JP2007052613A - Translation device, translation system and translation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable provision of accurate translation data in a short time without need of conversion time or translation time by providing data to a user using the translation data of an original which are preliminarily stored in a storage means. <P>SOLUTION: In this translation device, when translation data corresponding to an original text preliminarily exist in a storage means (step S4: YES), the translation data are used. When not the translation data but the original text data already exist, the original text data are translated to obtain translation data (step S7). When neither the original text data nor the translation data exit, the document obtained from an original are translated (step S9) to use the translation data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a translation apparatus, a translation system, and a translation method that can perform translation with high translation accuracy.

  Due to the spread of cross-border activities such as global information distribution and economic activities in recent years, there is an increasing demand for translation from one language (eg Japanese) to another (eg English, Chinese). . However, when a contractor (translator) is requested to translate, it is generally expensive and time-consuming, so that there is an increasing demand for automatic translation using a machine such as a computer.

In general, automatic translation obtains image data obtained by imaging a paper on which a document is printed with a scanner device, and character recognition (for example, OCR (Optical Character Recognition)) of the image data. The document data is converted into text data by the above-described method, and this document data is translated from the first language into the second language using machine translation software.
This machine translation software recognizes characters in word units from document data that has been converted to text by document analysis, derives the meaning of words by referring to a translation dictionary, and creates a translation sentence from a dependency relationship. Is going.

In machine translation, since the translation accuracy depends on the number of words stored in the translation dictionary, the accuracy of machine translation is set by the number of data corresponding to the storage capacity. Further, considering the case of translating Japanese into another language, Japanese may be described with the subject omitted, so that it is difficult to translate with the subject omitted in machine translation.
On the other hand, human translation can also translate new words that are not stored in the dictionary according to the experience of the translator, complementing the subject that is omitted from the previous and subsequent sentences. It is also possible.
Thus, since the translation accuracy by machine translation is determined by the accuracy by character recognition and the accuracy by machine translation software, the accuracy is lower than that by human translation. For example, when the translation data described in the second language is returned to the first language, the translation processing is performed twice, so that the obtained document data becomes a rather difficult sentence and may not be understood. .

  Here, as a technique for improving the translation accuracy, when a word recognized by character recognition is translated, if a corresponding word does not exist in the translation dictionary, a technique for automatically correcting the word (see Patent Document 1, Reference). ), A technique of printing a barcode in which a part or all of a character string is replaced on a document, and reading the barcode and converting it into a character string using a dedicated barcode reader as required (Patent Document 2) , See).

Japanese Patent Laid-Open No. 5-54070 JP-A-8-83283

  However, Patent Document 1 increases the accuracy of machine translation software, and Patent Document 2 increases the accuracy of character recognition for converting image data into text data. However, with any technique, the accuracy of translation by the person concerned cannot be reached.

  The present invention has been made in view of the above-described prior art. When there is translation data corresponding to the original text, a translation device, a translation system, and a translation method with high translation accuracy can be obtained by using the translation data. The purpose is to provide.

  In order to achieve the above object, a translation apparatus employed by the present invention includes a storage unit that stores translation data obtained by translating an original text in association with identification information of the original text, and original text input data that is input together with the identification information. Extraction means for extracting the identification information, determination means for determining whether or not translation data corresponding to the identification information extracted by the extraction means exists in the storage means, and the determination means includes translation data Translation data output means for outputting translation data in the storage means if determined, and for outputting translation data obtained by performing translation processing on the original input data if the determination means determines that there is no translation data It was characterized by comprising.

  According to this invention, when translation data corresponding to the original text exists in advance, it is not necessary to perform translation processing and character recognition processing by using the translation data, and translation from the original text input data to the translation data is possible. Regardless of accuracy, translation with high translation accuracy can be realized.

  Another translation apparatus employed by the present invention includes: storage means for storing at least one of original text data or translation data obtained by translating the original text in association with identification information of the original text; and identification information And extracting means for extracting the identification information from the original text input data input together with it, and determining whether the storage means has translation data or original text data corresponding to the identification information extracted by the extraction means When the determination means determines that the translation data is present, the translation data in the storage means is output. When the determination means determines that there is no translation data, the original data in the storage means is output. The translation data subjected to the translation processing is output, and if the determination means determines that neither the translation data nor the original text data exists, the original text input data Characterized in that anda translation data output means for outputting the translated data subjected to the translation processing for.

  According to this invention, when translation data corresponding to the original text exists in advance, it is not necessary to perform translation processing and character recognition processing by using the translation data, and translation from the original text input data to the translation data is possible. Regardless of accuracy, translation with high translation accuracy can be realized. In addition, if there is original data corresponding to the original text to be translated without the translation data, the original text data is translated to obtain the translation data. Can be prevented, and as a result, the translation accuracy is improved.

In the above configuration, when there is no translation data corresponding to the identification information of the original text in the storage means by the determination means, the translation data subjected to translation processing on the original text data corresponds to the identification information of the original text. It is desirable to provide data update storage means for updating and storing the translation data in the storage means.
In the above-described configuration, a translation data correction unit that corrects translation data according to an operation by an operator is provided, and the translation data updated and stored in the storage unit by the data update storage unit is corrected by the translation data correction unit. It is desirable to be data.
In the above configuration, it is desirable that the identification information extracted by the extraction unit is represented by a specific character string or a specific image.
In the above configuration, it is desirable that the translation data output means is a printing means, and the printing means prints the text of the translation data and prints the identification information of the translation data as a specific character string or a specific image.
In the above-described configuration, it is preferable that the image reading unit includes an image reading unit that reads an image from a document on which an original is printed, generates identification information and original text input data, and outputs the identification information and the original text input data to the extraction unit.

  According to the translation apparatus according to the present invention, when translation data corresponding to the original text exists in the storage means in advance, it is not necessary to perform translation processing and character recognition processing by using the translation data, and input the original text Regardless of the translation accuracy from data to translation data, a translated sentence with high translation accuracy can be obtained. In addition, when the stored translation data is used, since conversion time and translation time are not required, accurate translation data can be obtained in a short time.

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing an overall configuration of a translation system 1 according to the present embodiment. The system 1 includes a multifunction peripheral 100 and a translation server 200 connected via a network 300.
The multifunction device 100 has functions of a copy, a scanner, and a printer. The multifunction device 100 is connected to a translation server 200 or a user terminal (not shown) such as a personal computer via a network 300. The translation server 200 is a server device that specializes in translation processing, and has a function of receiving data corresponding to the original text from the multifunction peripheral 100, generating translation data for the received original text, and sending it back.

<1. Configuration of MFP 100>
FIG. 2 is a block diagram illustrating a functional configuration of the multifunction peripheral 100. The multi-function device 100 includes a reading unit 110, an image forming unit 120, a transmission / reception unit 130, a printing unit 140, and a control unit 150 that controls them. The control unit 150 includes a CPU, a RAM, and a ROM (not shown) and has a function for performing various processes.

  The transmission / reception unit 130 exchanges data with an external device (not shown) connected to the network 300 such as the translation server 200. The printing unit 140 receives the image data formed by the image forming unit 120 and prints an image on a sheet (recording material).

  The multi-function device 100 functions as a scanner device by outputting the image data read by the reading unit 110 to an external device via the transmission / reception unit 130. Further, the image printing unit 140 outputs an image corresponding to the image data read by the reading unit 110 to function as a copier. Further, the multi-function device 100 functions as a printer by outputting an image corresponding to data received from the outside via the data transmitting / receiving unit 130 in the image printing unit 140.

  Here, as shown in FIG. 1, the original M used in the present embodiment is printed with an original text and a barcode BM indicating the original text document ID at a predetermined position in the margin, and is output from the multifunction device 100. On the translated document T, a translated sentence and a barcode BT indicating the translated document ID are printed at a predetermined position in the margin.

The control unit 150 of the MFP 100 according to the present embodiment causes each unit to perform the following processing.
First, the control unit 150 receives an instruction of a language to be translated of the document M from an input instruction device (not shown). Next, the reading unit 110 reads the document M to acquire the original text image data and the image data of the identification information from the original text and the barcode BM printed on the original M, and the transmission / reception unit 130 receives the original text image. The data and the language to be translated are transmitted to the external translation server 200.
On the other hand, when translation data obtained through the processing described below is transmitted from the translation server 200, when the transmission / reception unit 130 receives the translation data, the image forming unit 120 converts the text-format translation data into an image-format translation image. Convert to data. At this time, if data representing identification information for printing a barcode is added to the translation data, the converted barcode image data is added to the translation image data. Then, the printing unit 140 prints the translated image data on a sheet and outputs a translated document T. As a result, the user obtains the translation document T translated into the translation target word.

<2. Configuration of Translation Server 200>
FIG. 3 is a block diagram illustrating a functional configuration of the translation server 200. The translation server 200 includes a document management unit 220, a translation dictionary unit 230, a transmission / reception unit 240, and a control unit 250 for controlling them, to which a document database 210 is connected. The control unit 250 includes a CPU, a RAM, and a ROM (not shown) and has a function for performing various processes.

  The transmission / reception unit 240 exchanges data with an external device (not shown) connected to the network 300 such as the multifunction peripheral 100. The translation dictionary unit 230 is used when the original text data is translated into a language to be translated based on an instruction from the control unit 250, and includes a dictionary for various languages.

  In the document database 210, text-format document data (file) is stored in association with each document ID. The storage process of the document data in the document database 210 is not limited to the process in step S8 of FIG. 5 to be described later, but when the scanner function, the copy function, and the printer function of the multifunction machine 100 are operated, or the user's operation process When it is done. The document data (original text data or translation data) of the manuscript is stored in the document database 210 together with the document ID, and the document data management unit 220 stores and updates the document ID and the document name of the translation data automatically or by user operation.

  The document management unit 220 manages the document data stored in the document database 210, and stores a plurality of data tables as shown in FIG. 4A for each reference document ID. The data table includes a first language version document ID, a document name (file name), a second language version document ID, a document name (file name), a third language version document ID, and a document name (file name). Is written in association with the document data stored in the document database 210. For example, this data table includes a Japanese XX instruction (document ID; 001-1), an English XX instruction (document ID; 001-2), and a Chinese XX instruction (document ID; 001-3). In this example, the first three characters are used as a common character string to indicate that each document is related, and the document description language format (first language, second language, third language) is indicated by a subscript.

The data table of FIG. 4A will be described. When the original text is a Japanese version document, the document ID of this document is “001-1”, which corresponds to this document ID “001-1”. The document is an English version document (document ID “001-2”) and a Chinese version document (document ID “001-3”).
If the original document is an English version document, the document ID of this document is “001-2”, and the document corresponding to this document ID “001-2” is a Japanese version document (document ID). “001-1”) and a Chinese version document (document ID “001-3”).
Further, if the original text is a Chinese version document, the document ID of this document is “001-3”, and the document corresponding to this document ID “001-3” is a Japanese version document (document ID “001-1”) and an English version document (document ID “001-2”).
Further, in the case of FIG. 4A, there are n such data tables.

On the other hand, the data table of FIG. 4B is a case where a “Japanese version” document is not stored in the document database 210. Even in this case, if the original document ID is “001-1,” this database can be extracted using the common basic name as a search term, and the English version of the document (document ID “001-2”) can be stored in the document database. ) And a Chinese version document (document ID “001-3”).
Further, since one data table is formed in the document management unit 220 for the document ID of one reference document (original text), the document management unit 220 includes a plurality of data tables.
In FIG. 4, for convenience, the Japanese document ID “001-1” is used as the original document ID. However, when the extracted document ID is “001-2”, the English document is the original document. The document corresponding to the extracted document ID is the original text for any document.

The control unit 250 of the translation server 200 according to the present embodiment performs a translation process described below.
The operation of the translation server 200 will be described below based on the flowchart shown in FIG.
The translation server 200 starts the operation by receiving the original text image data and the barcode image data transmitted from the multi-function device 100 by the transmission / reception unit 240.

  When the original image data from the multi-function device 100 is received by the transmission / reception unit 240, the control unit 250 converts the original image data in the image format into text format data by performing character recognition processing (for example, OCR processing). At this time, since there is image data indicating a barcode in the original text image data, this barcode is converted into character data, and information meaning the barcode, that is, a document ID is extracted (step S1).

  Next, the control unit 250 determines whether or not the document ID by the barcode is extracted by the process of step S1 (step S2). When the document ID is extracted (step S2; YES), the control unit 250 proceeds to step S3.

The control unit 250 searches the document data and translation data stored in the document database 210 by collating the document ID of the document data management unit 220 with the extracted document ID as a search key, and obtains the search result. Obtain (step S3).
In the example of the data table in FIG. 4A, when the extracted original document ID is “001-1”, the first three characters are searched as a search key to start Japanese document data. As a search result, the English translation data and the Chinese translation data are stored in the document database 210.

The control unit 250 determines whether there is translation data in the document database 210 from the search result (step S4). If the translation data of the translation target word is stored in the document database 210 (step S4; YES), the translation data corresponding to the document ID is read, and this translation data is temporarily stored in a storage area (not shown). Store (step S5).
For example, if the translation target language requested by the user is English, the document ID “001-2” of the English version of the translation data with respect to the original document ID “001-1” is obtained from the search result. Therefore, it can be seen that the English translation data is stored in the document database 210.

  On the other hand, when the translation data is not stored in the document database 210 in step S4 (step S4; NO), the control unit 250 determines whether or not the original text data exists in the document database 210 from the search result (step S4). S6). If the original text data is stored in the document database 210 (step S6; YES), the original text data corresponding to the document ID is read, and the translation processing based on the original text data is performed using the translation dictionary unit 230. The translation data obtained by this process is temporarily stored in the storage area (step S7).

  Further, the control unit 250 stores the translation data obtained in step S7 in the document database 210 and converts the document ID obtained by changing the subscript of the original document ID into the data table of the document data management unit 220. The database is updated by adding it in association with the name (file name) (step S8).

Further, when the document ID is not extracted from the received original text image data in the determination process of step S2 (step S2; NO), or when the original text data is not stored in the document database 210 in the determination process of step S6. (Step S6; NO), the control unit 250 performs a translation process on the text data of the original text image data, and temporarily stores the translation data obtained by this process in the storage area (Step S9).
After this processing, the original text data before translation may be stored in the document database 210 together with the document ID, and a new data table corresponding to the original text data may be created in the document data management unit 220.

  Then, the control unit 250 transmits the translation data obtained by any one of the three processes to the multi-function device 100 via the transmission / reception unit 240 (step S10). The transmitted translation data includes a document ID indicating the translated sentence as well as data representing the translated sentence.

  In the MFP 100 that has received the translation data, as described above, the control unit 150 of the MFP 100 causes the image forming unit 120 to convert the text format translation data into an image format including a barcode BT indicating the document ID of the translation data. Convert to translated image data. Then, the printing unit 140 prints the translated text and the barcode of the translated image data on a sheet and outputs the translated document T (see FIG. 1).

  As described above, in the translation system 1 according to the present embodiment, when the translation data indicating the translation corresponding to the original sentence is stored in the document database 210 of the translation server 200 in advance, the translation data is transmitted to the multifunction peripheral 100. By doing so, OCR processing and translation processing can be omitted, and a translated sentence with high translation accuracy can be provided to the user. In addition, since the translation process is not required, the translation time can be greatly shortened.

Even if the translation data is not stored in the document database 210, if the original document data is stored in the document database 210, the OCR process is performed by translating the document data. Therefore, erroneous recognition of characters due to OCR processing can be prevented, and erroneous translation can be reduced as compared with a translation obtained by performing OCR processing and translation processing on an original sentence.
As described above, the translation system 1 according to the present embodiment can provide a translated sentence with high accuracy to the user.

As mentioned above, although embodiment of this invention was described, this invention is not limited to the above-mentioned embodiment, A various aspect is possible.
In the above-described embodiment, the read original image data is transmitted to the translation server 200 by the reading unit 110 of the multifunction peripheral 100. However, in the present invention, the barcode is printed at a predetermined position of the document. Alternatively, only the barcode identification ID may be read from the document image data and transmitted to the translation server 200. As a result, the original image data that has a larger capacity than the barcode data capacity is not transmitted, and the transmission time via the network 300 can be greatly reduced.
When translation data and original text data are not stored in the document database 210 of the translation server 200, the original text image data may be transmitted again from the multifunction peripheral 100 to the translation server 200.
Alternatively, when the document M is read by the reading unit 110 of the multifunction peripheral 100, OCR processing may be performed to convert the original text image data in the image format into the original text input data in the text format. However, the original text input data is different from the document data stored in the document database 210 in advance.

  In the present embodiment, the translation system in which the multifunction peripheral 100 and the translation server 200 are connected via the network 300 has been described. However, the present invention is not limited to this, and the multifunction peripheral 100 includes the function of the translation server 200. 100 may be a translation device. Such a multifunction device 100 is preferably applied to a relatively small company that can support work with one multifunction device. At this time, the document data stored in the document database 210 is stored in the document database 210 when the scanner function, copy function, and printer function of the multi-function peripheral 100 are operated, as well as the storage update in step S8 of FIG. You may let them. At this time, the document data of the original to be scanned, copied, and printed may be stored in the document database 210 together with the document ID, and the document ID and document name of the translation data may be stored and updated in the document data management unit 220.

In the data table of the embodiment, as shown in FIG. 4, document data is associated as a document ID having a common character string. However, the present invention is not limited to this, and even with different document IDs, What is necessary is just to associate with a data table.
In the present invention, the translatable languages are not limited to three of the first to third languages, and it is needless to say that the present invention can deal with two languages, four languages or more.

  In the above embodiment, the original document information is used as a document ID, and the document ID is converted into a barcode. However, the present invention is not limited to this, and an IC tag or the like in which a character string existing in a designated area or a document ID is written. May be used instead of the barcode. For example, in the case of a character string, a character string existing in a predetermined area portion of a sentence printed on paper is used as a document ID or a document name, and this predetermined area is subjected to OCR processing so that the document ID or document name is converted into the original text. You may make it extract as identification information.

  In addition, when the user describes annotations such as annotations and sticky notes on the manuscript, when the manuscript is read by the reading unit 110, the characters described in the annotation and the position information thereof are calculated, and the translation server 200 When the translated data is returned, the printing unit 140 prints the translated annotation together with the translated text at the same position as the original document based on the translated data and the positional information of the annotation.

  The translation data stored and updated in the document database 210 in step S8 is stored as translation data by machine translation by the translation server 200, but here, it is not stored, but the translation corrected by the translator based on this translation data. Data or translation data translated from an original document by a translator may be newly stored in the document database 210.

In the above embodiment, the translation system 1 is composed of the multifunction peripheral 100 and the translation server 200. However, a translation in which a personal client terminal (for example, a personal computer) and the translation server 200 are connected via the network 300. It is good also as a system.
In this system, the translation data in the text format or the original text data and the translation data can be displayed on the display unit of the client terminal while being compared. Furthermore, various usage methods such as displaying data on a display unit, designating a text area by an area designating means such as a mouse, and displaying a translation corresponding to the text area can be applied.

  In FIG. 1, only one MFP 100 and one translation server 200 are shown in order to prevent the drawing from becoming complicated, but the translation system 1 has a plurality of MFPs 100 and translation servers 200. Of course, it may be.

  In the above-described embodiment, the MFP 100 is realized by the control unit 150 and the translation server 200 is executed by the control unit 250, respectively. However, a part or all of this is realized by hardware. It is good.

It is a figure which shows the structure of the translation system which concerns on embodiment of this invention. 1 is a block diagram illustrating a configuration of a multifunction machine used in an embodiment of the present invention. It is a block diagram which shows the structure of the translation server used for embodiment of this invention. It is a figure which shows the data table memorize | stored in the document data management part. It is a flowchart which shows the translation process by this embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Translation system 100 ... Multifunction machine 110 ... Reading part 120 ... Image forming part 130 ... Transmission / reception part 140 ... Printing part 150 ... Control part 200 ... Translation server 210 ... Document database 220 ... Document data Management unit 230 ... Translation dictionary unit 240 ... Transmission / reception unit 250 ... Control unit 300 ... Network

Claims (11)

Storage means for storing translation data obtained by translating the original text in association with the identification information of the original text;
Extracting means for extracting the identification information from the original text input data input together with the identification information;
A determination unit for determining whether or not the storage unit has translation data corresponding to the identification information extracted by the extraction unit;
When the determination means determines that there is translation data, the translation data in the storage means is output, and when the determination means determines that there is no translation data, the translation processing is performed on the original input data. A translation data output means for outputting data. Storage means for storing at least one of the original text data or the translation data obtained by translating the original text in association with the identification information of the original text;
Extracting means for extracting the identification information from the original text input data input together with the identification information;
Determining means for determining whether translation data or original text data corresponding to the identification information extracted by the extracting means exists in the storage means;
When the determination unit determines that there is translation data, the translation data in the storage unit is output, and when the determination unit determines that there is no translation data, a translation process is performed on the original text data in the storage unit. A translation data output means for outputting the translation data subjected to the translation process, and further, when the judgment means judges that neither the translation data nor the original text data exists, the translation data subjected to translation processing on the original text input data; A translation apparatus comprising: The translation apparatus according to claim 1 or 2,
When there is no translation data corresponding to the identification information of the original text in the storage means by the determination means, the translation data subjected to translation processing on the original text data is used as the translation data corresponding to the identification information of the original text. A translation apparatus comprising data update storage means for updating and storing in a storage means. The translation device according to claim 3, wherein
Translation data correction means for correcting the translation data according to the operation of the operator is provided.
The translation apparatus, wherein the translation data updated and stored in the storage means by the data update storage means is data corrected by the translation data correction means. The translation apparatus according to claim 1 or 2,
The identification information extracted by the extraction means is represented by a specific character string or a specific image. The translation apparatus according to claim 1 or 2,
The translation data output means is a printing means,
The translation device prints a sentence of translation data and prints identification information of the translation data as a specific character string or a specific image. The translation apparatus according to claim 1 or 2,
Image reading means,
The translation apparatus, wherein the image reading unit reads an image from a document on which an original text is printed, generates identification information and original text input data, and outputs the identification information and the original text input data to the extraction unit. A terminal device and a server device that communicates with the terminal device via a network;
The server device is
Storage means for storing translation data obtained by translating the original text in association with the identification information of the original text;
Extracting means for extracting the identification information from the original text input data input together with the identification information;
A determination unit for determining whether or not the storage unit has translation data corresponding to the identification information extracted by the extraction unit;
When the determination means determines that there is translation data, the translation data in the storage means is output, and when the determination means determines that there is no translation data, the translation processing is performed on the original input data. Translation data output means for outputting data,
The terminal device is
Image reading means for generating the original text input data and identification information;
A translation system comprising: data display means for displaying or printing translation data obtained by the translation data output means. A terminal device and a server device that communicates with the terminal device via a network;
The server device is
Storage means for storing at least one of the original text data or the translation data obtained by translating the original text in association with the identification information of the original text;
Extracting means for extracting the identification information from the original text input data input together with the identification information;
Determining means for determining whether translation data or original text data corresponding to the identification information extracted by the extracting means exists in the storage means;
When the determination unit determines that there is translation data, the translation data in the storage unit is output, and when the determination unit determines that there is no translation data, a translation process is performed on the original text data in the storage unit. A translation data output means for outputting the translation data subjected to the translation process, and further, when the judgment means judges that neither the translation data nor the original text data exists, the translation data subjected to translation processing on the original text input data; With
The terminal device is
Image reading means for generating the original text input data and identification information;
A translation system comprising: data display means for displaying or printing translation data obtained by the translation data output means. A storage step of storing the translation data obtained by translating the original text in the storage means in association with the identification information of the original text;
An extraction step of extracting the identification information from the original text input data input together with the identification information;
A determination step of determining whether or not translation data corresponding to the identification information extracted by the extraction step exists in the storage unit;
When the determination step determines that there is translation data, the translation data in the storage means is output, and when the determination step determines that there is no translation data, the translation processing is performed on the original input data. A translation data output step for outputting data. A storage step of storing at least one of the original text data or the translation data obtained by translating the original text in the storage means in association with the identification information of the original text;
An extraction step of extracting the identification information from the original text input data input together with the identification information;
A determination step of determining whether translation data or original text data corresponding to the identification information extracted by the extraction step exists in the storage unit;
When the determination step determines that there is translation data, the translation data in the storage means is output, and when the determination step determines that there is no translation data, a translation process is performed on the original text data in the storage means. A translation data output step that outputs the translation data that has been subjected to a translation process when the determination step determines that neither the translation data nor the original text data exists. A translation method characterized by comprising:

Images