. 1 is a principle diagram of the present invention.

FIG. 2 is a dictionary explanatory diagram (1) of the present invention.

FIG. 3 is an explanatory diagram (part 2) of the dictionary of the present invention.

FIG. 4 is an explanatory diagram of a search device and operation of the present invention.

FIG. 5 is an operation explanatory diagram of the input analysis unit.

FIG. 6 is an operation explanatory diagram of an object search unit.

FIG. 7 is an operation explanatory diagram of the object presenting unit.

FIG. 8 is an explanatory diagram (1) of an interactive search state.

FIG. 9 is an explanatory diagram (2) of an interactive search state.

[Explanation of symbols]

 1 user input / output unit 2 search unit 3 input analysis unit 4 object search unit 5 object presentation unit 6 word dictionary

Claims (4)

[Claims] 1. A multimedia word dictionary in which a headword and natural language information about the headword are registered as items, and image information about the headword is registered as an item. 2. A multimedia word dictionary in which a headword and natural language information about the headword are registered as items, and voice information about the headword is registered as an item. 3. A multimedia word dictionary in which a headword and natural language information related to this headword are registered as items, and image information and voice information related to this headword are respectively registered as items. . 4. A multimedia word dictionary (6) in which an entry word and natural language information regarding this entry word are registered as items, and at least one of image information and audio information regarding this entry word is registered as an item.
An input analysis means (3) for analyzing an input sentence input by a user to extract a request and a search target; and a multimedia word dictionary (6) based on the information extracted by the input analysis means (3). An object retrieval means (4) for retrieving and an object presentation means (5) for outputting the retrieval result in the object retrieval means (4) to the user are provided, and the input sentence is analyzed by the input analysis means (3). Based on this, the object searching means (4) searches the multimedia word dictionary (6) to extract predetermined information, and the object presenting section (5) answers this to the user. Characteristic search device.