JP2907840B2 - Translation method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial application field) The present invention relates to a translation method and apparatus for translating a document (data) in a first language into a document (data) in a second language. More particularly, the present invention relates to a translation method and apparatus for reducing time required for translation and processing related to translation.

(Prior Art) In recent years, various translation apparatuses have been proposed that automatically translate using a computer or the like.

In such a conventional translation method and device, when a first language document (data) composed of a plurality of sentences is input using an input device or the like, the input first language document is converted for each sentence. The documents are sequentially translated, and the translated sentences (data) in the second language and the original document in the first language are recorded in association with each sentence.

Further, in the conventional machine translation apparatus, the translation result, that is, the modification of the document in the second language is corrected while the correspondence between the document (data) in the first language and the document (data) in the second language is held. It was possible to do.

By the way, when a document to be newly translated (an untranslated document (data) in the first language) is to be translated into the second language, the first and second recorded documents already associated with each sentence as described above. There are cases where it is desired to use documents (data) in two languages. In such a case, for example, only a sentence in which the untranslated document and the document in the first language differ from each other in a document composed of a plurality of already translated documents is extracted, and only that part is newly translated, and another same sentence is extracted. It is conceivable to use the already translated text in the second language as it is.

(Problems to be Solved by the Invention) When translating a document (data) to be newly translated as described above, it is necessary to find a difference from a document (data) in the first language recorded in association with the document (data). In the past,
The operator needs to collate each sentence from the top in advance and extract the difference, and the translation takes time.

In addition, the extraction of the difference by the operator is performed by sequentially comparing and extracting the first sentence of the document for each sentence with the document in the first language recorded in association with the document to be translated. become. Therefore, for example, when sentences of the same sentence exist at positions in different orders, the corresponding translated sentence is rarely used, and the sentence to be newly translated is increased.

[Structure of the Invention] (Means for Solving the Problems) In order to achieve the above object, according to the translation method of the present invention, when a newly input untranslated document in the first language is translated, translation is already performed. As a result, the translated document of the first language composed of a plurality of blocks and the translated document obtained by translating the translated document into the second language are previously associated with each other on a block basis and stored in the storage unit. Using the data,
One document block of the untranslated document and the translated document in the storage data are compared in a unit of a block by a comparing unit, and the presence or absence of a matching block is detected. A translation block of the translated document corresponding to the original text block that matches the translated document is read from the storage data, and the document block of the untranslated document to be compared corresponds to the translated block of the translated document that has been read. And if the matching block is not detected in the comparison, the document block of the untranslated document to be compared is newly translated into a second language by the translation means, The document block of the untranslated document to be compared and the new translation block of the newly translated new translated document are associated with newly stored data. Characterized in that to store Te.

In order to achieve the above object, the translation device of the present invention
A storage unit in which a translated document of a first language composed of a plurality of blocks and a translated document obtained by translating the translated document into a second language are stored in association with each other in block units, A comparing unit that compares one document block of the untranslated document in the first language with the translated document stored in the storage unit on a block-by-block basis and detects the presence or absence of a matching block; When a matching block is detected, a translated text block of the translated document corresponding to the original text block that matches the translated document is read from the storage unit, and the document block of the untranslated document to be compared and the read text block are read. Means for associating the translated text block of the translated document with new storage data and, when the matching block is not detected by the comparing means, The translated document block of the untranslated document is newly translated into a second language by the translation means, and the document block of the untranslated document to be compared corresponds to the new translation block of the newly translated new translated document. And storing the data in the storage means.

(Operation) As described above, the translation method of the present invention compares the translated document (data) with the untranslated document (data) of the first language to be newly translated and determines whether there is a matching block. Since the detection is performed, the operator can save the trouble of comparing the full text of the untranslated document and the first language translated document one by one.

Further, the translation apparatus of the present invention includes the comparing means as described above, and compares the translated document with the untranslated document of the first language to be newly translated to detect the presence or absence of a matching block. As in the first aspect, the burden on the operator can be greatly reduced.

Embodiment An embodiment according to the present invention will be described below in detail with reference to the drawings.

First, the overall configuration of the translation method and apparatus according to the present invention will be described with reference to FIG. 2. The input unit 10 is composed of a keyboard, a disk device, and the like, and character information and information based on various key operations. Input various information from the storage medium.

The translation unit 11 translates an untranslated document (data) composed of a plurality of blocks in a first language, for example, English, into a document (data) in a second language, for example, Japanese for each block.

The display unit 12 is composed of, for example, a CRT display device, and displays character information requiring translation, translated character information, and the like.

The recording unit 13 stores the translated document for each block in the first language and the translated document (data) in the second language that is translated from the translated document.
Are recorded in association with each other. Further, the recording unit 13 can separately and independently record the translated document for each block in the first language and the translated document in the second language.

The editing unit 16 executes a process related to editing various documents (data) recorded in the recording unit 13 or these documents (data) and the document (data) input from the input unit 10.

The translation control unit 18 is connected to each of the input unit 10, the translation unit 11, the display unit 12, the recording unit 13, and the editing unit 16, and controls these devices as a whole. The translation control unit 18
A comparing unit for comparing, for each block, whether or not the untranslated document (data) L1A and the translated document (data) L1B input from the input unit 10 or the editing unit 16 match; A sending unit for sending the untranslated document L1A portion of the block that does not match based on the comparison result to the translating unit 11, and a translated unit corresponding to the translated document L1B of the matching block based on the comparison result of the comparing unit It is provided with a translation processing unit for reading out the document (data) L2B from the storage unit 13 and corresponding to the untranslated document L1A of the block.

Next, the schematic operation of the translation control unit 18 shown in FIG. 2 will be described with reference to FIG.

In step 101, when an untranslated document (data) L1A in the first language is input, the process proceeds to step 102, where the recording unit
It is determined whether or not the translated document (data) L2B exists in 13. If the translated document L2B does not exist in step 102, the process proceeds to step 103, where the untranslated document L1A in the first language is transferred to the translation unit 11 and instructed to be translated. When the translation is completed, as shown in FIG. 9 (B), for example, as shown in FIG. 9 (B), an untranslated document L1A in English as a first language and a
A document (data) L2A obtained by translating 1A into Japanese as a second language is displayed correspondingly. Subsequently, the process proceeds to step 104, where the document L1A and its translation result are recorded in the recording unit 13 in association with each other.

The document (data) recorded in this manner is referred to as a translated document (data) L1B in a first language and a corresponding translated document L2B in a second language, respectively. Used for translation.

Further, the translated document (data) may be corrected prior to the recording process in step 104. Ninth
FIG. (C) shows a final translation result document L2B obtained by modifying the above-described document L2A translated from “document A”. When the above correction is made, the translated document L2B is associated with the translated document L2B and the translated document L1B.
Record as

In step 102, if the translated document L2B exists in the recording unit 13, that is, the translated document L1B in the first language
If there is a corresponding translated document L2B in the second language, the process proceeds to step 105. In step 105, as shown in FIG. 9 (D), the display unit 12 inquires whether to use the past translation result, and selects to use the past translation result, for example, "Document A". Then
The translated document L1B in the first language corresponding to “Document A” recorded in the recording unit 13 is read.

Next, in step 106, for example, the untranslated document (data) L1A in the first language of "Document B" is divided into n blocks. Here, the size of each block is displayed on the display unit 12 as shown in FIG. 9 (A), and the size of each block may be a single sentence, a paragraph composed of a plurality of sentences, a chapter composed of a plurality of paragraphs, or the like. An appropriate size is selected. Steps
At 107, the value of the coefficient i for identifying the block is set to 1.

Subsequently, in step 108, it is determined whether or not any of the blocks of the translated document (data) L1B includes data that matches the data of the block i of the untranslated document (data) L1A of “document B”.

If no matching data is included in step 108, the flow advances to step 109 to transfer the data of the block i of the untranslated document L1A to the translation unit 11 and instruct it to translate the data. It is associated as data of block i of the untranslated document L1A.

If data matching the data of the block i of the untranslated document L1A exists in any block of the translated document L1B in step 108, the process proceeds to step 110. In step 110, the translated document L2B in the second language recorded in association with the translated document L1B in the first language is read from the recording unit 13, and the read translated document L2B is read.
As the data of the block i of the untranslated document L1A.

That is, in step 110, as shown in FIG. 9 (E), the translated document L2B of “Document A” is directly translated into the translated document of “Document B” for the untranslated document L1A of “Document B”. Corresponds as L2A.

In step 111, it is determined whether or not the block i is n, that is, whether or not the processing has been completed for all blocks of the untranslated document (data) L1A. Is Step 11
From 1 the process proceeds to step 112 to count up the value of the coefficient i by +1. Proceeding from step 112 to step 108 again,
Hereinafter, similarly, the comparison process and the translation process based on the comparison result are executed for all the blocks of the untranslated document (data) L1A.

If it is determined in step 111 that the above-described processing has been completed for all blocks of the untranslated document (data) L1A, the process proceeds to step 104, where the untranslated document (data) L1A of each block described above and the corresponding translation result Is recorded in the recording unit 13.

Next, with reference to FIG. 4, the comparison process for each block in steps 108 to 112 shown in FIG. 3 and the translation process based on the comparison result will be described in detail.

In step 201, it is determined whether or not documents (data) that have been translated and recorded in the recording unit 13 are arranged in blocks. That is, the word located at the end of each block is associated with the block, and the word located at the end is set as a keyword for classifying each block. Determine if the keyword has been set.

If it is determined in step 201 that the blocks have not been organized, the process proceeds to step 202, where both documents (data), that is, the translated document L1A in the first language and the translated document L1B in the first language are stored in each block. The comparison is made sequentially every time.
If it is determined in step 202 that there is no matching block among the documents in each block, the process proceeds to step 21.
Proceeding to 3, it is determined that the past document (data), that is, the translated document L1B stored in the storage unit 13 cannot be used for the subsequent translation processing. If there is a matching block among the documents (data) of each block in step 202, the process proceeds to step 215 via S1, and the translated document L1B recorded in the recording unit 13 is translated. It is determined that the process can be used.

In step 201, when the above-mentioned keywords are arranged in blocks, the process proceeds to step 203, where the input data, ie, the unit of each block of the untranslated document L1A and the past document (data), ie, recorded in the recording unit 13, It is determined whether the unit of each block of the translated document L1B is equal to each other. If the units of both blocks are different in step 203, the process proceeds to step 204, where the translated document L1B
The unit of the block is subjected to compound processing, and it is determined whether or not the block unit of the untranslated document L1A can be matched by, for example, setting a plurality of blocks as one unit of a new block. If the units of both blocks cannot be matched in step 204, the process proceeds to step 202. If it is determined in step 204 that the units of both blocks can be matched, the process proceeds to step 205 to execute the compound processing of the translated document L1B to match the units of each block of both data.

In step 203, when the unit of each block of the untranslated document L1A is equal to the unit of each block of the translated document L1B, the process proceeds to step 206, and corresponds to the first block to be compared, that is, the coefficient i = 1. Select a block. Subsequently, in step 207, it is determined whether or not the last word, that is, the keyword, is the same for each block of both documents to be compared. If the keywords located in the first block are equal to each other in step 207, step
Proceeding to 208, it is determined whether or not each keyword matches each other for the entire corresponding block. If the keywords of the first block do not match in step 207, or if the keywords of the entire block do not match each other in step 208, the process returns from step 209 to step 207 via step 210, and returns to step 207. The comparison process for each block is executed in the same manner as in. When the comparison process for the last block is completed in step 209, the process proceeds to step 213.

Next, the operation related to the translation process will be described using a specific example.

The translated document (data) L1B as shown in FIG. 5 and the translated document (data) L2B as shown in FIG.
The case where an untranslated document (data) L1A as shown in FIG. 7 is translated in the state recorded in 3 will be described.

Here, the translated document L1B shown in FIG. 5 is a guide for a specific device described in English, and includes L1B1, L1B2, and L1B.
It consists of 10 chapters up to 3, ..., L1B10. The sixth
The translated document L2B shown in the figure is a translation of the above-described translated document L1B into Japanese, and is composed of data consisting of ten chapters L2B1, L2B2, L2B3,..., L2B10. In the translated document L1B and the translated document L2B, data for each chapter is divided as a block unit, and the data for each block is recorded in association with each other. The untranslated document L1A shown in FIG. 7 is a guide written in English, and is composed of data consisting of ten chapters L1A1, L1A2, L1A3,..., L1A10. This untranslated document L1A
L1A1, L1A2, L1 corresponding to Chapters 1 to 4
A3 and L1A4 are the same as the data L1B1, L1B2, L1B3 and L1B4 corresponding to the first to fourth chapters of the translated document L1B.
The data L1A7, L1A8, L1A9, and L1A10 corresponding to Chapters 7 to 10 of the untranslated document L1A are the data L1B7, L1B8, L1B9, and L1B10 of the translated document L1B corresponding to Chapters 7 to 10.
Is the same as Therefore, only the data corresponding to Chapters 5 and 6 differ between the untranslated document L1A and the translated document L1B.

In step 101, when an untranslated document L1A as shown in FIG. 7 is input, the process proceeds to step 105 via step 102, and the translated document L1B as shown in FIG.

Subsequently, in step 106, the untranslated document L1A is divided into blocks for each chapter, that is, ten blocks. In steps 107 and 108, it is determined whether or not the data of the blocks corresponding to the first chapter in both the untranslated document L1A and the translated document L1B match. If both data match, the process proceeds to step 110, where the data L2B1 of the translated document L2B recorded in association with the data L1B1 corresponding to the first chapter of the translated document L1B is read from the recording unit 13. At the same time, the read data L2B1 is set as the translated data L2A1 of the data L1A1.

Similarly, the data L1A2, L1A3, L1A4 of the untranslated document L1A
Is the same as the data L1B2, L1B3, L1B4 of the translated document L1B, the data L2B2, L2B3, L2B4 of the translated document L2B is set as translated data L2A2, L2A3, L2A4. That is, as the translated data corresponding to the first to fourth chapters of the untranslated document L1A, the data corresponding to the first to fourth chapters of the translated document L2B shown in FIG. It is attached and set.

Next, since the data L1A5 and L1A6 corresponding to Chapters 5 and 6 of the untranslated document L1A are different from the data L1B5 and L1B6, the process proceeds from Step 108 to Step 109, where the data L1A5 and L1A6 are transmitted to the translation unit 11. Is done. This allows data L1A5 and L1A6
And are newly translated.

Next, data L1A7, ..., L1A10 from Chapter 7 to Chapter 10
As described above, the data of the corresponding block of the translated document L2B is sequentially allocated as it is, as described above.

Therefore, as shown in FIG. 8, the document L2A obtained by translating the untranslated document L1A is converted into data L2B1,... Of the translated document (data) L2B.
, L2B4 are set as translated data L2A1,..., L2A4, and translated data L2B7,..., L2B10 are set as translated data L2A7,. Only the data L1A5 and L1A6 corresponding to Chapters 5 and 6 of the untranslated document (data) L1A are newly translated and set as translated data L2A5 and L2A6, respectively.

In the case where the data of the final translation result corrected is recorded as the translated document L2B recorded in the recording unit 13, the final data thus modified is stored in the translated document (data ) Can be used as it is as L2A.

Next, a case in which a document edited using the editing unit 16 is further translated will be described.

For example, when further translating a document (data) edited using a part of the translated document (data) L1B recorded in the recording unit 13, the unedited document (data) is set as the translated document L1B. And the edited document (data)
Is set as an untranslated document (data) L1A, and the same processing as described above is performed to execute the untranslated document (data) L1A.
L1A can be sequentially translated. At this time, in the edited document (data), if specific editing information is added to the edited portion or the edited block, it is easy to extract the edited portion or the edited block based on the editing information. The range to be newly translated can be easily determined without dividing the edited document (data) into blocks.

If the range to be newly translated among the untranslated documents (data) composed of a plurality of blocks is known in advance, the range to be newly translated is designated by a key operation from the input unit 10 as in the related art. be able to.

For example, when the untranslated document L1A of the block to be translated is designated by a key operation of the input unit 10 among the untranslated documents (data) L1A composed of a plurality of blocks, the untranslated document L1A of the designated block is The translation is performed by the translation unit 11 into a second language. For the untranslated document L1A other than the blocks translated by the translation unit 11, the data recorded in the recording unit 13 is used. For example, when a block to be used is specified by a key operation of the input unit 10, a portion of the translated document (data) L 2 B and a corresponding portion of the translated document (data) L 2 B of the specified block are transmitted from the recording unit 13. Is read. The read translated document L2B is associated with the untranslated document L1A of the corresponding block.

In this way, it is also possible to specify the range requiring a new translation or the range to be used among the translated documents (data) recorded in the recording unit, thereby simplifying the processing related to translation.

In the comparison process for each block in step 203, it is configured to determine whether both documents (data) completely match. However, a configuration is adopted in which a predetermined margin is set for the determination of the matching. May be. For example, a specific sentence, such as (FIGURE1), of a document (data) that constitutes a sentence with a plurality of words as a block unit is represented by (F
In the case where the data has been changed to IGURE2), it is possible to judge that there is consistency between both data simply by changing the numeral 1 to the numeral 2. That is, only the changed sentence needs to be translated.

In the case where a paragraph is used as a block unit and only one sentence of a plurality of sentences forming a specific paragraph is different, the entire paragraph including the different sentence is translated. With this configuration, when the different sentence affects the context of the entire paragraph, the entire paragraph can be accurately translated.

Further, when translating the untranslated document L1A, the most similar translated document is selected from the plurality of translated documents L1B recorded in the recording unit 13, and translation is performed with reference to the selected translated document. With this configuration, the processing related to translation can be further simplified.

In the recording unit 13, the translated document L1B and the translated document L2B obtained by translating the translated document L1B are recorded in association with each word or each sentence, each paragraph, or each chapter. When the correspondence has been made, it is possible to facilitate the correspondence between each paragraph of a larger unit or each chapter by combining a plurality of sentences,
Efficient comparison processing can be performed by appropriately setting the unit size of the block for performing the comparison processing.

Further, in the example shown in FIG. 4, if the units of the blocks of both documents (data) to be compared do not match in step 203, the units of the blocks of one document (data) are combined to form the other. Although it is configured to determine whether or not the unit of the block of the document (data) matches, as another embodiment, the translated data recorded in the recording unit 13 only when the unit of the block of both documents matches. Is used, there is no need to perform a process relating to the combination of block units, and the configuration can be further simplified.

In step 207 shown in FIG. 4, a word located at the end of each block of both documents is set as a keyword, and whether or not these two keywords match is configured. The present invention is not limited to this, and another keyword or a plurality of types of keywords may be set.

With the above-described configuration, the document (data) comparison processing for each block can be performed more accurately and quickly.

[Effects of the Invention] As described above, according to the present invention, when translating an untranslated document (data), only the untranslated document in the block that does not match the document (data) recorded in the recording unit is compared. Is translated by the translation means, and the data recorded in the above-mentioned recording means is read and used for the untranslated document portion of the matching block, so that the translation time is reduced and the labor required for translation is reduced. Significant reduction can be achieved.

[Brief description of the drawings]

FIG. 1 is a claim correspondence diagram, FIG. 2 is a block diagram showing an overall configuration of a translation method and apparatus according to the present invention, FIG. 3 is a flowchart showing a schematic operation of FIG. 2, and FIG. FIG. 3 is a flowchart showing the details of the control process from step 108 to step 112. FIG. 5 is a translated document (data) L1B in the first language recorded in the recording means 3.
FIG. 6 shows an example of the translated document L1 shown in FIG.
FIG. 7 is an explanatory diagram showing a translated document (data) L2B recorded in the recording unit in association with B, FIG. 7 is an explanatory diagram showing an example of an untranslated document (data) L1A, and FIG. L1A
FIG. 9 is an explanatory diagram showing a translated document (data) L2A obtained by translating. FIG. 9 is an explanatory diagram showing a display example on a display unit. L1A: Untranslated document (data) L1B: Translated document (data) L2A: Document in second language (data) L2B: Translated document (data) 1: Recording means 3, Comparison means 5 translation means, 7 reading means

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Koichi Hasebe 1 Komukai Toshiba-cho, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture Toshiba Research Institute, Inc. (56) References JP-A-63-300361 (JP, A) JP-A Sho 60-20284 (JP, A) JP-A-62-297972 (JP, A) (58) Fields investigated (Int. Cl. ⁶ , DB name) G06F 17/28

Claims (2)

(57) [Claims] When a newly input untranslated document in a first language is translated, as a result of being already translated,
The translated document of the first language composed of a plurality of blocks and the translated document obtained by translating the translated document into the second language are associated with each other in units of blocks in advance using storage data stored in the storage unit. Comparing one document block of the untranslated document with the translated document in the storage data by a comparing unit in block units to detect the presence or absence of a matching block, and detecting a matching block in the comparison; Reading the translated document block of the translated document corresponding to the original text block that matches the translated document from the stored data, and comparing the translated untranslated document block and the translated translated document block read out. If the corresponding block is not newly detected as the storage data in association with the comparison,
The document block of the untranslated document to be compared is newly translated into a second language by a translation unit, and the document block of the untranslated document to be compared and a new translation of the newly translated new translated document A translation method characterized by storing a new storage data in association with a block. 2. A storage means for storing a translated document in a first language consisting of a plurality of blocks and a translated document translated from the translated document into a second language in block units in advance, and Comparing means for comparing one document block of the untranslated document of the first language to be newly translated and the translated document stored in the storage means in block units to detect the presence or absence of a matching block; When the matching block is detected by the comparing means, the translated text block of the translated document corresponding to the original text block matching the translated document is read out from the storage means, and the document of the untranslated document serving as the comparison target is read out. Means for associating the block with the translated text block of the read translated document and newly storing it as storage data; and when the matching block cannot be detected by the comparing means, The document block of the untranslated document to be compared is newly translated into a second language by a translation means, and the document block of the untranslated document to be compared and a new translation of the newly translated new translated document Means for storing in the storage means in association with a block.