JP4001133B2 - Automatic translation device and automatic translation system - Google Patents

Description

  The present invention relates to an automatic translation apparatus and an automatic translation system.

  With the spread of activities across borders such as global information distribution and economic activities in recent years, the demand for translation from one language (for example, English) to another language (for example, Japanese) is increasing. However, when a translation is requested from a trader (translator), the price is generally high and it takes time. Therefore, the demand for an automatic translation (machine translation) apparatus that automatically translates using a machine such as a computer is increasing.

Machine translation is generally less accurate than human translation, but for example, it complicates the translation process, provides learning functions, increases the vocabulary of the dictionary, or sets appropriate translation selection rules. Techniques for improving translation accuracy have been developed, such as increasing it. Furthermore, not only a stand-alone translation device but also a network translation system in which a translation server device having a high processing capability and a terminal device are connected via a network has been developed (see Patent Documents 1 to 3).
JP-A-6-243210 JP 2001-357033 A JP 2003-162353 A

  However, in the translation systems described in Patent Documents 1 to 3, when a very simple sentence is translated or a sentence with a small amount is translated, data is transmitted and received through the network one by one, and the translation server device It was the structure which performs processing. In other words, even though the accuracy of translation may be low but the user wants to obtain a translated sentence as soon as possible, the translation processing is performed by the translation server. Therefore, there has been a problem that such a user's request cannot be completely met.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide an automatic translation apparatus in which a user can select the accuracy and processing time of an output translation.

  In order to solve the above-described problem, the present invention specifies original text data described in a first language and a translation processing allowable time for translating the original text into a translated text described in a second language. Storage means for storing information; translation means capable of executing a plurality of translation processes for the original sentence; selection means for selecting one of the plurality of translation processes; and A translation time predicting means for calculating a predicted time of processing when a translated sentence is generated by one translation process selected by the selecting means, and a translation by the first translation process based on the predicted time and the allowable time There is provided an automatic translation apparatus having a judging means for judging whether sentence generation can be processed within the allowable time. According to this automatic translator, it is possible to determine whether a translation can be generated within an allowable time.

In a preferred aspect, the automatic translation apparatus according to the present invention is configured such that when the determination unit determines that the generation of the translated sentence by the first translation process cannot be performed within the allowable time, the original sentence data Transmitting means for transmitting a part or all of the data, receiving means for receiving translated text data for part or all of the original text data transmitted by the transmitting means, and translated text data received by the receiving means Output means for outputting the reflected translation. According to this automatic translation apparatus, when it is impossible to generate a translated sentence within an allowable time, it is possible to transmit data to an external apparatus and receive a result translated by the external apparatus.
In this aspect, the automatic translation apparatus according to the present invention optically reads a document and generates image data corresponding to the document, and performs character recognition processing on the image data generated by the reading unit. And character recognition means for storing the obtained character string data in the storage means as the original text data. According to this automatic translation apparatus, the user can obtain a translated sentence directly from the original sentence described in the manuscript.
In this aspect, the automatic translation apparatus according to the present invention may further include an image forming unit in which the output unit forms an image corresponding to the translated sentence on a sheet-like recording material. According to this automatic translation apparatus, the user can obtain a translation printed on a recording material such as paper.

  In addition, the present invention includes a terminal device and a server device that communicates with the terminal device via a network, and the terminal device stores the original text data described in the first language and the original text in the second language. Storage means for storing information for designating a permissible time for translation processing to be translated into a translation sentence described in the language, first translation means capable of performing a plurality of translation processes on the original sentence, and Selection means for selecting one translation process among a plurality of translation processes, and a translation time for calculating a predicted time of a process when a translation sentence is generated by one translation process selected by the selection means for the original sentence Based on the prediction means, based on the prediction time and the allowable time, a determination means that determines whether the generation of the translation sentence by the first translation processing can be processed within the allowable time; Translated text by processing When it is determined that the generation cannot be performed within the allowable time, original text transmission means for transmitting part or all of the original text data and one of the original text data transmitted by the original text transmission means A translated sentence receiving means for receiving data of a translated sentence for all or part, and an output means for outputting a translated sentence reflecting the translated sentence data received by the translated sentence receiving means. Generated by the original sentence receiving means for receiving the original sentence data transmitted by the original sentence transmitting means, the second translation means for generating a translation sentence from the data received by the original sentence receiving means, and the second translation means There is provided an automatic translation system comprising translation text transmission means for transmitting translation text data to the terminal device. According to this automatic translation system, the terminal device determines whether the translation can be generated within the allowable time, and if the translation cannot be generated within the allowable time, the data is transmitted to the server device. Then, the server device performs the translation process, and the terminal device can receive the result.

  According to the automatic translation apparatus of the present invention, translation accuracy and processing time can be selected as necessary.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
<A. Configuration of translation system 1>
FIG. 1 is a diagram showing a configuration of a translation system 1 according to an embodiment of the present invention. A multifunction machine 100 is an automatic translation apparatus according to the present embodiment, and is a multifunction machine having functions of an automatic translator, a copy, a scanner, and a printer. The multifunction device 100 is connected to the translation server 200 via the Internet 300. The translation server 200 is a server device that specializes in translation processing, and has a function of receiving original text data from the multifunction peripheral 100 as a client device, generating a translated text for the received original text, and returning the translated text. In FIG. 1, only one multifunction device 100 and one translation server 200 are shown in order to prevent the drawing from becoming complicated. However, the translation system 1 has a plurality of multifunction devices 100 and translation servers 200. May be.

<1. Configuration of MFP 100>
FIG. 2 is a diagram illustrating a functional configuration of the multifunction peripheral 100. The multi-function device 100 includes a reading unit 110 that reads an image displayed on a document, an image forming unit 120 that forms an image on a sheet (recording material) according to image data output from the control unit 140, and a translation It is comprised from the translation part 130 which processes. The multi-function device 100 functions as a scanner by outputting the image data read by the reading unit 110 to an external device. Further, the image forming unit 120 outputs an image corresponding to the image data read by the reading unit 110 to function as a copier. Furthermore, the multi-function device 100 functions as a printer by outputting an image corresponding to data input from the outside in the image forming unit 120.
FIG. 3 is a diagram illustrating a hardware configuration of the multifunction peripheral 100. The multifunction peripheral 100 generally includes a reading unit 110 that optically reads an image from a document, and an image forming unit 120 that forms an image corresponding to the image read by the reading unit 110 on a sheet (recording material). .

<1-1. Reading unit 110>
FIG. 4 is a diagram illustrating a configuration of the reading unit 110 of the multifunction machine 100. The ADF 11 is an automatic document feeder that conveys documents one by one to a reading position. The user can automatically scan a plurality of documents one by one to the document reading position by using the ADF 11, or can manually scan the documents one by one without using the ADF 11. You can also scan it on top. When scanning using the ADF 11, the user sets a document on the tray 13. The tray 13 is provided with a sensor (not shown), detects the presence of a document set in the tray 13, and outputs a signal indicating that. The documents are conveyed one by one to the conveying roller 15 by the drawing roller 14. The transport roller 15 transports the document toward the platen glass 16 while changing the document transport direction. The document conveyed in this manner is pressed against the platen glass 16 by the back platen 17 and is finally discharged from the ADF 11 by the discharge roller 18. On the platen glass 12, four document reading positions are provided from the upstream to the downstream of the conveyance path. At each of these reading positions, the document is conveyed at a constant speed. Reflected light (original image) irradiated from the light source 20a at the reading position and reflected by the original is changed in optical path by the mirror 20b, mirror 21a, and mirror 21b, condensed by the lens 22, and CCD (Charge Coupled Devices). An image is formed on the sensor 34. The CCD sensor 23 is, for example, a 4-line CCD sensor, and outputs analog image signals R, G, B, and K to the subsequent circuit in accordance with the reflected light (original image) input at each reading position. When the reading is completed, the document is conveyed by the conveyance roller 16 and is discharged to the discharge tray 19 via the discharge roller 18.

  When scanning without using the ADF 11, the user sets originals one by one on the platen glass 12. When a scan instruction is input by a method such as pressing the start button of the operation input unit 34, the first CRG 20 including the light source 20a reads a document image while moving in the direction A in FIG. The CCD sensor 23 outputs analog image signals R, G, B, and K to the subsequent circuit as in the case of using the ADF 11.

  Analog image signals R, G, B, K output from the CCD sensor 23 are converted into digital image data R, G, B, K by the A / D conversion circuit 30. These digital image data R, G, B, and K are corrected by a shading correction circuit or the like (not shown) corresponding to the sensitivity variation of the CCD sensor 23 and the light quantity distribution characteristics of the optical system. The digital image data R, G, B, and K corrected in this way are output to the subsequent image processing circuit 31. The image processing circuit 31 converts the digital image data R, G, B, and K into image data Y, M, C, and K corresponding to each color of yellow (Y), magenta (M), cyan (C), and black (K). Convert to K. Hereinafter, part or all of the digital image data Y, M, C, and K is simply referred to as “image data”.

The CCD drive circuit 33 controls the CCD sensor 23 according to a signal (command) output from the CPU 50. The ADF drive circuit 32 controls the ADF 11 according to a signal (command) output from the CPU 50. The ADF drive circuit 32 has a function of measuring the number of conveyed documents. The CPU 50 reads and executes various programs stored in the ROM 52 or the HDD 53 using the RAM 51 as a work area. The display unit 36 displays an execution screen or the like corresponding to the processing of the program executed by the CPU 50. The operation input unit 34 includes a touch panel formed on the display unit 36, a numeric keypad, a start key, a stop key, and the like. The operation input unit 34 outputs a signal corresponding to the image displayed on the display unit 36 and a user operation input to the CPU 50.
<1-2. Image Forming Unit 120>
With reference to FIG. 3 again, the image forming unit 120 of the multifunction peripheral 100 will be described. The image forming unit 120 rotates in the direction of an arrow a in FIG. 3 and forms a toner image of each color of Y, M, C, and K, and charging for uniformly charging the photosensitive drum 71. , An exposure device 73 that forms an electrostatic latent image of each color of YMCK by irradiating the charged photosensitive drum 71 with laser light with exposure light modulated based on image data of each color of YMCK, and each color of YMCK A developing unit 74 that develops an electrostatic latent image formed on the photosensitive drum 71 with toner, thereby forming a toner image of each color of YMCK on the photosensitive drum 71, and a photosensitive after transfer. A drum cleaner 75 for collecting toner remaining on the surface of the body drum 71.

  Further, the image forming unit 120 includes an intermediate transfer belt 76 on which a toner image formed on the photosensitive drum 71 is temporarily transferred. The intermediate transfer belt 76 is an endless belt member, is stretched by a drive roller 77, a primary transfer roller 78, a secondary transfer roller 79, and the like, and is circulated and moved by the drive roller 77 in the direction of arrow b in the figure. On the intermediate transfer belt 76, a belt cleaner 80 for collecting toner remaining after the secondary transfer is provided.

  The primary transfer roller 78 forms a nip region with the photosensitive drum 71 across the intermediate transfer belt 76, and in this nip region, the toner image on the circumferential surface of the photosensitive drum 71 is transferred to the intermediate transfer belt 76 by pressure contact force and electrostatic force. To do. The toner image transferred onto the intermediate transfer belt 76 is conveyed by the rotation of the intermediate transfer belt 76, and is conveyed while the timing is adjusted by the registration roller 82 in the nip region formed by the secondary transfer roller 79 and the backup roller 81. Secondary transferred onto the incoming paper.

  The paper is supplied by either the cut paper tray 85 or the manual feed tray 86 and is transported along a paper transport path 88 having a plurality of transport rolls 87. On the paper conveyance path 88, a conveyance belt 89, a fixing device 90, and a discharge roll 93 are provided. The sheet on which the toner image is transferred in the nip region formed by the secondary transfer roller 79 and the registration roller 82 is conveyed to the fixing device 90 by the conveyance belt 89. The fixing device 90 fixes the toner image on the paper by heating and pressurizing the paper on which the toner image is fixed. The discharge roll 93 discharges the paper on which the toner image is fixed to the paper discharge tray 94.

<1-3. Translation unit 130>
FIG. 5 is a block diagram illustrating a functional configuration of the translation unit 130 of the multifunction peripheral 100. The translation unit 101 refers to the translation dictionary 107 stored in the storage unit 106 and performs a translation process from the original sentence to the translated sentence. The timer 105 has a function for measuring time and a function for outputting time in response to a request. The processing time prediction unit 102 has a function of predicting the time required for the translation processing based on the original text data and the time data output from the timer 105. Based on the processing time predicted by the processing time prediction unit 102, the determination unit 103 determines whether the multifunction peripheral 100 can process the requested translation process. The output unit 104 outputs the translation text data generated by the translation unit 101. Each functional component described above is realized by the CPU 50 of the multifunction peripheral 100 reading and executing the translation program stored in the HDD 53. Details of each function will be described later.

In the present embodiment, the multi-function device 100 basically executes a translation process according to the following procedure.
(1) Pre-edit processing First, the CPU 50 of the multifunction peripheral 100 performs simple pre-processing on the original text in order to facilitate the translation processing. For example, for the sentence “This is a pen.”, A space of one character is inserted between the last word “pen” and a period. As a result, the sentence to be processed becomes “This is a pen.”. In the next dictionary lookup step, when a dictionary is looked up, a space is used as a mark to recognize a break between words, and this processing is performed to facilitate recognition of a break between words.

(2) Dictionary Lookup Subsequently, the CPU 50 regards the original text to be processed from blank to blank as one word and performs dictionary lookup with reference to the translation dictionary 107. Idioms composed of a plurality of words are also adopted as translation candidates as long as they are registered in the dictionary. The CPU 50 preferentially adopts the longest idiom as a translated word. When the dictionary lookup is completed, the j-th candidate word of the i-th English word in the sentence to be processed is assigned to the variable aj (i, j), and the attributes of those candidate words are assigned to the variable aa (i, j). . As an attribute stored in the variable aa (i, j), for example, there is a part of speech.

(3) Selection of proper translation Subsequently, the CPU 50 selects a proper translation based on the proper translation selection rule stored in the HDD 53, and stores the proper translation selected for the i-th English word in the variable cj (i). .

(4) Phrase Detection and Phrase Translation Sentence Generation Subsequently, the CPU 50 detects a phrase and generates a phrase translation sentence. Specifically, the CPU 50 reads out the grammar rule file stored in the HDD 53, and the part-of-speech code arrangement pattern (part-of-speech code string) recorded in the grammar rule file and the part-of-speech of words constituting the original sentence. The phrase is detected by comparing the code string. The CPU 50 generates a translation sentence by rearranging them, changing the ending, or supplementing the case particles based on the words constituting the phrase. At this time, a translated sentence can be generated by referring to information on a phrase before and after the phrase, or by referring to context information of a sentence before the target sentence.

The multi-function device 100 performs the translation process by executing the processes (1) to (4) described above. In this embodiment, the multi-function device 100 translates the translated text with different accuracy according to the user's request. Can be generated. That is, the multi-function device 100 can perform translation processing at five levels from level 1 (low accuracy) to level 5 (high accuracy). In each processing step described above, five levels of processing from level 1 to level 5 are defined. For example, in the (2) dictionary lookup step, it is defined that only a general dictionary is used in level 1 processing, and all specialized dictionaries stored in the HDD 53 are used in level 5. Also, for example, (4) in the phrase detection and phrase translation generation step, this process is skipped in the level 1 process, and the translation is performed with reference to only the target phrase information in the level 2 process. Level 5 stipulates that translation is performed with reference to information on a sentence two before the subject sentence. In this way, the multi-function device 100 can generate translated sentences with different accuracy.
Note that the above translation algorithm is merely an example, and a translation process may be performed using an algorithm other than this.

<2. Configuration of Translation Server 200>
FIG. 6 is a block diagram illustrating a hardware configuration of the translation server 200. The CPU 201 executes a program read from the ROM 203 or the HDD 204 using the RAM 202 as a work area. The HDD 204 stores data of n dictionaries (dictionaries 206-1 to 206-n) used for translation processing in addition to various programs and data. These n dictionaries have different vocabularies, for example, the dictionary 206-1 is a general-purpose dictionary, the dictionary 206-2 is a physics and chemistry term specialized dictionary, and the dictionary 206-3 is a patent term specialized dictionary. The translation server 200 can exchange data with an external device via the I / F 205. Further, the translation server 200 has a function of generating and sending back translated text data for the received original text when receiving the original text data from an external device.

  Similarly to the multifunction machine 100, the translation server 200 can also provide translations at levels 1 to 5. Furthermore, since the translation server 200 has a higher processing capacity of the processor and a storage capacity of the storage device such as the HDD 204 than the multi-function device 100, it can provide higher level level 6 to level 10 translated sentences. That is, the translation server 200 can provide 10 levels of translation sentences of level 1 to level 10. The multifunction peripheral 100 transmits original text data to the translation server 200 as necessary, as will be described later. The translation server 200 performs a translated sentence generation process on the received original sentence data, and returns the generated translated sentence data to the MFP 100. In this way, the multi-function device 100 can provide the user with translated sentences with 10 levels of accuracy from level 1 (minimum accuracy) to level 10 (maximum accuracy). As the level of translation processing increases, the MFP 100 or the translation server 200 is required to perform processing with a heavy load, and the processing time increases.

<B. Operation of translation system 1>
Next, the operation of the translation system 1 according to this embodiment will be described. In the following operation example, a user reads an English document printed on paper into the multifunction device 100, the multifunction device 100 performs a translation process, and prints a Japanese translation on paper (recording material). A mode of outputting will be described.

  FIG. 7 is a flowchart showing the operation of the translation system 1. First, the user sets a document on which the original English text is printed on the tray 13 of the ADF 11. The user operates the touch panel of the operation input unit 34 to display a menu screen. The user selects “Translate” from the menu. When “Translate” is selected, a screen as shown in FIG. The user can change the setting level of translation accuracy by operating the touch panel. By default, translation accuracy is set to level 5. Further, the user can designate the language of the original sentence and the language of the translated sentence by setting the “source language” or “translated sentence language” field to an appropriate one on this screen.

When the user touches the “Start” button on the execution screen shown in FIG. 8, the CPU 50 updates the variable Lv indicating the translation level to a value shown on the screen and stores it in the RAM 51. Further, the CPU 50 sets a variable L ₁ indicating the language of the original sentence and a variable L ₂ indicating the language of the translated sentence to values shown on the screen, and stores them in the RAM 51. Subsequently, the CPU 50 conveys the original set on the ADF 11 to the reading position one by one and outputs an instruction to scan. As a result, the originals are read one by one, and corresponding image data is generated (step S101). The CPU 50 stores the generated image data in the HDD 53.

  A description will be given with reference to FIG. 7 again. The CPU 50 measures the time required for the translation processing for a partial image of a part of the image data (step S102). Specifically, the CPU 50 first reads the current time from the timer 105 and stores the read time in the RAM 51 as the processing start time. Subsequently, the CPU 50 reads the image data, and performs character recognition processing on a partial image of a part of the image data. In this embodiment, the part to be subjected to character recognition processing is determined in advance as “the lower half of the first page” by the translation program. Therefore, the CPU 50 performs character recognition processing on the data corresponding to the lower half of the first page of the image data, and stores the extracted character string in the RAM 51.

  Next, the CPU 50 performs a translation process on the extracted character string. The CPU 50 reads the variable Lv indicating the translation level from the RAM 51, and performs the processes of the steps (1) to (4) according to the translation level specified by the variable Lv. When the translation is completed, the CPU 50 stores the generated translated text data in the HDD 53. When the above processing is completed, the CPU 50 reads the current time from the timer 105 and stores the read time in the RAM 51 as the processing end time.

Next, the CPU 50 calculates a translation processing prediction time (step S103). Specifically, the CPU 50 calculates a time t ₁ required for the translation process of the partial image from the difference between the process end time and the process start time. The CPU 50 calculates the processing time per unit page by dividing t ₁ by the number of pages of partial images (0.5 pages in the present embodiment). CPU50 is, the processing time per unit page, and the translation processing prediction time t ₂ is multiplied by the number of pages n of the read image. That is,
t ₂ = t ₁ /0.5×n (1)
It is. The CPU 50 stores the calculated translation processing prediction time t ₂ in the RAM 51.

Next, the CPU 50 determines whether or not processing is possible within a predetermined time (step S104). Specifically, the CPU 50 reads a constant t _th indicating the allowable time for translation processing from the HDD 53. In the present embodiment, the processing allowable time t _th is predetermined by the translation program. The CPU 50 reads the translation processing predicted time t ₂ and the processing allowable time t _th from the RAM 51 and compares the two. When t ₂ > t _th , the CPU 50 determines that processing is impossible within the processing allowable time t _th . When t ₂ ≦ t _th , the CPU 50 determines that processing is possible within the processing allowable time t _th .

When it is determined that the processing is impossible within the processing allowable time t _th (step S104: NO), the multi-function device 100 causes the translation server 200 to perform the translation processing. That is, the CPU 50 transmits the original text image data to the translation server 200 (step S107). When transmitting data, the multi-function device 100 adds information such as an ID for identifying itself to the image data. The CPU 201 of the translation server 200 that received the original text image data performs character recognition processing and translation processing on the received data, and generates translated text string data (step S108). The CPU 201 transmits the generated character string data to the multi-function device 100 specified by the ID added to the image data (step S109). Upon receiving the character string data of the translated sentence, the multi-function device 100 processes it into an appropriate form and outputs it (step S106).

When it is determined that the processing is possible within the processing allowable time t _th (step S104: YES), the CPU 50 performs character recognition processing and translation processing on the entire original image data to generate character string data of the translated text. (Step S105). The CPU 50 stores the generated character string data in the HDD 53. The CPU 50 processes the character string data stored in the HDD 53 into an appropriate form and outputs it (step S106).

  The character string data of the translation sentence output from the CPU 50 is formed as an image on a sheet (recording material) by the image forming unit 120 and output. That is, when the user sets an original document in which the original text is described in the multifunction peripheral 100 and presses a button to instruct the translation process, the translation process is performed. The multi-function device 100 prints the generated translation sentence on paper and outputs it. Therefore, the user does not need to type in English from the keyboard or store the original text data in a recording medium such as a CD-ROM to input the original into the translator, and the original is printed. You can get the translation if you have paper (manuscript).

  If the multifunction peripheral 100 according to the present embodiment is installed at, for example, a store of a convenience store, the user can obtain a translation of the document as if making a copy. Also, during translation processing, the MFP 100 determines whether the requested translation processing can be processed by itself, and causes only the processing that is determined to be unprocessable to be processed by a translation server having high processing capability. Since only a simple sentence is processed by itself, a translation result can be obtained quickly for a sentence that requires only a small load such as a very short sentence or a simple sentence. Furthermore, in the MFP 100 according to the present embodiment, the user can select the translation accuracy. For example, a user who wants to know only the meaning of a word appearing in a sentence sets the translation level to “1”. Even if the accuracy is poor, the translation result can be obtained quickly. Conversely, a user who wants a translation that is as accurate as possible may take time, but setting the translation level as high as “10” will increase the processing time, but it will be accurate using a dedicated translation server. You can get a translation. In this way, according to the multifunction peripheral 100 according to the present embodiment, it is possible to meet various requests for user translation.

<C. Modification>
As mentioned above, although one Embodiment of this invention was described, this invention is not limited to the above-mentioned embodiment, A various deformation | transformation implementation is possible.
In the above-described embodiment, the aspect in which the multifunction peripheral 100 is used as the automatic translation apparatus has been described. However, the automatic translation apparatus according to the present invention is not limited to the multifunction peripheral 100, and may be an apparatus such as a personal computer or a PDA. The network that mediates between the multifunction peripheral 100 and the translation server 200 is not limited to the Internet 300, and may be a network such as a LAN or a WAN. In addition, these networks may include a radio area in part or all thereof. Moreover, the original text and the translated text are not limited to English and Japanese, and can be applied to any language.

Further, in the above-described embodiment, the aspect has been described in which the MFP 100 performs processing when the MFP 100 determines that its processing capability is insufficient. However, the MFP 100 automatically sets the translation level. It is good also as a structure to reduce to. That is, when it is determined that the estimated translation processing time t ₂ exceeds the allowable processing time t _th (step S104 in FIG. 7: NO), the CPU 50 decreases the translation level Lv by 1 and again estimates the processing time (step S102). To S103). In this way, the CPU 50 may reduce the translation level until it is determined that processing is possible.

  In the above-described embodiment, the configuration including the MFP 100 and the translation server 200 to which the translation system 1 is connected via the network has been described. However, the translation system 1 includes only the stand-alone MFP 100. It is good also as a structure. In this configuration, when the multi-function device 100 determines that its own processing capability is insufficient, the configuration may be such that the translation level is automatically lowered as described above.

  In the above-described embodiment, the mode in which the original image data is transmitted to the translation server 200 when the MFP 100 determines that the processing capability of the multifunction peripheral 100 is insufficient has been described. At the time of transmission, processing for reducing the load on the network such as compression processing may be performed. Alternatively, the data may be encrypted and transmitted in order to improve security. Alternatively, the MFP 100 may extract the character string of the entire original text by performing character recognition processing on the image data of the entire original text, and transmit the extracted character string data to the translation server 200. In this case, it is not necessary to perform character recognition processing in the translation server 200, and the translation server 200 may perform translation processing on the received data.

Further, in the above-described embodiment, the aspect in which t _th is determined in advance by the translation program has been described. However, as shown in FIG. 9, the configuration may be such that the user can specify t _th on the translation processing instruction screen. Good. That is, when the “start” button is pressed by the CPU 50, the value displayed in the “processing allowable time” field of the execution screen is stored in the RAM 51 as the value of t _th . Similarly, as shown in FIG. 9, the calculated processing predicted time may be displayed on the display unit 36.

  Further, the translation server 200 may have a known learning function. Alternatively, the translation server 200 may store user dictionary data corresponding to the user and perform translation using this user dictionary. In this case, the CPU 50 adds the user dictionary ID for designating the user dictionary to the image data in step S107 in FIG. The user dictionary ID may be configured to be manually input by the user by operating the operation input unit 34, or may be configured to be input by extracting a character string described on the document by character recognition processing.

  Further, in the above-described embodiment, in a certain translation algorithm, a translation sentence with different accuracy is provided by changing the processing content in each step according to the translation level. It is not limited to. For example, a plurality of translation algorithms with completely different processing contents may be prepared, and these algorithms may be switched according to the required translation accuracy. Or it is good also as a structure which prepares the translation algorithm comprised from many steps, and changes a translation system by abbreviate | omitting some steps among these according to the required translation precision.

  In the above-described embodiment, the image of the region (lower half of the first page) determined in advance by the translation program is extracted as a partial image serving as a reference for calculating the predicted processing time. It is not limited to. The CPU 50 may be configured to perform a well-known layout extraction process on the document image data and extract a region satisfying a specific condition as a partial image. As the specific condition, for example, information on the area of the region and the position of the region in the page can be used.

  Further, in the above-described embodiment, the process prediction time is calculated using Expression (1), but the method of calculating the translation process prediction time is not limited to this. For example, the CPU 50 may estimate the number of characters in the original text from the density information of the image data, and calculate by multiplying the processing time per unit character obtained from the time required for processing the partial image by the number of characters in the entire document. .

Alternatively, the translation processing prediction time may be calculated by the following method. That is, a table in which the translation level Lv and / or the identifier of the dictionary used and the translation time t ₃ per unit character (time / number of characters, time / number of pages, or time / area) are recorded in the HDD 53 in association with each other. May be stored in advance, and the translation processing prediction time may be calculated by multiplying this by the amount of translation (the number of characters, the number of pages, or the area of the character portion). In this case, the translation amount may be specified by a user's instruction input. The user may specify the number of pages, or may specify the number of pages and the amount of characters, and the CPU 50 may calculate the processing amount by multiplying them. The character amount may be specified by allowing the user to select one from the choices of “large”, “normal”, and “small”. In this case, the CPU 50 calculates the translation fee by multiplying the number of characters associated with each option and the number of pages. Alternatively, the ADF 11 may be used to read all originals and perform character recognition processing, and the number of recognized characters may be used as the translation amount. Also, character recognition processing is performed on one page of the document, and the amount of translation is calculated by multiplying the number of recognized characters by the number of pages specified by the user (or the number of pages measured by the ADF 11). Also good.
Among such aspects, for those that do not need to measure the processing time of a part of the document, the processing corresponding to step S102 in FIG. 7 such as extraction of partial images may not be performed.

In the above-described embodiment, the MFP 100 automatically selects the subject of translation processing. However, the user may select the subject of translation processing. In this case, the CPU 50 may calculate a parameter reflecting both the translation accuracy and the processing time, and display the parameter on the display unit 36. The user can determine the processing subject based on this parameter. As a parameter for this purpose, for example, a value obtained by standardizing the translation level (1 to 10) divided by the predicted processing time t ₂ can be used.

  In addition, when the multifunction device 100 performs character recognition processing and translation processing of a partial image, the MFP 100 extracts information specifying the genre of the original text, adds the information to the data, and transmits the information to the translation server 200. Good. The translation server 200 can specify the dictionary used for the translation process according to the information and perform the translation process. As a method for specifying the genre of the original text, for example, a database in which a document genre and a word characteristic of a document of the genre are registered in advance in the HDD 53 is stored and registered in the database. When a word is present in a document in a certain ratio or more, the document may be determined to be a genre document corresponding to the characteristic word.

  In the above-described embodiment, the functions of the translation unit 130 of the multifunction peripheral 100 are realized by the CPU 50 executing the translation program. However, a part or all of the functions may be realized by hardware. That is, the translation dictionary 107 may have an electronic circuit that performs processing corresponding to each functional component shown in FIG.

It is a figure which shows the structure of the translation system 1 which concerns on one Embodiment of this invention. 2 is a diagram illustrating a functional configuration of the multifunction peripheral 100. FIG. 2 is a diagram illustrating a hardware configuration of the multifunction machine 100. FIG. 2 is a diagram illustrating a configuration of a reading unit 110 of the multifunction peripheral 100. FIG. 3 is a block diagram illustrating a functional configuration of a translation unit 130 of the multifunction peripheral 100. FIG. 3 is a block diagram showing a hardware configuration of a translation server 200. FIG. 3 is a flowchart showing the operation of the translation system 1. It is a figure which illustrates the execution screen of a translation program. It is a figure which illustrates the execution screen of the translation program which concerns on another embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Translation system, 11 ... ADF, 12 ... Platen glass, 13 ... Tray, 14 ... Pull-in roller, 15 ... Conveyance roller, 16 ... Platen glass, 17 ... Back platen, 18 ... Discharge roller, 19 ... Discharge tray, 20 ... 1st CRG, 22 ... lens, 23 ... CCD sensor, 30 ... A / D conversion circuit, 31 ... image processing circuit, 32 ... ADF drive circuit, 33 ... CCD drive circuit, 34 ... operation input unit, 36 ... display unit, 50 ... CPU, 51 ... RAM, 52 ... ROM, 53 ... HDD, 71 ... photosensitive drum, 72 ... charger, 73 ... exposure device, 74 ... rotary developer, 75 ... drum cleaner, 76 ... intermediate transfer belt, 77 ... Drive roller 78 ... Primary transfer roller 79 ... Secondary transfer roller 80 ... Belt cleaner 81 ... Backup roller 82 ... Registration roller 8 ... Cut paper tray, 86 ... Bypass tray, 87 ... Conveying roll, 88 ... Paper conveying path, 89 ... Conveying belt, 90 ... Fixing device, 93 ... Discharging roll, 94 ... Discharging tray, 100 ... Multifunction machine, 101 ... Translation , 102 ... processing time prediction unit, 103 ... determination unit, 104 ... output unit, 105 ... timer, 106 ... storage unit, 107 ... translation dictionary, 110 ... reading unit, 120 ... image forming unit, 130 ... translation unit, 140 ... Control unit, 200 ... Translation server, 201 ... CPU, 202 ... RAM, 203 ... ROM, 204 ... HDD, 205 ... I / F, 300 ... Internet.

Claims (5)

Storage means for storing original text data described in a first language and information designating an allowable time for translation processing for translating the original text into a translated text described in a second language;
Translation means capable of executing a plurality of translation processes on the original text;
Selecting means for selecting one of the plurality of translation processes;
A translation time prediction means for calculating a prediction time of a process when a translation sentence is generated by one translation process selected by the selection means for the original sentence;
An automatic translation apparatus comprising: a determination unit configured to determine whether generation of a translation sentence by the first translation process can be performed within the allowable time based on the predicted time and the allowable time. A transmission means for transmitting a part or all of the original text data when it is determined by the determination means that the generation of the translated text by the translation process of 1 cannot be processed within the allowable time;
Receiving means for receiving translated sentence data for part or all of the original sentence data transmitted by the transmitting means; and output means for outputting a translated sentence reflecting the translated sentence data received by the receiving means. The automatic translation apparatus according to claim 1. Reading means for optically reading a document and generating image data corresponding to the document;
The character recognition means further comprising: character recognition means for performing character recognition processing on the image data generated by the reading means, and storing the obtained character string data in the storage means as the original text data. Translation device. The automatic translation apparatus according to claim 3, wherein the output unit includes an image forming unit that forms an image corresponding to the translated sentence on a sheet-like recording material. A terminal device;
A server device that communicates with the terminal device via a network;
The terminal device is
Storage means for storing original text data described in a first language and information designating an allowable time for translation processing for translating the original text into a translated text described in a second language;
First translation means capable of executing a plurality of translation processes on the original text;
Selecting means for selecting one of the plurality of translation processes;
A translation time prediction means for calculating a prediction time of a process when a translation sentence is generated by one translation process selected by the selection means for the original sentence;
A determination means for determining whether the generation of the translated sentence by the first translation processing can be processed within the allowable time based on the predicted time and the allowable time;
An original sentence transmitting means for transmitting a part or all of the original sentence data when it is determined by the determining means that the generation of the translated sentence by the one translation process cannot be processed within the allowable time;
A translated sentence receiving means for receiving translated sentence data for a part or all of the original sentence data transmitted by the original sentence transmitting means;
An output means for outputting a translated sentence reflecting the translated sentence data received by the translated sentence receiving means;
Have
The server device is
Original text receiving means for receiving data of the original text transmitted by the original text transmitting means;
Second translation means for generating a translation sentence from the data received by the original sentence reception means;
An automatic translation system comprising: a translated sentence transmission means for transmitting the translated sentence data generated by the second translation means to the terminal device.

Images