JPH0991296A - Machine translation system and translation result output method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate and output a document in a conversational form in the range required by a user without destroying the essential forms of an original document and a translation result document at all. SOLUTION: When a translation processing control part 63 executes translation while dividing the original document stored in an original document storage buffer 51 into sentences, position information of a sentence as the translation object in the original document is written in a pertinent column of an original/translation result correspondence table buffer 53 at each time of translation, and position information of the sentence as the translation result is written in a pertinent column of the correspondence table buffer 53 at each time of writing this sentence as the translation result in a translation result document storage buffer 52; and at the time of display in a conversational form, a display processing control part 61 successively segments corresponding sentences in an externally designated range or in all of the document from the original document and the translation result document in accordance with registered contents of the correspondence table buffer 53 and sets them to a buffer 54 for picture display while pairing them to display them in the conversational form.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a machine translation device and a translation result output method for translating a document to be translated while dividing it into partial units.

[0002]

2. Description of the Related Art Conventionally, two methods, a batch translation method and a sequential translation method, are known as translation methods in a machine translation device. In either method, the first language document (translation source document) to be translated is a part (hereinafter referred to as a sentence).
It is the same in that it is divided into units and translated into the target second language.

In this type of machine translation device, the user can check the translation result while comparing it with the original document.
It is common to display the source document and the translated document, for example, by arranging them on the left and right sides or the upper and lower sides of the screen. However, with such a display format (output format), it is difficult to check the translation result by sentence, such as which sentence in the source document is translated and how. It was

Therefore, in recent years, when translating a source document of a first language to be translated into a second language while dividing the original document into sentence units, a sentence number (translation number) is attached to the corresponding sentence and bilingual translation is performed. In addition to translating into an expression, the same sentence number is also attached to the sentence of the translation result, and after the translation is finished, while removing the sentence number, the pair of the source sentence and the sentence of the translation result of the same sentence number (consisting bilingual sentence) A machine translation device has been developed which can output texts in a bilingual format, for example, can be displayed on the screen by connecting the texts in the order of sentence numbers.

However, with this method, once the sentence numbers in the document have been removed to make it a normal document, it is not possible to create and output a sentence in the bilingual form again. On the other hand, if sentence numbers are left in the source document and the translated document, it is possible to create and output (eg, display, print, or save to another file) bilingual text at any time. However, the translation source document and the translation result document are different from the original image (shape) of the original document, and when these documents are displayed or printed, the sentence number unnecessary for the user is displayed. There was a problem that the presence of was disturbing the appearance.

In addition, in the conventional machine translation device, when it is desired to display (print) differently from the translation source sentence, in order to make the displayed (printed) parallel translation easier for the user to see, the translation is automatically performed. No means existed, and the user had to manually perform the operation for each sentence.

Further, in the conventional machine translation device, even if the sentence of the translation source document and the document of the translation result are left attached with the sentence numbers, the translation source document can be output, for example. When editing operations such as correction (correction), deletion, and addition (insertion) were performed on, the bilingual sentence could not be output thereafter. In other words, in order to be able to output bilingual text, not only leave the sentence numbers in the original document and the translated document, but also edit operations such as correction, deletion, and addition. Had to ban.

[0008]

As described above, in the conventional machine translation device, in order to create and output a bilingual sentence based on the source document and the translated result document, the sentence (part) unit is used. It is necessary to process the document itself, such as by adding consecutive sentence numbers with, and without destroying the original shape of the source document and the translated document at all, and for users like sentence numbers. There is a problem in that it is not possible to create and output an interactive sentence of a portion required by the user without displaying unnecessary information.

Further, in the conventional machine translation device, there is a problem that the translation source sentence and the translation result sentence in the bilingual sentence cannot be automatically displayed or printed in different display formats or print formats.

Further, in the conventional machine translation apparatus, even if sentence numbers are left in the source document and the translated result document, editing operations such as correction, deletion, addition, etc. are performed on the source document. After that, there was also a problem that it was not possible to create and output an interactive document.

The present invention has been made in consideration of the above circumstances, and an object thereof is an interactive format within a range required by a user without destroying the original shapes of a source document and a translated document. To provide a machine translation device and a translation result output method capable of creating and outputting a document.

Another object of the present invention is to provide a simple operation.
It is an object of the present invention to provide a machine translation device and a translation result output method capable of displaying or printing a translation source sentence and a translation result sentence in a bilingual sentence in a display format or a printing format desired by a user.

Still another object of the present invention is to provide a machine translation device and a translation device capable of creating and outputting an interactive document even after performing editing operations such as correction, deletion and addition on a source document. It is to provide a result output method.

[0014]

A machine translation apparatus according to a first aspect of the present invention divides a source document in a first language to be translated into partial units and translates them into a second language. A machine translation apparatus including translation processing means for acquiring a translation result document, wherein a translation source for storing a translation source part and position information on a corresponding document of the translation result in partial units When the translation result correspondence table buffer means and the translation processing means perform translation on a sub-unit basis, a set of position information on a corresponding document of a translation source part and its translation result is associated with that part. The above translation source /
Correspondence within the range externally designated from the translation source document and the translation result document or in the entire document based on the registration means registered in the translation result correspondence table buffer means and the registered contents of the translation source / translation result correspondence table buffer means. It is characterized by further comprising a translation result output means for creating and outputting a bilingual document by sequentially cutting out the portions to be processed and forming a pair.

Next, a machine translation apparatus according to a second aspect of the present invention has the same structure as the machine translation apparatus according to the first aspect.
When the output of bilingual format documents is displayed or printed, an output format specification accepting means that accepts external specifications about the display format or print format for each source document and translation result document is added. According to the display format or print format designation accepted by the means, the translation result output means creates and displays or prints out a bilingual format document.

Next, a machine translation apparatus according to a third aspect of the present invention is the same as the machine translation apparatus according to the first or second aspect, except that the source document is edited from outside the editing position. Edit position designation accepting means for accepting the designation in units of parts registered in the translation source / translation result correspondence table buffer means, and a designated portion of the translation source document accepted by this acceptance means and a corresponding portion of the translation result document. On the other hand, when the editing processing means performs an externally specified editing operation and the translated portion of the translation source document needs to be translated, the translation processing means performs translation, and the translation result sentence is translated. Is used to edit the corresponding portion of the translation result document, and the positional relationship between the corresponding partial units of the translation source document and the translation result document after the editing operation by this editing means is reflected. Characterized in that adding and updating means for updating the contents of the translation source / translation result correspondence table buffer means.

In the machine translation device according to the first aspect, the position information on the document of each part of the translation result document corresponding to each part of the source document (for example, the start character position and the end character position of the corresponding part are The position information (shown) is registered in the translation source / translation result correspondence table buffer means as a set corresponding to that part in units of parts. Therefore, based on the registered contents of the translation source / translation result correspondence table buffer means, corresponding portions in the externally designated range or in the entire document are sequentially cut out from the translation source document and the translation result document, and are translated into pairs. By creating and outputting this document, the source document and the translation result document are not processed at all (that is, the original image of the document is not destroyed at all), and it is not convenient for the user such as the sentence number used in the past. Without displaying the necessary information, it is possible to obtain the bilingual format document in the range required by the user (or the entire document) and display, print, or save (as an independent document) output.

In the machine translation apparatus according to the second aspect, the bilingual format document is simply designated by specifying the display format or printing format of the bilingual format document in each document unit of the translation source document and the translation result document. Since the translation source part and the translation result part of each of the bilingual sentences constituting the are displayed or printed in the specified display format or print format, respectively.
This is convenient when it is desired to display or print in a different display format or print format so that the translation source and the translation result can be clearly identified (and one is treated as a quotation).

Next, in the machine translation apparatus according to the third aspect, after the translation process is completed, the translation operation is performed on the translation source document in partial units registered in the translation source / translation result correspondence table buffer means. Then, the editing operation is automatically performed on the corresponding part of the translation result document, and the positional relationship of the corresponding partial units of the translation source document and the translation result document after the editing operation is reflected. The registered contents of the translation source / translation result correspondence table buffer means are updated. Therefore, even if an editing operation such as partial insertion, deletion, or correction is performed on the translation source document after the translation processing, based on the updated content registered in the translation source / translation result correspondence table buffer means, It is possible to output the translation source document and the translation result document after the editing process in a bilingual format.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a machine translation device according to an embodiment of the present invention.

The machine translation apparatus of FIG. 1 mainly includes an input unit 1, an external storage device 2, a display unit 3, a printing unit 4, a RAM (Ra
ndom Access Memory) 5 and control device 6. The input unit 1 is for inputting information intended by a user according to a user operation, and is a keyboard, a tablet and a pen, or an OCR (Optical Character Reader).
It is composed of an input device such as.

The external storage device 2 is capable of recording (writing) / reproducing (reading) information such as documents, and is nonvolatile such as FDD (floppy disk device), HDD (hard disk device), MD (mini disk device), etc. It is a storage device.

The display unit 3 displays a translation source document, a translation result document,
Color CR used for displaying bilingual documents
The display device is a T (Cathode Ray Tube) display, a color liquid crystal display, or the like.

The printing unit 4 includes a translation source document, a translation result document,
A printer device such as a thermal transfer type color printer or an ink jet type color printer, which is used for displaying bilingual documents and the like.

The RAM 5 is a readable / writable memory. A translation source document storage buffer 51 for storing a translation source document (translation source document), a translation result document storage buffer 52 for writing a document composed of translated sentences (translation result document), and a RAM 5, A translation source for storing the position of the translation source sentence and its translation result (sentence) on the corresponding document in association with each other, for example, by a sentence start position and a sentence end position.
The translation result correspondence table buffer 53 is secured.

The RAM 5 also has a screen display buffer 54 for writing display data to be displayed on the display unit 3, a print buffer 55 for writing print data to be output to the printing unit 4, and a screen display designated by the user. Display setting information storage area 56 for storing various setting information for printing, and print setting information storage area 5 for storing various setting information for screen display designated by the user
7 is secured.

In the present embodiment, in the display setting information storage area 56, when the bilingual format document is displayed, the display setting information (here, the display format is independently specified for each of the translation source and the translation result). , Setting information that specifies the type of font to use, color to use, etc. is set,
In the print setting information storage area 57, when a bilingual document is printed, for each of the translation source and the translation result,
Print setting information for which the print format is specified independently (here,
Setting information that specifies the type of font used, the color used, etc. is set.

The control device 6 is responsible for main processing control in the machine translation device. For example, a ROM (Read Only Memor) in which various processing control programs, dictionaries, etc. are stored in advance.
y) and the CPU that executes the program in the ROM.

The control device 6 controls the display processing control unit 61 and the print buffer 55 for displaying the contents of the screen display buffer 54 on the display unit 3 according to the setting information in the display setting information storage area 56. It has a print processing control unit 62 for printing the content by the printing unit 4 according to the setting information in the print setting information storage area 57.

The control device 6 also sets the translation source document of the first language stored in the translation source document storage buffer 51 as the first translation source document.
Translation processing control unit 63 for translating into a second language different from the above language, and well-known dictionary information (words, parts of speech, translation rules, etc.) used in the translation processing by the translation processing control unit 63 are registered. It has an existing translation dictionary 64 and an edit processing control unit 65 for performing edit processing such as insertion (addition), deletion, and correction on a translation source document and a translation result document in accordance with a user operation.

Next, the operation of the machine translation device configured as shown in FIG. 1 will be described. First, a case where the translation processing control unit 63 performs translation on, for example, the entire sentence of the translation source document designated by the user will be described.

Now, the translation source document in the first language designated by the user, for example, the translation source document in English, is read from the external storage device 2 by R
After being read into the translation source document storage buffer 51 in the AM 5, the screen display data of this translation source document is created and written in the screen display buffer 54, and the screen display data is output to the display unit 3 for screen display. Be present. Here, the user operates the input unit 1 to specify, for example, a range to be translated on the screen, and the translation processing control unit 63 in the control device 6 recognizes the translation range specified by the user. In addition, when the whole sentence is set as the translation range, or when the range is specified page by page, it is possible to specify directly using a menu or the like.

Here, it is assumed that the whole sentence is designated as the translation range. When the translation range is specified by the user operation, the translation processing control unit 63 starts the translation processing described below according to the flowchart of FIG.

First, the translation processing control unit 63 checks whether or not there is an untranslated portion within the designated range (here, the entire document) of the translation source document stored in the translation source document storage buffer 51. (Step S1).

If there is a portion that has not been translated yet, the translation processing control unit 63 first cuts out the maximum translatable portion from the untranslated portion in the translation source document as a translation target sentence and translates it. It is translated into a second language, for example, Japanese with reference to the dictionary 64 (step S2).

In the example of the translation target document (translation source document) shown in FIG. 3 (a), "It's lovely t" is selected from among a plurality of sentences.
"be going home and know it's home," was cut out as the first sentence, and as shown in Figure 3 (b), "going home,
And it's great to know that it's a home. Is translated as ".

The translation process itself for translating the source document into the target language while dividing the source document into sentences (parts) can be realized by a well-known technique. When the translation process control unit 63 completes the translation of one sentence (one portion) in step S2, the translation process control unit 63 determines that the first character of the sentence to be translated (here, the first sentence) in the source document is translated. The position and the position of the last character are written in the translation source / translation result correspondence table buffer 53 as the translation source / sentence start position and the translation source / sentence end position, respectively (step S3).

Here, the translation source / translation result correspondence table buffer 53 has a tabular data structure, as shown in FIG. 3C, and has a translation source / sentence start position and a translation source / sentence end position. The translation result / sentence start position and the translation result / sentence end position have columns (fields) set for each corresponding sentence.

Therefore, in the example of FIG. 3, in the first processing of step S3, "1 line 3 in the translation source / sentence start position" column of the first sentence of the translation source / translation result correspondence table buffer 53 is displayed.
Both data of "column" and "translation source and sentence end position" are written in "2 lines, 15 columns".

Upon executing step S3, the translation processing control section 63 writes the translation result obtained in step S2 in the translation result document storage buffer 52, and further writes the first character (in the translation result document) of the written translation result sentence. The position and the position of the last character are written in the translation source / translation result correspondence table buffer 53 as the translation result / sentence start position and the translation result / sentence end position, respectively (step S4). Here, both "1 line 1 digit" in the translation result / sentence start position column of the first sentence of the translation source / translation result correspondence table buffer 53 and "2 lines 16 digits" in the translation result / sentence end position column. Data is written.

The translation processing control unit 63 repeats the above steps S1 to S4, and in step S1 there is no untranslated portion, that is, the translation of the range specified by the user (here, the whole sentence) is completed. When it is detected, a series of translation processing ends. At this time, in the translation result document storage buffer 52, the translation result document as shown in FIG. 3B, which is the translation result of the translation source document in FIG. 3A, is written.

Next, based on the translation source document storage buffer 51, the translation result document storage buffer 52, and the translation source / translation result correspondence table buffer 53, the display processing control unit 61 displays the bilingual format screen display data on the screen. A case of creating in the buffer 54 for screen display will be described with reference to the flowchart of FIG.

First, the display processing control unit 61 allows the user to specify the range in which the bilingual document should be displayed by operating the input unit 1, and accepts the range specification (step S11). When the display processing control unit 61 receives the range designation designated by the user, it refers to the translation source / translation result correspondence table buffer 53 to check which position the designated range (here, the entire sentence) corresponds to (step S12).

The display processing control section 61 always creates the screen display buffer 54 in the screen display buffer 54 in parallel display format before displaying the screen display buffer in the range checked in step S12 (user's It is checked whether or not the screen display buffer has been created (for the specified range) (step S13).

If the creation has not been completed yet, the display processing control unit 61 performs the following processing. First, the display processing control unit 61 takes out the setting information of the translation source / sentence start position and the translation source / sentence end position of the first sentence from the translation source / translation result correspondence table buffer 53 (in the case of the first processing), and The contents of the translation source document in the portion designated by the position information are set in the screen display buffer 54 from the beginning (in the case of the first processing) (step S14).

In the example of FIG. 3, from the translation source / translation result correspondence table buffer 53, "1 line 3 digits" is set as the translation source / sentence start position of the first sentence, and "2 lines 15" is set as the translation source / sentence end position.
Digit's data is taken out. Then, from the translation source document storage buffer 51, starting from this “1 line 3 digits”,
"It's lovely to be going hom"
"eand know it's home," is taken out and set in the screen display buffer 54.

Next, the display processing control unit 61 takes out the setting information of the translation result / sentence start position and the translation result / sentence end position of the first sentence from the translation source / translation result correspondence table buffer 53 (1
In the case of the process of the second time), the content of the translation result document of the portion designated by the position information is set from the next line of the screen display buffer 54 set in step S14 (step S15).

In the example of FIG. 3, from the translation source / translation result correspondence table buffer 53, "1 line 1 digit" is set as the translation result / sentence start position of the first sentence, and "2 lines 16" is set as the translation source / sentence end position.
Digit's data is taken out. Then, from the translation result document storage buffer 52, starting from this "1 line 1 digit",
The sentence "It's wonderful to go home and know that it is a home." That ends with "2 lines, 16 digits" is retrieved, and "It's lovely to" previously set in the screen display buffer 54. be going home and know it's
Set from the line following home, ".

Next, the display processing control section 61 determines in step S1.
A blank line (space line) is set to the next line of the screen display buffer 54 set in step 5 (step S16). The display processing control unit 61 repeats the above steps S13 to S16. In the i-th (i is an integer of 2 or more) step S14, the translation source / translation start / end position setting information of the i-th sentence in the translation source / translation result correspondence table buffer 53 is extracted, and the position The content of the translation source document in the portion designated by the information is set from the line next to the blank line of the screen display buffer 54 set in the previous step S16.
Similarly, the i-th (i is an integer of 2 or more) step S15
In the process of, the i of the translation source / translation result correspondence table buffer 53
The setting information of the translation result / sentence start / end position of the sentence is taken out.

The display processing controller 61 starts from step S13.
When S16 is repeated and it is detected in step S13 that the creation of the screen display buffer for the range designated by the user is completed, a series of screen display buffer creation processing is ended, and the data in the screen display buffer 54 at that time is displayed. It is displayed on the section 3 and the process is terminated (step S1).
7).

Thus, in the example of FIG. 3, the translation source document screen of FIG. 3A and the translation result document of FIG. 3B are displayed interactively as shown in FIG. The result is obtained. Here, a blank line 51 (following step S16) is set between adjacent bilingual sentences so that each bilingual sentence can be easily distinguished.

In addition to the above-mentioned bilingual document display, the translation source document stored in the translation source document storage buffer 51 and the translation result document stored in the translation result document storage buffer 52 are individually displayed. It is naturally possible to display both documents by arranging them on the left and right sides of the screen, for example.

Next, based on the translation source document storage buffer 51, the translation result document storage buffer 52, and the translation source / translation result correspondence table buffer 53, bilingual print data is created in the print buffer 55 and printed. Figure 6
This will be described with reference to the flowchart in FIG.

First, the print processing control unit 62 allows the user to specify the range in which the bilingual document should be printed, and also prints the font type and color of the translation source document (sentence) and the translation result document (sentence). A print condition for determining a print format, such as a font type and a color to be used, is designated, and the designation is accepted (step S21).

When the print processing control unit 62 receives the range designation by the user operation and the print condition designation of each of the translation source document and the translation result document, the print setting information storing area 57 stores the print setting information in which the designated printing condition is set. And the translation source / translation result correspondence table buffer 53 to check which position the specified range (here, the whole sentence) corresponds to (step S22).

Before printing the bilingual format document, the print processing control unit 62 always checks whether or not the printing for the range checked in step S22 (for the range specified by the user) has been completed (step S23). ).

If the printing is not completed yet, the print processing control unit 62 carries out the following processing. First, the print processing control unit 62 takes out the setting information of the translation source / sentence start position and the translation source / sentence end position of the first sentence from the translation source / translation result correspondence table buffer 53 (in the case of the first processing), and The buffer 5 for printing prints the contents of the translation source document specified by the position information.
5 is set from the beginning (after clearing the previous information), and the printing unit 4 prints with the user-specified font and color of the source document indicated by the print setting information set in the print setting information storage area 57. (Step S
24).

In the example of FIG. 3, from the translation source / translation result correspondence table buffer 53, "1 line 3 digits" is set as the translation source / sentence start position of the first sentence, and "2 lines 15" is set as the translation source / sentence end position.
Digit's data is taken out. Then, from the translation source document storage buffer 51, starting from this “1 line 3 digits”,
"It's lovely to be going hom"
"eand know it's home," is taken out, set in the print buffer 55, and printed in red using, for example, a Gothic font.

Next, the print processing control unit 62 takes out the setting information of the translation result / sentence start position and the translation result / sentence end position of the first sentence from the translation source / translation result correspondence table buffer 53 (1
In the case of the second processing), the contents of the translation result document in the portion designated by the position information are set in the print buffer 55 from the beginning (after the previous information is cleared), and set in the print setting information storage area 57. The printing unit 4 prints with the font and color specified by the user for the translation result document indicated by the print setting information (step S25).

In the example of FIG. 3, from the translation source / translation result correspondence table buffer 53, the translation result / sentence start position of the first sentence is "1 line 1 digit", and the translation source / sentence end position is "2 lines 16".
Digit's data is taken out. Then, from the translation result document storage buffer 52, starting from this "1 line 1 digit",
The sentence ending in "2 lines 16 digits""It is wonderful to go home and know that it is a home." Is taken out and set in the print buffer 55, and the original sentence that was printed earlier Is printed in blue from the next line of, using, for example, a brush font.

The print processing control unit 62 executes step S24,
When a pair of bilingual forms (one bilingual sentence) can be printed in S25, the printing unit 4 is controlled to feed one line of paper,
One blank line is secured (step S26).

The print processing controller 62 starts from step S23.
S26 is repeated, and when it is detected in step S23 that printing for the range designated by the user is completed, a series of printing processes is completed.

As a result, in the example of FIG. 3, the translated document print result as shown in FIG. 7 in which the translation source document of FIG. 3A and the translation result document of FIG. 3B are interactively printed. Is obtained. Here, since the translation source sentence and the translation result sentence have different fonts and colors, the user can clearly identify the translation source and the translation result in the bilingual sentence. Moreover, the print format is specified for each document of the translation source document and the translation result document.
The user operation is not complicated because it is not specified for each sentence.

Even when the bilingual document is displayed on the screen, the font type, color, and translation result document (for the translation source document) for displaying the translation source document Text type, font type, color, etc.
By allowing the user to specify the display condition for determining the display format for each document and storing the display setting information in which the specified display condition is set in the display setting information storage area 56, the translation source sentence and its translation It is also possible to display the resulting sentence and the screen in the font type and color that the user has separately designated.

Also, the font type as in the above example,
In addition to the color, of the pair of (translated sentences) of the sentence of the translation source document and the sentence of the translation result document, which side of the translation source and the translation result is to be placed first, whether to insert a line feed, It is also possible to specify whether or not to attach a parenthesis to be treated as a quotation to either one, which side it is attached to, and the type of parentheses to be attached. Accordingly, it is possible to create a parallel translation document in which the original text is quoted in the translation result and a parallel translation document in which the translation result is quoted in the original text.

Next, a description will be given of a case where the translation source document stored in the translation source document storage buffer 51 is subjected to editing operations such as insertion, deletion and correction. First, when an editing operation is performed on a translation source document stored in the translation source document storage buffer 51, the translation source document and the translation result document (stored in the translation result document storage buffer 52) are displayed on the screen. The data is displayed on the display unit 3 through the display buffer 54 (for example, arranged on the left and right sides of the screen). In this state, the edit processing control unit 65 controls the input unit 1 in units of sentences registered in the translation source / translation result correspondence table buffer 53.
It is possible to accept the user's position designation using.

When the position designation is accepted, the user inserts a new English (first language) sentence immediately before the sentence at the designated position, deletes the sentence at the designated position, and deletes the designated position. Operations for correcting sentences can be selectively performed. In the present embodiment, the input section 1 is provided with a [delete] key, an [insert] key, and a [correction] key. When the [delete] key is pressed after the above-mentioned position designation, the sentence is deleted by [insert]. You can insert a sentence by pressing the key and correct it by pressing the [Correct] key.

First, the insertion of a sentence will be described. When the user wants to insert a desired sentence in the translation source document, the user specifies the insertion position (the position of the sentence next to the sentence) and presses the [Insert] key, and then the sentence to be inserted (here, the first sentence). (English sentence as the language) is input from the input unit 1.

Then, the edit processing control unit 65 inserts the input sentence immediately before the designated sentence in the translation source document. When inserting this new sentence, the sentences after the specified position are shifted later by the length of the new sentence.

When the translation source sentence is inserted by the edit processing control unit 65, the translation processing control unit 63 automatically translates the inserted sentence and translates the translated sentence (translated sentence). Insert in the result document. The insertion position of this translation is determined from the specified position and the translation source / translation result correspondence table buffer 53. For example, when the designated position is the second sentence in the example of FIG. 3, a new translated sentence is inserted immediately before the second sentence in the translation result document, that is, from the second line and the 17th digit.

When a new translation sentence is inserted in the translation result document, the edit processing control section 65 uses the translation source / translation result correspondence table buffer 53 for each sentence corresponding to the current translation source document and the translation result document. Update so that the positional relationship is reflected.

Next, the deletion of sentences will be described. When the user wants to delete a desired sentence in the translation source document, the user specifies the position of the sentence to be deleted and presses the [Delete] key. Then, the edit processing control unit 65 deletes the sentence at the designated position from the translation source document. When deleting this sentence, the succeeding sentence is filled up by the length of the sentence. At the same time, the edit processing control unit 65 deletes the sentence corresponding to the designated position from the translation result document. For example, in the example of FIG. 3, if the designated position is the second sentence, the second sentence in each of the translation source document and the translation result document is deleted and the succeeding sentence is padded forward, and the third sentence before deletion is The second sentence, the fourth sentence, the third sentence, and so on.

When the edit processing control unit 65 executes the sentence deletion from the translation source document and the translation result document, the edit source control unit 65 stores the translation source / translation result correspondence table buffer 53 for each sentence corresponding to the current translation source document and the translation result document. Update so that the positional relationship is reflected.

Next, the correction of a sentence will be described. When the user wants to correct a desired sentence in the translation source document, he or she specifies the correction target sentence position and presses the [Correction] key, and then performs a correction operation on the sentence at the specified position from the input unit 1.

Then, the edit processing control section 65 corrects the sentence at the designated position in the translation source document in accordance with the correction operation by the user. At this time, if the number of characters in the uncorrected sentence is larger than that in the corrected sentence, the sentence after the difference is moved forward, and in the opposite case, the sentence after the difference is moved backward. .

When the editing process control unit 65 corrects the sentence at the specified position, the translation process control unit 63 automatically translates the corrected sentence and converts the corresponding sentence in the translation result document to the corresponding sentence. Correct the translation (corrected translation). The correction target sentence position in the translation result document is determined from the designated position and the translation source / translation result correspondence table buffer 53. For example, when the designated position is the second sentence, in the example of FIG. 3, the sentence from the second line, 17 digits to the third line, 6 digits in the translation result document is replaced with the corrected translation sentence. At this time, the translated sentences from the third sentence onward are moved forward or backward by the difference in the number of characters between the uncorrected translated sentence and the corrected translated sentence.

When the translation result document is corrected, the edit processing control section 65 makes a translation source / translation result correspondence table buffer 53.
Is updated so as to reflect the positional relationship between the corresponding sentences of the current translation source document and the translation result document.

Next, the details of the updating process of the translation source / translation result correspondence table buffer 53 accompanying the above editing process will be described with reference to the flowchart of FIG. 8 taking the case of sentence deletion as an example.

Here, the sentence 91 (second sentence) in the translation source document shown in FIG. 9A stored in the translation source document storage buffer 51 (corresponding to FIG. 3A) is deleted, In response to this, the sentence 92 (second sentence) in the translation result document shown in FIG. 9B stored in the translation result document storage buffer 52 (corresponding to FIG. 3B) is deleted. Figure 9
It is assumed that it becomes like (c) and (d). At this time, the contents of the translation source / translation result correspondence table buffer 53 are as shown in FIG.
It is as shown in FIG. 10 (a) (corresponding to (c)).

In this case, the edit processing control unit 65 checks the number of the deleted sentence from the specified position and the translation source / translation result correspondence table buffer 53, and the data corresponding to the sentence position is set. All the information in the correspondence table buffer 53 (here, the translation source / sentence start / end position and translation result / sentence start / end position setting values of the third sentence) are cleared (step S31).

Next, the edit processing control section 65 carries out the following processing for the sentence following the deleted sentence to the last sentence. First, the edit processing control unit 65, for the translation source document, the "translation source / sentence start position" in the translation source / translation result correspondence table buffer 53 from the sentence next to the deleted sentence to the last sentence.
Each setting value of "translation source / sentence end position" is reduced by the number of characters of the deleted translation source sentence (step S32).

Similarly, for the translation result document, the edit processing control unit 65 also selects the "translation result / translation result / sentence" in the translation source / translation result correspondence table buffer 53 from the sentence following the deleted sentence to the last sentence. The setting values of "sentence start position" and "translation result / sentence end position" are reduced by the number of characters of the deleted translation result sentence (step S33).

Finally, the edit processing control section 65 sets the "translation source / translation source / translation source / translation result correspondence table buffer 53, which has been set in the corresponding column in the translation source / translation result correspondence table buffer 53, from the sentence following the deleted sentence to the last sentence. "Sentence start / end position" and "Translation result / Sentence start /
The data of "end position" is preceded by one sentence (that is, 1
(To the previous column) (step S34).

The above translation source / translation result correspondence table buffer 5
A specific example of the updating process of No. 3 will be described with reference to FIG. 9 when the second sentence, that is, the sentences 91 and 92 are deleted. In this case, first, the translation source / sentence start position = 2 lines and 17 digits, the translation source / sentence end position = 2 lines and 25 digits, registered as the second sentence data in the translation source / translation result correspondence table buffer 53,
Each data of the translation result / sentence start position = 2 lines and 17 digits and the translation result / sentence end position = 3 lines and 6 digits is cleared (step S31).

Next, since the translation source document has been deleted by 10 characters, the data of the third sentence in the translation source / translation result correspondence table buffer 53 is translated source / sentence start position = 2 lines, 27 digits. → 2 lines, 17 digits, translation source / end of sentence =
3 rows and 17 digits → amended to 3 rows and 7 digits (step S
32). In this step S32, the same data correction is carried out for the subsequent fourth character sentence to the last sentence.

Since the translation result document has been deleted by 8 characters, the data of the third sentence in the translation source / translation result correspondence table buffer 53 is translated result / sentence start position = 3 lines, 7 digits. -> 2 lines, 17 digits, translation result / sentence end position = 4 lines, 11 digits → 4 lines, 3 digits (step S33). In this step S33, the same data correction is carried out for the subsequent fourth character sentence to the last sentence.

Finally, the data in the third character column of the translation source / translation result correspondence table buffer 53 subjected to the processing in steps S31 to S33 is transferred to the second character column in the fourth character column data. Is moved to the front of the third character by one sentence, for example (step S34).

In this way, the translation source / translation result correspondence table buffer 53 is updated from the state of FIG. 10A to the state of FIG. 10B so that the result of the editing process designated by the user is reflected. To be done. Then, thereafter, the translation source document storage buffer 51 (the edited translation source document therein) and the translation result document storage buffer 52 are subjected to the procedure as described above.
(The translation result document after editing in), and the translation source after updating /
Based on the translation result correspondence table buffer 53, it is possible to display and print a document in a bilingual format reflecting the editing result.

The translation source / translation result correspondence table buffer 5
The contents of 3 (translation source / translation result correspondence table) are saved as one document in the external storage device 2 together with the translation source document in the translation source document storage buffer 51 and the translation result document in the translation result document storage buffer 52. In addition, by calling the translation source document and the translation result document into the RAM 5 as necessary,
It is possible to perform the above-described editing operation at the desired time and display / print the document in the parallel translation format after the editing operation. Further, by linking the related translation source document, the translation result document, and the translation source / translation result correspondence table (documents), for example, only by designating the translation source document, the corresponding translation result document and the translation source / translation source / It is also possible to call the translation result correspondence table together.

Although the case of Japanese-to-English translation has been described above, it is needless to say that it can be generally applied to the case of translating an arbitrary first language into another second language, such as the case of English-Japanese translation.

[0091]

As described above in detail, according to the present invention, the original shape of the original document and the original document of the translation result is destroyed, and information unnecessary for the user such as the sentence number is added to the document (in It is possible to display or print an interactive document in the range required by the user without adding it to the beginning of each sentence.

Further, according to the present invention, the translation source sentence and the translation result sentence in the bilingual sentence can be displayed in a display format desired by the user by a simple designation operation for each document unit of the translation source document and the translation result document. It can be displayed or printed in print format.

Further, according to the present invention, the translation source document and the translation result document after the editing process are performed even after the editing operation such as insertion, deletion, correction, etc. is performed on the translation source document in sentence units (partial units). Can be output in a bilingual format.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a machine translation device according to an embodiment of the present invention.

FIG. 2 is a flowchart for explaining an operation during translation processing according to the same embodiment.

FIG. 3 is a diagram showing an example of the contents of a translation source / translation result correspondence table buffer 53 after translation processing in comparison with the corresponding translation source document and translation result document.

FIG. 4 is a flowchart for explaining an operation when a translation source and a translation result are displayed in a bilingual format in the same embodiment.

FIG. 5 is a diagram showing an example of a screen display in a parallel translation format according to the same embodiment.

FIG. 6 is a flowchart for explaining an operation when a translation source and a translation result are printed in a bilingual format using different fonts (and colors) in the embodiment.

FIG. 7 is a diagram showing an example of a print result in a bilingual format according to the first embodiment.

FIG. 8 is a flowchart for explaining an update process for the translation source / translation result correspondence table buffer 53 at the time of sentence deletion in the same embodiment.

FIG. 9 is a view showing an example of a translation source document and a translation result document before and after sentence deletion in the same embodiment in comparison.

10 is a diagram showing an example of contents before and after updating a translation source / translation result correspondence table buffer 53 corresponding to before and after sentence deletion in the example of FIG. 9;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Input part, 2 ... External storage device, 3 ... Display part, 4 ... Printing part, 5 ... RAM, 6 ... Control device, 51 ... Translation source document storage buffer, 52 ... Translation result document storage buffer, 53 ... Translation source / Translation result correspondence table buffer, 54 ... Screen display buffer, 55 ... Printing buffer, 61 ... Display processing control section (translation result output means, output format designation receiving means), 62 ... Print processing control section (translation result output means) , Output format designation receiving means), 63 ... translation processing control section (registering means), 64 ... translation dictionary, 65 ... editing processing control section (editing position designation receiving means, updating means).

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Hideki Nagasaki, 2-9 Suehiro-cho, Ome-shi, Tokyo Inside the Ome Plant of Toshiba Corporation (72) Inventor Chizuru Sato 2--9, Suehiro-cho, Ome-shi, Tokyo Stock Association Inside the Toshiba Ome Factory (72) Inventor Izumi Sakata 2-9 Suehiro-cho, Ome-shi, Tokyo Stock Company Toshiba Inside Ome Factory (72) Inventor Tomotaka Mori 2-9 Suehiro-cho, Ome City Tokyo Metropolitan Government Toshiba Corporation In the Ome Factory (72) Masanori Gunji 2-9 Suehiro-cho, Ome-shi, Tokyo Inside the Ome Factory, Toshiba Corporation (72) Inventor Masakazu Saito 1385 Shinmachi, Ome-shi, Tokyo Inside Toshiba Software Engineering Co., Ltd.

Claims (6)

[Claims] 1. A machine translation apparatus comprising translation processing means for translating a translation source document in a first language to be translated into partial units and translating it into a second language to obtain a translation result document. Translation source / translation result correspondence table buffer means for storing the position information on the document corresponding to the original part and its translation result in partial units, and the translation processing unit executes translation in partial units. In this case, registration means for registering, in the translation source / translation result correspondence table buffer means, a set of position information on a document corresponding to a translation source portion and its translation result, respectively, in association with the portion, Based on the registered contents of the source / translation result correspondence table buffer means, corresponding portions within the externally designated range or the entire document are sequentially cut out from the translation source document and the translation result document to form pairs. A machine translation device comprising: a translation result output unit for creating and outputting a bilingual document. 2. When the output of the bilingual format document by the translation result output unit is a display or a print output, an external designation of the display format or the print format is performed for each of the translation source document and the translation result document. The translation result output means further comprises an output format designation receiving means for receiving, and creates or displays or prints the bilingual format document in accordance with the display format or print format designation accepted by the output format designation receiving means. The machine translation apparatus according to claim 1, wherein the machine translation apparatus outputs. 3. An edit position designation accepting unit for accepting designation from outside of the edit position of the translation source document in units of sub-units registered in the translation source / translation result correspondence table buffer unit, and this editing Edit processing means for performing an externally specified editing operation on the designated portion of the translation source document received by the position designation receiving means and the corresponding portion of the translation result document, which is designated after editing the translation source document. When the translation of the part is necessary, the translation processing means performs the translation, and the translation result is used to perform the editing operation on the corresponding portion of the translation result document, and the editing operation by the editing processing means. Updating means for updating the contents of the translation source / translation result correspondence table buffer means so that the positional relationship of the corresponding partial units of the subsequent translation source document and translation result document is reflected. Further machine translation apparatus according to claim 1 or claim 2, characterized in that it comprises. 4. A translation result output method in a machine translation device that obtains a translation result document by translating a translation source document of a first language to be translated into partial units into a second language, When performing translation while dividing the translation source document into partial units, a set of position information on the corresponding document of the translation source part and its translation result is registered in partial units corresponding to the parts. And a translation source / translation result correspondence table is created in advance. Based on the translation source / translation result correspondence table, corresponding portions within the range externally designated from the translation source document and the translation result document or in the entire document. A method for outputting a translation result, characterized in that a bilingual format document is created and output by sequentially cutting out and pairing. 5. When the output of the bilingual format document is a display or a print output, an external designation of the display format or the print format is accepted for each of the translation source document and the translation result document, and this is accepted. Depending on the display format or print format specification,
The translation result output method according to claim 4, wherein the bilingual format document is created and displayed or printed out. 6. A designation from the outside of an editing position for the translation source document is accepted for each partial unit registered in the translation source / translation result correspondence table, and the designated portion of the accepted translation source document is accepted. And an externally specified editing operation is performed on the corresponding portion of the translation result document, and at that time, if translation of the edited designated portion of the translation source document is required, the designated portion is translated, and then An editing operation is performed on the corresponding portion of the translation result document using the translation result, and the translation source / translation is performed so that the positional relationship of the corresponding partial units of the translation source document and the translation result document after the editing operation is reflected. The result correspondence table is updated, and the result correspondence table is updated.
Alternatively, the translation result output method according to claim 5.