JP2020071470A - Information processing system and transcription method - Google Patents

Abstract

To enhance confidentiality of contents of transcription target speech.SOLUTION: A management device 12 divides target speech data, in which a transcription target speech is recorded, into multiple sectional speech data pieces over multiple sections. The management device 12 provides an operator device 16 with a set of at least one of the multiple sectional speech data pieces and dummy speech data. The management device 12 receives text data transcribed from the at least one of the multiple sectional speech data pieces, and/or text data transcribed from the dummy speech data. The management device 12 generates text data transcribed from the target speech by using the text data transcribed from the at least one of the multiple sectional speech data pieces among the received text data.SELECTED DRAWING: Figure 1

Description

  The present invention relates to data processing technology, and more particularly to an information processing system and a transcription method.

  A transcription system for raising a character from a conversation recorded as voice has been proposed (for example, see Patent Document 1). In the transcription system of Patent Document 1, the server divides the voice data in which the conversation is recorded into voice data relating to a plurality of voice sections, and transmits the voice data of each voice section to a plurality of information terminals. Each information terminal outputs a character string transcribed from voice data to the server, and the server combines the individual character strings to construct sentence data in which the entire conversation of the original voice data is converted into a sentence.

JP, 2008-107624, A

  It may not be desirable for the voice to be transcribed to be transmitted as it is to the operator who performs the transcription, such as when confidential information is included. The present inventor considered that there is room for improvement in order to increase the confidentiality of the content of the voice to be transcribed.

  The present invention has been made on the basis of the present inventors' recognition of the above problems, and one object thereof is to improve the confidentiality of the content of the voice of the transcription target.

  In order to solve the above-mentioned problems, an information processing system according to an aspect of the present invention includes a first storage unit that stores target voice data in which a voice of a transcription target is recorded, and a dummy voice in which a dummy voice is recorded. A second storage unit that stores data, a dividing unit that divides target voice data into a plurality of section voice data related to a plurality of sections, and a set of at least one of the plurality of section voice data and dummy voice data to an external device And a reception unit that receives the text data transcribed based on at least one of the plurality of section voice data and the text data transcribed based on the dummy voice data, and the text received by the reception unit. Text data obtained by transcribing the target voice using text data obtained by transcribing at least one of a plurality of section voice data among the data. Comprising a generation unit for generating data.

  Another aspect of the present invention is a transcription method. According to this method, an information processing system that stores target voice data in which a voice to be transcribed is recorded and dummy voice data in which a dummy voice is recorded are stored in a plurality of sections in which the target voice data is divided into a plurality of sections. Dividing into voice data; providing a set of at least one of the plurality of section voice data and dummy voice data to an external device; and text data transcribed based on at least one of the plurality of section voice data. A step of receiving both the text data transcribed based on the dummy voice data, and the text data transcribed based on at least one of the plurality of section voice data among the received text data, And a step of generating text data in which the voice is transcribed.

  It should be noted that any combination of the above constituent elements and the expression of the present invention converted between an apparatus, a computer program, a recording medium storing the computer program, and the like are also effective as an aspect of the present invention.

  According to the present invention, it is possible to improve the confidentiality of the content of the voice of the transcription target.

It is a figure which shows the structure of the transcription system of an Example. It is a block diagram which shows the functional block of the management apparatus of FIG. It is a figure which shows the example of division | segmentation of audio | voice data. It is a figure which shows the example of audio | voice data. It is a figure which shows the example of allocation of section audio | voice data. It is a figure which shows the example of the work result by a worker. It is a figure which shows the example of division | segmentation of audio | voice data.

  The transcription system of the embodiment provides a worker who performs transcription with a set of at least a part of a voice to be transcribed (a voice relating to a user and a voice that may include secret information) and a dummy voice. Provide both of those sounds to the operator to transcribe. As a result, it is possible to reduce the risk of leakage of the entire content of the voice to be transcribed and to increase the confidentiality of the content of the voice to be transcribed.

  FIG. 1 shows the configuration of a transcription system 10 according to an embodiment. The transcription system 10 is an information processing system that supports transcription, and includes a management device 12, a plurality of user terminals 14, and a plurality of worker devices 16. Each device of the transcription system 10 is connected via a communication network 18 including a LAN, WAN, the Internet, and the like. Transcription is the conversion of audio content into text, which can also be called tape transcription.

  The management device 12 is an information processing device that provides a web service of transcription to a plurality of user terminals 14. Detailed functions of the management device 12 will be described later.

  The plurality of user terminals 14 are information processing devices operated by a user who uses the transcription service. The plurality of user terminals 14 are operated by a user terminal 14a operated by a user a belonging to Company A, a user terminal 14b operated by a user b belonging to Company B, and a user c belonging to Company C. The user terminal 14c is included. The user terminal 14 may be a PC, a tablet terminal, or a smartphone.

  The plurality of worker devices 16 are information processing devices that mainly perform transcription. In the embodiment, a human listens to a voice and converts the voice into text. The plurality of worker devices 16 include a worker device 16x operated by a worker x, a worker device 16y operated by a worker y, and a worker device 16z operated by a worker z. The worker device 16 may be a PC, a tablet terminal, or a smartphone.

  FIG. 2 is a block diagram showing functional blocks of the management device 12 of FIG. Each block shown in the block diagram of this specification can be realized by an element such as a processor, a CPU, and a memory of a computer, an electronic circuit, and a mechanical device in terms of hardware, and by a computer program or the like in terms of software. However, here, the functional blocks realized by those collaborations are drawn. Therefore, it will be understood by those skilled in the art that these functional blocks can be realized in various ways by a combination of hardware and software.

  The management device 12 includes a control unit 20, a storage unit 22, and a communication unit 24. The control unit 20 executes various data processing for providing a transcription service. The storage unit 22 stores data referred to or updated by the control unit 20. The communication unit 24 communicates with an external device according to a predetermined communication protocol. The control unit 20 transmits / receives data to / from the user terminal 14 and the worker device 16 via the communication unit 24.

  The storage unit 22 includes a target voice storage unit 30, a dummy voice storage unit 32, an allocation rule storage unit 34, a distribution data storage unit 36, a work result storage unit 38, a text storage unit 40, a correct answer storage unit 42, and an evaluation storage unit 44. Including. At least a part of the data stored in the storage unit 22 may be stored in a storage device (not shown) different from the management device 12, and the management device 12 stores data stored in an external storage device. Of course, you may refer to and update.

  The target voice storage unit 30 stores target voice data which is voice data received from the user terminal 14 and in which a voice to be transcribed (hereinafter also referred to as “target voice”) is recorded. The dummy voice storage unit 32 stores not the voice data received from the user terminal 14, but the dummy voice data in which a dummy voice (hereinafter also referred to as “dummy voice”) is recorded. The dummy voice is a voice having contents predetermined by the administrator of the management device 12 (for example, the person in charge of the company that provides the transcription service).

  The correct answer storage unit 42 stores text data indicating the content of the dummy voice. In the embodiment, the dummy voice data is preliminarily divided into voice data for each section (hereinafter also referred to as “section voice data”) related to a plurality of sections, and the dummy voice storage unit 32 stores a plurality of pieces based on the dummy voice data. It is assumed that the section voice data of is stored. Further, the correct answer storage unit 42 is supposed to store text data (hereinafter also referred to as “correct answer data”) indicating the contents of each of the plurality of section voice data based on the dummy voice data.

  The assignment rule storage unit 34 stores the section voice data in which the target voice data is divided and the rule (hereinafter, also referred to as “assignment rule”) for assigning the section voice data in which the dummy voice data is divided to the worker. .. The allocation rule will be described later in relation to the configuration of the allocation unit 56.

  The distribution data storage unit 36 is data to be distributed to each of a plurality of workers, and stores distribution data including one or more section voice data. For example, the distribution data storage unit 36 stores the distribution data for the worker x (worker device 16x), the distribution data for the worker y (worker device 16y), and the distribution data for the worker z (worker device 16z). Memorize

  The work result storage unit 38 stores text data as a result of transcription by the worker x (worker device 16x), text data as a result of transcription by the worker y (worker device 16y), and worker z (work). The text data that is the result of the transcription by the worker device 16z) is stored.

  The text storage unit 40 stores text data (hereinafter also referred to as “sentence data”) indicating the content of the entire target voice, which is generated by the text generation unit 62 described below. The evaluation storage unit 44 stores the evaluation results regarding a plurality of workers, which are generated by the evaluation unit 66 described below.

  The control unit 20 includes a request reception unit 50, a conversion unit 52, a division unit 54, an allocation unit 56, a distribution unit 58, a work result reception unit 60, a text generation unit 62, a text provision unit 64, and an evaluation unit 66. A computer program that implements the functions of the plurality of functional blocks may be stored in the storage unit 22. The processor of the management device 12 may exhibit the functions of the plurality of functional blocks of the control unit 20 by reading the computer program into the main memory and executing the computer program.

  The request receiving unit 50 receives a plurality of request data for requesting transcription of voice from a plurality of user terminals 14. The request receiving unit 50 stores the received request data in the target voice storage unit 30 in association with the request source user or the user terminal 14. The request data received from the user terminal 14a includes target voice data in which a voice (company's remark, a conference voice, etc.) regarding the company A is recorded. Further, the request data received from the user terminal 14b includes target voice data in which the voice regarding the company B is recorded. Further, the request data received from the user terminal 14c includes target voice data in which the voice regarding the company C is recorded.

  The conversion unit 52 converts at least one target voice data out of the plurality of target voice data received by the request receiving unit 50 by a known voice conversion function, so that the voice quality (pitch, pitch) of the plurality of target voice data is converted. Pressure, tone, etc.). This makes it difficult for an operator who has heard a plurality of section voice data to distinguish whether or not the original target voices of those section voice data are the same, and the confidentiality of the content of the target voice can be improved.

  In the embodiment, the conversion unit 52 converts the voice qualities of the plurality of target voice data received by the request reception unit 50 so as to be the same or similar to the voice qualities of the dummy voice data. This makes it difficult for an operator who has heard a plurality of section audio data to distinguish whether or not the original target sounds of the section audio data are the same, and also it is difficult to distinguish whether or not they are dummy sounds. Therefore, the confidentiality of the content of the target voice can be further enhanced.

  The division unit 54 divides the target voice data stored in the target voice storage unit 30 into a plurality of section voice data related to a plurality of sections. FIG. 3 shows an example of division of audio data. The dividing unit 54 divides the target voice data Aa of the company A into three pieces of section voice data Aa-1, section voice data Aa-2, and section voice data Aa-3. Further, the dividing unit 54 divides the target voice data Ab of the company A into three, that is, the section voice data Ab-1, the section voice data Ab-2, and the section voice data Ab-3. Similarly, the dividing unit 54 divides the target voice data Ba and the target voice data Bb of the company B.

  As described above, in the embodiment, the dummy voice data is divided into a plurality of section voice data in advance. For example, in FIG. 3, the dummy audio data Ca is divided into two, that is, interval audio data Ca-1 and interval audio data Ca-2. As a modification, the dividing unit 54 may divide the dummy voice data into a plurality of section voice data when dividing the target voice data.

  The division unit 54 stores, for each of the plurality of section voice data, position information (for example, the order from the beginning or the time position) in the target voice data or the dummy voice data before the division in the storage unit 22. For example, the division unit 54 stores information indicating that the section audio data Aa-1 is the first section of the target audio data Aa, and the section audio data Aa-2 includes 2 of the target audio data Aa. Information indicating the second section may be stored.

  FIG. 4 shows an example of audio data. The figure shows the waveform of the voice, and specifically, the horizontal axis of the figure shows the elapsed time from the start of the voice, and the vertical axis shows the volume. The dividing unit 54 holds a predetermined minimum time and maximum time (in other words, maximum length) that the section that divides the voice can take. In the embodiment, the minimum time of the section is 10 seconds (end range start point 70 in FIG. 4) and the maximum time is 20 seconds (end range end point 72 in FIG. 4). The shorter the section, the higher the confidentiality of the voice content, but the lower the accuracy of transcription. The minimum time and the maximum time of the section may be determined as appropriate values by weighing the confidentiality of the voice content and the accuracy of the transcription.

  When determining the end position of one section in the target audio data, the dividing unit 54 ends the section at a time point within a predetermined minimum time or more and maximum time or less and the sound volume is less than a predetermined threshold value. Determine as position. For example, in the example of FIG. 4, the time point of 15.5 seconds from the voice start is determined as the end position (division point 74) of the section. For the next section, the dividing unit 54 sets the division point 74 of FIG. 4 as the start position, and within the range of 10 seconds to 20 seconds from the division point 74 and when the volume is less than the predetermined threshold value, the next section. Is determined as the end position of. The volume threshold may be a volume value that is considered to be silent, or may be a volume value that is assumed in a quiet room. For example, the volume threshold may be 0.002 pascals (40 decibels).

  In the target voice, the volume of word breaks or meaning breaks tends to be low. In the embodiment, by setting the position where the volume is less than the threshold as the end position of the section, it is possible to set the word break or the break of the meaning as the end position of the section, and the accuracy of the transcription can be improved.

  Returning to FIG. 2, the assigning unit 56 sets a set of the section voice data based on the target voice data and the section voice data based on the dummy voice data according to the assignment rule stored in the assignment rule storage unit 34. Assign to each. The assigning unit 56 stores, in the distribution data storage unit 36, a set of section voice data based on the target voice data assigned to each worker and section voice data based on the dummy voice data.

  The allocation rule of the embodiment is set so that a plurality of section voice data allocated to one worker are temporally and spatially separated from each other. Specifically, (1) the allocation rule is that, when a plurality of worker devices 16 exist, one of the plurality of section voice data originating from one target voice data is set for one worker device 16. It is decided to allocate some section audio data. That is, the allocation rule prohibits allocation of all section audio data based on one target audio data to one worker device 16. Thereby, the confidentiality of the content of the target voice can be improved.

  Further, (2) allocation rule prohibits allocation of a plurality of time-sequential section voice data in one target voice data to one worker device 16. In other words, the allocation rule defines that a plurality of time-sequential section voice data in one target voice data is assigned to different workers. For example, the section voice data Aa-1 and the section voice data Aa-2 in FIG. 3 are prohibited from being assigned to the same worker, and are defined to be assigned to different workers. Thereby, the confidentiality of the content of the target voice can be further enhanced.

  Further, (3) the allocation rule is that when a plurality of section voice data is provided to one worker device 16, by providing a plurality of section voice data originating from one target voice data, It is determined that the provision of a plurality of section voice data originating from different target voice data has priority. The plurality of section voice data originating from different target voice data are likely to be unrelated in content, and thus the confidentiality of the content of each target voice can be further enhanced.

  Further, (4) the allocation rule is that when a plurality of section voice data originating from different target voice data is provided to one worker device 16, a plurality of section voice data originating from different target voice data regarding the same organization are provided. The provision of a plurality of section voice data originating from different target voice data related to different organizations is prioritized over the provision of the section voice data of. The plurality of section voice data originating from different target voice data relating to different organizations are more likely to be unrelated in content, and thus the confidentiality of the content of each target voice can be further enhanced.

  FIG. 5 shows an example of allocation of section voice data. The section voice data shown in the figure corresponds to the section voice data shown in FIG. In the example of the figure, the allocation unit 56 gives the worker x section voice data Aa-1 originating from the target voice data Aa and section voice data Ca-1 originating from the dummy voice data Ca. And the section voice data Bb-2 originating from the target voice data Bb are assigned. Further, the assigning unit 56 assigns a plurality of section voice data originating from different target voice data of different organizations to the worker y and the worker z, respectively.

  Returning to FIG. 2, the distribution unit 58 provides the section voice data to each worker according to the allocation result by the allocation unit 56. Specifically, the distribution unit 58 provides the worker device 16 of each worker with a set of section sound data based on the target sound data stored in the distribution data storage unit 36 and section sound data based on the dummy sound data. To do.

  In the embodiment, the delivery unit 58 transmits a web page (hereinafter also referred to as “work page”) for performing the transcription work to the plurality of worker devices 16 and displays the web page. The distribution unit 58 provides a work page for the worker x to the worker device 16x, a work page for the worker y to the worker device 16y, and a work page for the worker z to the worker device 16z. provide. The delivery unit 58 may notify the URL of the work page for each worker to the worker device 16 of each worker by e-mail or the like.

  The distribution unit 58 adds the section voice data (section voice data Aa-1, section voice data Ca-1, in the example of FIG. 5) assigned to the worker x by the assigning unit 56 to the data of the work page for the worker x. The section voice data Bb-2) is included. Similarly, the distribution unit 58 includes the section audio data assigned to the worker y (worker z) by the assigning unit 56 in the data of the work page for the worker y (worker z). Note that the distribution unit 58 arranges a button for reproducing each section audio data, an area for inputting a text as a result of transcribing the voice of each section audio data, and a transmission button on the work page for each worker. ..

  The work result reception unit 60 receives the work result of each worker transmitted from the worker device 16 of each worker. In the embodiment, the work result reception unit 60 receives the transcription result by the worker x input on the work page for the worker x, and the transcription result by the worker y input on the work page for the worker y. Is received, and the transcription result by the worker z input on the work page for the worker z is received. The work result acceptance unit 60 stores the work results of each worker in the work result storage unit 38.

  FIG. 6 shows an example of the work result by the worker. The figure shows the work result based on the allocation of FIG. For example, the work result by the worker x includes text data Aa-1, text data Ca-1, and text data Bb-2. The text data Aa-1 is a transcription of the voice of the section voice data Aa-1 which originates from the target voice data Aa. Further, the text data Ca-1 is a transcription of the voice of the section voice data Ca-1 originating from the dummy voice data Ca. Further, the text data Bb-2 is a transcription of the voice of the section voice data Bb-2, which originates from the target voice data Bb.

  Returning to FIG. 2, the sentence generator 62 receives characters from the text data received by the work result receiver 60 and stored in the work result storage 38 based on the section voice data originating from the target voice data. Using the transcribed text data, sentence data in which the entire target voice is transcribed is generated. When generating the sentence data, the sentence generator 62 does not use the text data transcribed based on the section voice data that originates from the dummy voice data.

  The sentence generation unit 62 combines text data based on a plurality of section voice data according to the position information of each section voice data stored in the storage unit 22 by the division unit 54 (that is, the position information in the target voice data). Generates text data by. The sentence generation unit 62 stores the sentence data for a certain target voice in the sentence storage unit 40 in association with the user (or the user terminal 14) who has requested the transcription of the target voice.

  The text data Aa-1 included in the work result of the worker x in FIG. 6 is the text of the section voice data Aa-1 corresponding to the first section of the target voice data Aa. The text data Aa-2 included in the work result of the worker y in FIG. 6 is the text of the section voice data Aa-2 corresponding to the second section of the target voice data Aa. Further, the text data Aa-3 included in the work result of the worker z in FIG. 6 is the text of the section voice data Aa-3 corresponding to the third section of the target voice data Aa. The sentence generator 62 synthesizes the text data Aa-1, the text data Aa-2, and the text data Aa-3 in this order to generate the sentence data Aa in which the entire target voice data Aa is converted to text.

  Returning to FIG. 2, the text providing unit 64 transmits the text data stored in the text storage unit 40 to the user (user terminal 14) that is the source of the transcription request. For example, the text providing unit 64 transmits the text data Aa obtained by transcribing the target voice data Aa shown in FIG. 6 to the user a (user terminal 14a) who has requested the transcription. Note that the text providing unit 64 receives the text data providing request from the user terminal 14a and, when the text data providing request is received, selects the text data associated with the user a from among the plurality of text data stored in the text storage unit 40. You may transmit to the user terminal 14a.

  The evaluation unit 66 includes correct answer data stored in advance in the correct answer storage unit 42, and a dummy transcription result of each worker (text data received by the work result receiving unit 60 and stored in the work result storing unit 38). Evaluate each worker by comparing and. For example, the evaluation unit 66 compares the correct answer data of the section voice Ca-1 originating from the dummy voice data Ca with the text data Ca-1 which is the transcription result of the section voice Ca-1 by the work x. Thus, the worker x is evaluated.

  In the embodiment, the evaluation unit 66 extracts the morphemes forming the correct answer data and the morphemes forming the text data of the work result by morpheme analysis, and the more morphemes that match between them, the more the conversion of the worker. It is evaluated that the accuracy is high, and a high evaluation value is given to the worker. The evaluation unit 66 may refer to the synonym dictionary and consider that the morphemes forming the correct answer data and the morphemes forming the text data of the work result do not match but may match as long as they are synonyms. As described above, in the embodiment, the worker can be objectively evaluated based on the transcription result of the dummy voice.

  The evaluation unit 66 stores the evaluation result (evaluation value) of each of the plurality of workers in the evaluation storage unit 44. The management device 12 may further include an evaluation result output unit (not shown) that provides the external device with the evaluation result of each worker stored in the evaluation storage unit 44. In this case, the external device may be, for example, a terminal of a person in charge who makes price negotiations or contracts with workers.

  The operation of the transcription system 10 having the above configuration will be described. Each of the plurality of users of the transcription system 10 uploads the target voice data from the user terminal 14 to the management device 12. The request receiving unit 50 of the management device 12 receives a plurality of target voice data transmitted from a plurality of user terminals 14. The conversion unit 52 of the management device 12 converts the voice qualities of the plurality of target voice data into a predetermined reference voice quality (in the embodiment, the voice quality is the same as the dummy voice data and may be the voice quality of the synthesized voice).

  The dividing unit 54 of the management device 12 divides each of the plurality of target voice data into a plurality of section voice data. The allocation unit 56 of the management device 12 allocates a set of the section voice data of the target voice data and the section voice data of the dummy voice data to each worker. The distribution unit 58 of the management device 12 presents each worker with a set of the section voice data of the target voice data and the section voice data of the dummy voice data on the web page for each worker.

  The worker reproduces the section voice data assigned to himself / herself on the web page for himself / herself and inputs the text indicating the voice content into a predetermined area of the web page. When the worker presses the send button on the web page, the worker device 16 transmits the text data input by the worker in the predetermined area to the management device 12.

  The work result acceptance unit 60 of the management device 12 outputs the text data obtained by transcribing the section voice of the target voice data transmitted from the worker apparatus 16 of each worker and the text data obtained by transcribing the section voice of the dummy voice data. Accept. The sentence generation unit 62 of the management device 12 synthesizes text data obtained by transcribing the section voice of the target voice data transmitted from the worker device 16 of each worker to synthesize a sentence in which the voice of the entire target voice data is converted into a text. Generate data.

  The text providing unit 64 of the management device 12 transmits the text data corresponding to each target voice data to the user terminal 14 that uploaded each target voice data. The user obtains the sentence data corresponding to the target voice data uploaded by himself / herself and advances the business. The evaluation unit 66 of the management device 12 determines the evaluation value of each worker based on the transcription result for the dummy voice data whose content is predetermined.

  The present invention has been described above based on the embodiments. This embodiment is merely an example, and it is understood by those skilled in the art that various modifications can be made to the combination of each constituent element and each processing process described in the embodiment, and such modifications are also within the scope of the present invention. This is where Hereinafter, modified examples will be shown.

  A first modification will be described. The dividing unit 54 of the management device 12 divides the target voice data to generate the first section voice data related to the first section and the second section voice data related to the second section immediately after the first section. , Part of the first section and part of the second section may overlap. In other words, the division unit 54 may provide a time region in which the first section and the second section have a margin.

  FIG. 7 shows an example of division of audio data. Here, the target voice data 80 is divided into four pieces of section voice data 82a, section voice data 82b, section voice data 82c, and section voice data 82d. The dividing unit 54 provides the overlapping period 84a and the overlapping period 84b in the section sound data 82a and the section sound data 82b. Further, the division unit 54 provides the overlapping period 84c and the overlapping period 84d in the section sound data 82b and the section sound data 82c. The division unit 54 also provides an overlap period 84e and an overlap period 84f in the section audio data 82c and the section audio data 82d. Here, each of the overlapping period 84a to the overlapping period 84f is 2.5 seconds.

  In the example of FIG. 7, the section audio data 82a is the section 15 seconds from the start point of the target audio data 80. The end position of the section may be determined by the method described in the example. In this section, 5 seconds before the end is the overlap period (the overlap period 84a + the overlap period 84b). The section voice data 82b is a section voice from the start point of the target voice data 80 to 10 seconds to 25 seconds. In this section, the overlap period is 5 seconds after the start and 5 seconds before the end. The section voice data 82a and the section voice data 82b are assigned to different workers, but the voices of the overlapping period 84a and the overlapping period 84b are transcribed by both different workers.

  In addition, the section voice data 82c is a section voice from the start point of the target voice data 80 to 20 seconds to 35 seconds. In this section, the overlap period is 5 seconds after the start and 5 seconds before the end. The section voice data 82d is the voice of the section from the start point of the target voice data 80 to 30 seconds to 45 seconds. In this section, the overlap period is 5 seconds after the start.

  The sentence generation unit 62 sets a predetermined number of text data of the first section voice data (for example, section voice data 82a) and text data of the second section voice data (for example, section voice data 82b) that are temporally consecutive in a predetermined number in the overlapping period. The text data of both are combined so that the characters (may be a predetermined number of morphemes) match.

  In addition, the text generation unit 62 adopts the text of the section voice data farther from the end as the text data in the overlapping period. In other words, the text data (for example, a predetermined number of characters or morphemes) corresponding to the end of each section voice data is not reflected in the sentence data of the synthesized word. For example, the sentence generation unit 62 adopts the text data based on the section voice data 82a for the overlap period 84a, while adopting the text data based on the section voice data 82b for the overlap period 84b. Similarly, the sentence generator 62 adopts the text data based on the section voice data 82b for the overlap period 84c, while adopting the text data based on the section voice data 82c for the overlap period 84d.

  The present inventor has recognized that, while implementing PoC (Proof of Concept) relating to transcription, the accuracy of text conversion decreases at the start and end of the section voice data. Therefore, in this modified example, an overlapping period is provided for the first section voice data and the second section voice data that are temporally continuous, and the text data of the first section voice data and the text data of the second section voice data are converted into text. The accuracy of the text data can be increased by reflecting the portion considered to have high accuracy in the text data.

  A second modification will be described. In the above-described embodiment, the person transcribes the target voice and the dummy voice when hearing the target voice, but as a modification, the computer (at least some of the worker devices 16) may automatically execute the transcription process. In this case, the distribution unit 58 calls the transcription requesting API that the worker device 16 publishes on the network, and transmits one or more section voice data (for example, the distribution data in FIG. 5) to the worker device 16. May be. The work result reception unit 60 may receive the text data of the transcription result as a return value of the transcription request API of the worker device 16.

  A third modified example related to the second modified example will be described. Transcription may be performed by both computer and human transcription. Specifically, the distribution unit 58 of the management device 12 first transmits one or more section voice data (for example, the distribution data of FIG. 5) to the first worker device that automatically executes the transcription process, and the work is performed. The result receiving unit 60 may obtain the result of the transcription process from the first worker device. Next, the delivery unit 58 transmits the result of the transcription process by the first worker device to the second worker device that performs the transcription manually, and the work result acceptance unit 60 causes the manual transcription ( Here, the result of inspection / editing) may be acquired from the second worker device. With this configuration, a person plays a role of inspecting and editing the result of the transcription by the computer, so that it is possible to improve the accuracy of the transcription while suppressing the labor cost.

  A fourth modification will be described. Although not mentioned in the above embodiment, the assigning unit 56 may assign the transcription of the section voice data in preference to the worker having a higher evaluation value stored in the evaluation storage unit 44. Further, the distribution unit 58 may distribute the section audio data by giving priority to a worker having a higher evaluation value stored in the evaluation storage unit 44. In other words, the section voice data may be assigned or distributed to a worker having a relatively high evaluation value stored in the evaluation storage unit 44 in preference to a worker having a relatively low evaluation value. .. This makes it easier to increase the accuracy of transcription.

  A fifth modification will be described. Although not mentioned in the above embodiment, the evaluation of the worker by the evaluation unit 66 may be executed before the generation of the sentence data by the sentence generation unit 62. When the evaluation value of a certain worker is less than a predetermined threshold value, the sentence generator 62 generates sentence data using the transcription result (text data) by the worker (hereinafter referred to as “low evaluator”). May be canceled. In this case, the allocation unit 56 may reallocate the section voice data allocated to the low-rated person to another worker (a worker whose evaluation value is the threshold value or more). The distribution unit 58 may provide the section voice data provided to the low-rated person to the other worker to request the transcription. As a result, it is possible to avoid providing the user with a transcription result with low accuracy and further improve the accuracy of transcription.

  A sixth modification will be described. Although not mentioned in the above embodiment, it is desirable that the allocation unit 56 provide different dummy voice data (at least different as segment voice data) for at least a predetermined period in the distribution data for the same worker. For example, the assigning unit 56 may store the identification information of the assigned dummy voice data (the section voice data) for each worker, and change the dummy voice data (the section voice data) each time the request is made. This makes it difficult for the operator to distinguish the dummy voice and enhances the confidentiality of the content of the target voice.

  A seventh modification will be described. The management device 12 of the above-mentioned embodiment provides the web page for each worker to the worker device 16, causes each worker to reproduce the section voice data on the web page for each worker, and performs each work. I asked the person to input the transcription result of the section voice data. In a modification, the distribution unit 58 of the management device 12 may transmit the encrypted data obtained by encrypting the section voice data of the target voice and the section voice encrypted data of the dummy voice to the worker apparatus 16. The work result acceptance unit 60 of the management device 12 may accept from the worker device 16 encrypted data obtained by encrypting the text data of the transcription result of each worker.

  An eighth modified example will be described. The user requests how many dummy voice section voice data are inserted into the target voice section voice data, or the ratio between the target voice section voice data and the dummy voice section voice data to be assigned to the worker. It may be determined by the confidentiality level of the target voice. That is, the allocation rule may be set such that the greater the required confidentiality strength, the greater the number of dummy voice section voice data inserted into the distribution data. Alternatively, the allocation rule may be set such that the ratio of the dummy voice section voice data to the target voice section voice data in the distribution data is high. Note that the greater the number of insertions or the ratio, the higher the confidentiality, and thus the selling price of the transcription service may be set higher.

  A ninth modification will be described. The configuration of the transcription system 10 described in the above embodiments is an example, and it goes without saying that the physical configuration (the number of housings, etc.) is not limited. For example, the function of dividing the original voice data and providing the section voice data to the worker device 16 and the function of collecting the transcription result by the worker, generating the sentence data, and providing it to the user are different from each other. It may be realized by the device.

  Any combination of the above-described examples and modifications is also useful as an embodiment of the present invention. The new embodiment resulting from the combination has the effects of the combined examples and modifications. It is also understood by those skilled in the art that the function that each constituent element described in the claims should fulfill is realized by the individual constituent elements shown in the embodiment and the modified examples or by their cooperation.

  10 character transcription system, 12 management device, 14 user terminal, 16 worker device, 52 conversion unit, 54 division unit, 58 distribution unit, 60 work result reception unit, 62 sentence generation unit, 64 sentence provision unit, 66 evaluation unit.

Claims (1)

A first storage unit for storing target voice data in which a voice of a transcription target is recorded;
A second storage unit for storing the dummy voice data in which the dummy voice is recorded;
A dividing unit that divides the target voice data into a plurality of section voice data relating to a plurality of sections,
A providing unit for providing an external device with a set of at least one of the plurality of section voice data and the dummy voice data;
A receiving unit for receiving text data transcribed based on at least one of the plurality of section voice data, and text data transcribed based on the dummy voice data;
A generation unit that generates, from the text data received by the reception unit, text data that has been transcribed based on at least one of the plurality of section voice data, and that generates text data that has transcribed the target voice. ,
An information processing system comprising:

Images