JPH09237270A - Device and method for supporting sentence elaboration - Google Patents

Abstract

PROBLEM TO BE SOLVED: To read sentence by speech sound synthesis, to set an editing mode during the reading and to support the sentence elaboration of a user. SOLUTION: In a read aloud mode, a sentence is sequentially segmented and analyzed, read-aloud data are generated from the analysis result and are read aloud by speech synthesis (A11-A15). When a correction instruction is given from the user during the read-aloud (A17), the editing mode is set and the pertinent sentence can successively be edited. At that time, an error in the pertinent sentence is detected and instructed. Thus, the sentence elaboration of the user is supported (A18-A22).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a text drafting support device used in a document creation device such as a Japanese word processor, and more particularly to a text drafting support device for drafting a text by utilizing a text reading function by voice synthesis. And a method for supporting text revision.

[0002]

2. Description of the Related Art Conventionally, when a sentence in a document created by a document creating device such as a Japanese word processor is selected, the sentence is read by oneself or by having another person read the sentence. Modifying is a common practice.

On the other hand, there is a device in which the reading of such a sentence is automated by voice synthesis. By using such a device, the text can be refined while listening to the voice generated from the speaker without the prepared text being read by the creator himself or others.

[0004]

As described above, conventionally, there has been a system that reads out a sentence by voice synthesis. However, this is only for reading the created text, and it is not possible to assist the user in text writing by setting the edit mode during the reading and detecting text errors. There wasn't.

The present invention has been made in view of the above points, and a sentence selection support capable of supporting the user's sentence selection by reading a sentence by voice synthesis and setting an edit mode during the reading. The present invention relates to a device and a text revision support method.

[0006]

[Means for Solving the Problems]

(1) The present invention sequentially cuts out a sentence in predetermined units from document information consisting of a plurality of sentences, analyzes the cut-out sentence to create reading data, and creates the reading data of the created sentence. When reading aloud by voice synthesis, and when text editing is instructed during the reading,
The editing mode is set so that the text can be edited one after another.

Further, when the edit mode is set, an error in the sentence is detected using the analysis result, and the detection result is presented. (2) The present invention sequentially cuts out a sentence in a predetermined unit from document information consisting of a plurality of sentences, analyzes the cut-out sentence to create reading data, and creates the reading data of the created sentence. When reading aloud by voice synthesis, and when text editing is instructed during the reading,
A marker is added to the sentence at that time, and the edit mode is set at a predetermined timing so that each sentence to which the marker is added can be collectively edited later. Further, when the edit mode is set, the analysis result is used to detect an error in each sentence, and the detection result is presented.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a text revision support device according to an embodiment of the present invention. In the figure, 11 is a control unit that controls the entire apparatus.
The control unit 11 is composed of, for example, a microprocessor, and executes image output processing, audio output processing, and the like in addition to key input determination processing, document creation and editing control processing.

An input unit 12 is used by the user to input characters and commands.
The input unit 12 includes an input device such as a keyboard, a mouse, and a pen. A display unit 13 displays a text, a message, etc. to be edited. This display unit 13
Are, for example, LCD (Liquid Crystal Display) and CRT
It consists of a display device such as (Cathode Ray Tube).

A document storage unit 14 stores document information composed of a plurality of sentences. The document storage unit 14 includes a memory such as a floppy disk device (FDD), a hard disk device (HDD), and a RAM. An analysis dictionary storage unit 15 stores an analysis dictionary used for sentence morphological analysis processing and syntax analysis processing. The analysis dictionary storage unit 15 is composed of a memory such as a ROM.

Reference numeral 16 denotes a one-sentence acquiring unit, which cuts out a sentence from the document information to be edited in units of one sentence. Reference numeral 17 denotes a sentence analysis unit, which performs morphological analysis and syntax analysis on the sentence cut out by the one sentence acquisition unit 16. This sentence analysis unit 1
The analysis result obtained by 7 is used not only as data for performing speech synthesis, but also for detecting an error in a sentence.

Reference numeral 18 is a reading data creating unit, which creates reading data mainly composed of reading kana based on the analysis result obtained by the sentence analyzing unit 17. A voice synthesizer 19 synthesizes voice data based on the read-out data created by the read-out data creator 18 and reads it out through the speaker 21.

Reference numeral 20 denotes an error detection unit, which is under the control of the control unit 11 when the editing mode is set.
The error in the sentence is detected using the analysis result obtained by, and the error result is presented.

Next, the operation of the embodiment will be described. When a user revises a sentence and makes corrections and changes while reading the sentence with synthesized voice, (a) a sequential edit type processing method for sequentially editing the sentence, and (a) There is a batch edit type processing method in which editing is carried out collectively after the reading of sentences is completed, and each case will be described here.

The sequential edit type / collective edit type can be arbitrarily designated by the user by initial setting or the like. Further, each processing is roughly divided into a text reading mode and an editing mode. The reading mode is a mode for reading a sentence by voice synthesis. Edit mode is
This is a mode for editing sentences.

(A) Sequential edit type FIG. 2 is a flowchart showing the sequential edit type processing operation. When the user selects a sentence, the control unit 11 first sets the reading mode (step A1).
1). In this reading mode, the 1-sentence acquisition unit 16
Cuts out each sentence in the document stored in the document storage unit 14 one by one (step A12). The cut-out sentence is given to the display unit 13 and displayed on the screen (step A13).

In this case, a mark is added to the beginning and the end of the sentence, and the marked sentence, that is, the sentence that is currently read out is highlighted to distinguish it from other sentences, or the sentence is displayed on the screen. Performs display control so that it is centered. As a result, the user can always easily confirm the current target sentence.

FIG. 4 shows a display example. In the figure, "I went to play with him." Is the current reading target sentence 31. The reading target sentence 31 is highlighted and displayed in the center of the screen.

The sentence cut out by the one-sentence acquisition unit 16 is also given to the sentence analysis unit 17. The sentence analysis unit 17 analyzes the sentence using the analysis dictionary stored in the analysis dictionary storage unit 15 (step A14). The reading data creating unit 18 creates reading data of the sentence based on the analysis result (step A1).
5).

In this case, Japanese kanji has on-phonetic readings and kun-yomi readings, and there are some hiragana characters whose pronunciation changes due to particles. Therefore, when creating the reading data, the sentence analysis unit 17 divides the sentence into morpheme units, obtains part-of-speech information from the connections before and after, and further performs syntactic analysis to determine the co-occurrence relationship. It is necessary to uniquely determine the reading. The reading information necessary for speech synthesis is obtained from such information.

FIG. 5 shows an example sentence and its morphological analysis result, syntactic analysis result, and reading information. For example, in a sentence such as "I went to school.", "I", "ha", "school",
Morphological analysis results such as "he", "go", and "ta" are obtained. The morphological analysis result includes part-of-speech information and reading information. This information is obtained from the analysis dictionary. Also, depending on the result of the syntactic analysis, the verb "go"
Against the noun “school” (where) and the pronoun “me”
You can see that (who) co-occurs. Thus, in the above example sentence, reading information such as "watashi / wa / qakkou / e / itta" is obtained.

For example, whether "I" is read as "watashi" or "watakushi" is determined in advance by priority setting or whether or not a honorific word is used at the end of the sentence by syntax analysis. And read "watakushi" when honorifics are used.

Further, whether or not to read symbols such as punctuation marks or parentheses can be designated by the user in advance by initial setting or the like. When the read-aloud data of the sentence is created, the read-aloud data creating unit 18 synthesizes the read-aloud data by voice and reads it out through the speaker 21 (step A16).

When the user hears the read sentence and feels that the sentence needs to be corrected, the input unit 12 gives a correction instruction. When the control unit 11 receives this correction instruction while reading the text (Y in step A17).
es), reading is stopped, and the edit mode is set (step A18).

By switching from the reading mode to the editing mode in this manner, the user can edit the text. Further, here, when the edit mode is set, the error detection unit 20 is activated and the following error detection processing is executed.

When the edit mode is set, the error detector 2
0 detects a sentence error using the above analysis result (step A19). In this case, in consideration of the delay when the trigger is applied in step A17 (when the user gives a correction instruction), at least two sentences, that is, the sentence read at that time and the sentence before that are targeted. Error detection. That is, in the example of FIG. 4, if a trigger is given when the sentence "I went to play with him." Is read aloud, "Wake up early and ate breakfast" before that. Error detection is performed including the sentence.

When the error detection result is obtained from the error detection unit 20, the control unit 11 displays it on the display unit 13 (step A20). In this example, a message stating that the part "I have him" in the sentence "I went to play with him" is incorrect is displayed in the message box 32 shown in FIG. The user can use this message box 3
The text can be corrected or edited by referring to the message displayed in 2 (step A21).

After that, when the processing is to be continued, the user gives an instruction to that effect through the input unit 12 (step A22).
No). As a result, the processing similar to the above is restarted from a sentence of which reading is not completed in the same document.

As described above, in the sequential edit type processing, the reading of the text can be stopped for each text and the edit mode can be entered. Therefore, when the user hears the read sentence and determines that the sentence should be corrected, the sentence can be corrected at that time. Also, at that time,
Since the error of the sentence is detected by the error detection and presented to the user, the user can easily correct the sentence by referring to it.

(B) Batch edit type FIG. 3 is a flowchart showing the batch edit type processing operation. When the user selects a sentence, the control unit 11 first sets the reading mode (step B1).
1). In this reading mode, the 1-sentence acquisition unit 16
Cuts out each sentence in the document stored in the document storage unit 14 one by one (step B12). The cut-out sentence is given to the display unit 13 and displayed on the screen (step B13).

In this case, a mark is added to the beginning and the end of the sentence, and the marked sentence, that is, the sentence that is currently read out is highlighted to distinguish it from other sentences, or the sentence is displayed on the screen. Performs display control so that it is centered. As a result, the user can always easily confirm the current target sentence.

The sentence cut out by the one-sentence acquisition unit 16 is also given to the sentence analysis unit 17. The sentence analysis unit 17 analyzes the sentence using the analysis dictionary stored in the analysis dictionary storage unit 15 (step B14). The reading data creating unit 18 creates reading data of the sentence based on the analysis result (step B1).
5).

When the reading data of the sentence is created, the reading data creating unit 18 synthesizes the reading data by voice and outputs it through the speaker 21 (step B16).

Here, when the user hears the sentence read aloud and feels that the sentence needs to be corrected, the input unit 12 gives a correction instruction. When the control unit 11 receives this correction instruction while reading the text (Y in step B17).
es), a marker is added to the target sentence (step B18).

In this case as well, in consideration of the delay when the trigger is applied (when the user gives a correction instruction), at least the sentence read at that time and the sentence before that are read.
A marker is added to one sentence. That is, in the example of FIG. 4, if a trigger is given when the sentence "I went to play with him." Is read aloud, "Wake up early and ate breakfast" before that. Add a marker including the sentence.

In this way, this input is accepted at any time while reading aloud, and at the time of input, a marker is added to the sentence currently being read and the sentence before it. The position of this marker is stored and referred to when the edit mode is entered.

However, when the user gives an instruction to finish reading through the input unit 12 (Yes in step B19).
s), the control unit 11 finishes reading and sets the edit mode (step B20).

By switching from the reading mode to the editing mode in this way, the user can edit the text. In addition, here, when the edit mode is set, a jump is made to a sentence with a marker (step B21), and the error detection unit 20 uses the above analysis result,
An error in the sentence with the marker is detected (step B
22). In the above example, a marker is added to the sentence "I went to play with him." And the sentence "I got up early in the morning and ate breakfast." Error detection is performed on the text.

When the error detection result is obtained from the error detection unit 20, the control unit 11 displays it on the display unit 13 (step B23). In this example, a message stating that the part "I have him" in the sentence "I went to play with him" is incorrect is displayed in the message box 32 shown in FIG. The user can use this message box 3
By referring to the message displayed in No. 2, the sentence can be corrected or edited (step B24).

After that, when the processing is to be continued, the user gives an instruction to that effect through the input unit 12 (step B25).
No). This jumps to the next sentence with a marker in the same document. If all the markers are referenced, the process ends.

As described above, in the batch edit type processing, when the reading of the text is completed, the correction target texts having the markers can be collectively corrected. Therefore, it is possible to memorize a sentence that the user has decided to correct without interrupting the reading of the sentence, and make a correction in consideration of the balance of the whole sentence after the reading of one sentence is finished. be able to. At that time, an error in the sentence is detected by the error detection and presented to the user, so that the user can easily correct the sentence by referring to the error.

[0042]

As described above, according to the sequential edit type processing of the present invention, it is possible to stop the reading of a sentence in units of one sentence and shift to the edit mode. When the user determines that it should be done, the sentence can be corrected at that time. Further, at this time, by presenting the error of the sentence to the user by error detection, the user can easily correct the sentence by referring to the error.

Further, according to the batch edit type process, when the reading of the text is finished, the correction target texts with the markers can be corrected in a batch, so that the text reading is not interrupted. It is possible to store a sentence that the user has decided to correct, and to make a correction in consideration of the balance of the entire sentence after reading a series of sentences is completed. Further, at this time, by presenting the error of the sentence to the user by error detection, the user can easily correct the sentence by referring to the error.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a text selection support device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a sequential edit type processing operation.

FIG. 3 is a flowchart showing a batch edit type processing operation.

FIG. 4 is a diagram showing a display screen when reading a sentence.

FIG. 5 is a diagram for explaining a process until obtaining reading information of a sentence.

[Explanation of symbols]

 11 ... Control unit 12 ... Input unit 13 ... Display unit 14 ... Document storage unit 15 ... Analysis dictionary storage unit 16 ... 1 sentence acquisition unit 17 ... sentence analysis unit 18 ... reading data creation unit 19 ... voice synthesis unit 20 ... error detection unit 21 ... Speaker 31 ... Sentences to be read 32 ... Message box

Claims (8)

[Claims] 1. A document storage unit for storing document information composed of a plurality of sentences, a sentence acquisition unit for sequentially cutting out sentences in predetermined units from the document information stored in the document storage unit, and the sentence acquisition unit. Text analysis means for analyzing the cut out text, reading data creation means for creating reading data for the text based on the analysis result of the text analysis means, and reading data for the text created by the reading data creation means A voice synthesizing means for reading aloud by voice synthesis, an instructing means for instructing text editing, and when the text editing is instructed by the above-mentioned instructing means during the text reading, the edit mode is set so that the text at that time can be sequentially edited. A text revision support device comprising an editing control means. 2. When the editing mode is set by the editing control means, an error detection means is provided for detecting an error in the sentence using the analysis result of the sentence analysis means and presenting the detection result. The sentence revision support device according to claim 1. 3. A document storage unit for storing document information composed of a plurality of sentences, a sentence acquisition unit for sequentially cutting out sentences in predetermined units from the document information stored in the document storage unit, and the sentence acquisition unit. Text analysis means for analyzing the cut out text, reading data creation means for creating reading data for the text based on the analysis result of the text analysis means, and reading data for the text created by the reading data creation means A voice synthesizing means for reading aloud by voice synthesis, an instructing means for instructing text editing, and when a text editing is instructed by the above instructing means during text reading, a marker is added to the text at that time,
A text revision support device comprising: an edit control unit that sets an edit mode at a predetermined timing so that the texts to which the markers are added can be collectively edited. 4. An error detecting means for detecting an error in each sentence using the analysis result of the sentence analyzing means when the edit mode is set by the edit control means and presenting the detection result. The sentence revision support device according to claim 3, characterized in that. 5. A sentence is sequentially cut out from document information consisting of a plurality of sentences in a predetermined unit, the cut-out sentence is analyzed to create reading data, and the reading data of the created sentence is voice-synthesized. A text revision support method characterized in that when a text edit is instructed during reading, the edit mode is set so that the text can be sequentially edited. 6. The sentence revision support according to claim 5, wherein when the edit mode is set, an error in the sentence is detected using the analysis result and the detection result is presented. Method. 7. A sentence is sequentially cut out from document information consisting of a plurality of sentences in a predetermined unit, the cut-out sentence is analyzed to create reading data, and the reading data of the created sentence is voice-synthesized. When a text edit is instructed during the reading, a marker is added to the text at that time, the edit mode is set at a predetermined timing, and each text to which the above marker is added is added later. A text revision support method characterized by enabling batch editing. 8. The sentence according to claim 7, wherein when the edit mode is set, an error in each sentence is detected using the analysis result and the detection result is presented. Elaboration support method.