Automatic report generation system

Abstract

Provided is an automatic report generation system capable of automatically generating reports for various purposes. A report automatic generation system is provided with a speech recognition unit (3) that recognizes that a pre-stored start term has been input, a timing unit (7) that determines a start time on the basis of the time at which the speech recognition unit (3) recognizes that the start term has been input, and a term dictionary (9) that determines a start time on the basis of the time at which the speech recognition unit (3) recognizes that a pre-stored end term has been input. A timing unit (7) determines an end time on the basis of the time at which the speech recognition unit (3) recognizes that an end term has been input, and the speech recognition unit (3) stores speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores an utterance keyword in terms included in the recorded speech, and stores the utterance keyword in the terms included in the recorded speech. Wherein the utterance keyword is an utterance included in the utterance dictionary (9), and the automatic report generation system generates a file that includes information relating to the start time, the end time, and the utterance keyword and that is associated with the recorded voice.

Description

This application is a divisional application of the invention patent application with application number 202080006101.3. The international application date of the original application is February 10, 2020, the priority date is May 14, 2019, and the date of entering the Chinese national stage is May 10, 2021. The name of the invention of the original application is “Automatic Report Generation System”. Technical Field

[0001] The present invention relates to an automatic report creation system. Background Art

[0002] Japanese Patent Publication No. 2015-75792 describes a report generation system using a mobile terminal held by a worker who performs maintenance and inspection work on an item. The report generation system summarizes the maintenance and inspection results of each work object item of the item by information obtained from voice input by the worker. This report generation system automatically collects information inputted by voice for each work target item of an article, so its use is limited to the maintenance and inspection work of the article. In addition, this system cannot make a report with arbitrary contents. [Prior art literature] [Patent Document]

[0003] Patent Document 1: Japanese Patent Publication No. 2015-75792 Summary of the invention [Technical problem to be solved by the invention]

[0004] An object of the invention described in this specification is to provide an automatic report generation system capable of automatically generating reports for various purposes. [Technical solutions for solving technical problems]

[0005] The above technical problem is based on the following insight, for example: by using a term dictionary to extract the required terms from the voice information included in the conversation, a report can be automatically generated. In addition, the preferred technical solution is based on the following insight: having a location information confirmation unit such as GPS, by confirming the location information, determining the location and time of the meeting, etc., it is possible to more automatically grasp the location and start time of the meeting, etc.

[0006] One of the inventions described in this specification relates to an automatic report generation system. The system 1 includes a speech recognition unit 3, a position information confirmation unit 5, a timing unit 7, and a term dictionary 9. The position information confirmation unit confirms the position of the automatic report generation system and confirms that the automatic report generation system has stayed within a specified geographical range for a certain period of time. Then, the position information confirmation unit obtains a meeting place in association with the position of the automatic report generation system. The obtained meeting place can be appropriately stored in the storage unit of the system.

[0007] In this case, the timing unit calculates the start time based on the time when the automatic report generation system starts to stay within the specified geographical range, the time when voice is input to the voice recognition unit after the automatic report generation system starts to stay within the specified geographical range, or the time when the voice recognition unit recognizes that a pre-stored start term has been input.

[0008] In addition, the timing unit calculates the end time based on the time when the automatic report generation system leaves the specified geographical range, the time when the voice recognition unit last inputs the voice before the time when the automatic report generation system leaves the specified geographical range, and the time when the voice recognition unit recognizes that a pre-stored end phrase is input.

[0009] The speech recognition unit stores the speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores utterance keywords from the terms included in the recorded speech, wherein the utterance keywords are terms included in the term dictionary. In addition, the report automatic generation system generates a file, which contains information related to the meeting place, start time, end time and speech keywords, and is associated with the recorded voice. In this way, the system automatically generates a report.

[0010] Preferred examples of the above system are as follows. That is, the position information confirmation unit displays the candidate meeting places based on the position information of the automatic report generation system, and stores the place input as the meeting place as the meeting place.

[0011] Preferred examples of the above system are as follows. Linked with schedule management tools or conference room reservation systems, Receive meeting information stored in the calendar management tool or conference room reservation system, When the position information confirmed by the position information confirming unit is associated with the information included in the meeting information, one or both of the meeting place information and the attendee information included in the meeting information are read out.

[0012] As another example of the automatic report generation system 1, The automatic report generation system 1 includes a speech recognition unit 3, a timer unit 7, and a term dictionary 9. The speech recognition unit 3 recognizes that a pre-stored start phrase is input. The timing unit 7 obtains the start time based on the time when the speech recognition unit 3 recognizes that the start word is input. The speech recognition unit 3 recognizes that a pre-stored ending phrase is input. The timing unit 7 obtains the end time based on the time when the voice recognition unit 3 recognizes that the end word is input. The speech recognition unit 3 stores the speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores speech keywords from the terms contained in the recorded speech, wherein the speech keywords are terms contained in the term dictionary 9. The automatic report generation system generates a file that contains information related to the start time, end time and utterance keywords, and is associated with the recorded speech. [Effects of the invention]

[0013] The automatic report generation system can automatically generate reports for various purposes. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a block diagram for explaining the automatic report generation system. Figure 2 It is a block diagram showing the basic structure of a computer. Figure 3 This is a conceptual diagram showing an example of a report format. Figure 4 This is a conceptual diagram showing a state where a plurality of company names are read out from a storage unit and the read company names are displayed as candidate locations. Figure 5 This is a conceptual diagram showing an example of a display screen of a schedule management tool. DETAILED DESCRIPTION

[0015] Hereinafter, a mode for implementing the present invention will be described using the drawings. The present invention is not limited to the mode described below, and also includes modes obtained by appropriately modifying the following mode within a range obvious to those skilled in the art.

[0016] Figure 1 is a block diagram used to illustrate the automatic report generation system. Figure 1 As shown, the system 1 includes a voice recognition unit 3, a position information confirmation unit 5, a timing unit 7, and a term dictionary 9. Preferably, the system is a system that is automatically processed by a computer. In addition, the system can also be a system that can process the input from the user as one of the information when there is input.

[0017] Figure 2 It is a block diagram showing the basic structure of a computer. As shown in the figure, the computer has an input unit 11, an output unit 13, a control unit 15, a computing unit 17 and a storage unit 19, and each structural element is connected by a bus 21 or the like so as to be able to exchange information. For example, the storage unit can store a control program or various information. When the specified information is input from the input unit, the control unit reads out the control program stored in the storage unit. Then, the control unit appropriately reads out the information stored in the storage unit and transmits it to the computing unit. In addition, the control unit appropriately transmits the input information to the computing unit. The computing unit uses the various information received to perform computing processing and stores the computing results in the storage unit. The control unit reads out the computing results stored in the storage unit and outputs them from the output unit. Various processing is performed in this way. The various processing is performed by each mechanism.

[0018] When the automatic report generation system 1 (the user carrying the automatic report generation system 1) moves, the location information confirmation unit 5 such as GPS stores the location of the automatic report generation system 1 in the storage unit at any time. The timing unit 7 is a mechanism that can measure time in addition to being able to find and store the moment. For example, when the automatic report generation system 1 stays within a certain range (for example, within 10m) for a specified time or more (for example, more than 1 minute), the place where it stays is stored as a meeting place. Therefore, information related to a certain time and a certain range is stored in the storage unit. Then, the system 1 receives the location information of the system 1 from the location information confirmation unit 5, and the system 1 reads out the information related to a certain time and a certain range from the storage unit. Then, the time information from the timing unit 7, the location information of the system 1 of the location information confirmation unit 5, and the information related to a certain time and a certain range read out are used to determine whether the system 1 exists within a certain time and a certain range. Then, when the system 1 determines that the system 1 exists within a certain range for more than a certain time, the location where the system 1 exists is stored in the storage unit as a meeting place. Thus, the location information confirmation unit 5 confirms the location of the automatic report generation system 1 and confirms that the automatic report generation system 1 has stayed in a predetermined geographical area for a certain period of time. Then, the location information confirmation unit 5 stores the interview point in association with the location of the automatic report generation system 1 .

[0019] In this case, the timing unit 7 obtains the start time based on the following time: (1) Report the time when the automatic generation system 1 starts to stay within the specified geographical area, (2) the time at which speech is input to the speech recognition unit 3 after the automatic report generation system 1 begins to stay within the prescribed geographical area, or (3) The speech recognition unit 3 recognizes the time when the pre-stored start word is input. Any of these times may be preset in the system.

[0020] The above (1) is explained. As described above, the timing unit 7 of the system 1 stores the time. In addition, the position information confirmation unit 5 stores the position of the system 1. Therefore, when the system 1 determines that the system 1 exists within a certain range for more than a certain time, the time when the system 1 begins to exist within the certain range can be read from the storage unit in the timing unit 7 as the start time.

[0021] The above (2) is explained. The voice recognition unit 3 has, for example, a sound collection function. In addition, the voice recognition unit 3 appropriately records the voice input into the system 1 and stores it in the storage unit. At this time, the voice recognition unit 3 can also analyze the various terms contained in the voice. The system 1 can also obtain information related to the time from the timing unit 7, and store the various terms together with the time when they are input into the system 1 in the storage unit. As described above, when the system 1 determines that the report automatic generation system 1 stays within a specified geographical range for more than a certain time, the voice recognition unit 3 (automatically) starts recording, or the voice recognition unit 3 (automatically) becomes turned on. When the volume of the voice (especially the voice in a specific frequency range) stored in the storage unit is above a certain volume, the system 1 can determine that a meeting has started, and store the time when the volume of the voice is above a certain volume as the start time. In this case, for example, information related to a specific frequency range and information related to a certain volume are stored in advance in the storage unit, and are read out appropriately, and the voice information stored in the storage unit is compared with the information related to the specific frequency range and the information related to the certain volume read out from the storage unit, and it is determined that the volume of the voice in the specific frequency range in the voice information stored in the storage unit is above the certain volume. In this way, the system 1 can store the time when the voice is input to the voice recognition unit 3 after the automatic report generation system 1 starts to stay in the specified geographical range as the start time in the storage unit.

[0022] Finally, the above (3) is explained. The description of the same parts as the above (2) in this case is omitted. The term dictionary 9 stores one or more starting terms associated with the start of a conversation, for example. The starting terms are, for example, terms such as "Sorry for the wait", "Please", "Take care", "Thank you", "Begin", and "Please allow me to begin". The speech recognition unit 3 stores the speech in the storage unit. At this time, for example, information related to the time when each term is input into the system 1 is received from the timing unit 7, and the information related to the time is also stored in the storage unit together with the recognized term. The system 1 reads the term stored as the starting term from the term dictionary 9. Then, the system 1 determines whether the term (conversational term) stored in the storage unit is consistent with the read starting term. Then, when the term (conversational term) stored in the storage unit is consistent with one of the starting terms, the time when the conversational term is input into the system 1 is read from the storage unit. At this time, the timing unit 7 of the system 1 can also calculate the time. Then, the read time is stored in the storage unit as the starting time. In this way, the system 1 can obtain the start time based on the time when the speech recognition unit 3 recognizes that the pre-stored start word is input.

[0023] In another example of (3), for example, the user arrives at the meeting place, turns on the power of system 1 just before the meeting begins, and speaks a start word (e.g., "start"), so that system 1 recognizes that the meeting has begun, receives the time when the start word is input to system 1 from the timing unit, and stores the time as the start time in the storage unit. In this case, the location information confirmation unit 5 may also store the location where system 1 is located as the meeting place in the storage unit.

[0024] In the example (3) above, when a term for a meeting or the like is generated, the time at which the term is generated can be used as the start time. Therefore, for example, it is possible to prevent the system from erroneously judging that a meeting has started because the vehicle stops for a predetermined time or longer due to a traffic jam while moving to the meeting place. In addition, it is possible to prevent the system from erroneously judging that a meeting has started because the person has to wait in a waiting room for a predetermined time or longer to transfer to a train or bus.

[0025] In addition, the timing unit 7 calculates the end time based on (1) the time when the report automatic generation system 1 leaves the specified geographical range, (2) the time when the voice recognition unit 3 last inputs the voice before the report automatic generation system 1 leaves the specified geographical range, and (3) the time when the voice recognition unit recognizes that a pre-stored end phrase is input. Any one of these times may be pre-set in the system.

[0026] The above (1) is explained. As described above, the timing unit 7 of the system 1 stores the time. In addition, the location information confirmation unit 5 stores the location of the system 1. Therefore, when it is determined that the system 1 has left a certain geographical range, the system 1 determines that the meeting is over, and reads the time of leaving the certain geographical range from the storage unit of the timing unit 7 as the end time.

[0027] The above (2) is explained. For example, the system 1 obtains information related to the time from the timing unit 7, and stores each term analyzed by the voice recognition unit together with the time when the voice is input to the system 1 (therefore, the voice recognition unit 3) in the storage unit. That is, the system 1 stores each term and the time when the term is spoken in the storage unit. The position information confirmation unit 5 stores the position of the system 1, and the system 1 analyzes the above position information to recognize that the system 1 has left the specified geographical range. Since the system 1 stores each term and the time when the term is spoken, the system 1 uses the time when the last term is spoken before leaving the specified geographical range as the end time. In this way, the system 1 can store the time when the voice is last input into the voice recognition unit 3 before the report automatic generation system 1 leaves the specified geographical range as the end time in the storage unit.

[0028] Finally, the above (3) is explained. The description of the parts that are the same as the above (2) is omitted in this case. The term dictionary 9 stores, for example, one or more ending terms associated with the start of a conversation. Ending terms are, for example, terms such as "thank you", "end", "thank you for your hard work", and "thank you for your personal visit". The speech recognition unit 3 stores the speech in the storage unit. The system 1 reads out the term stored as the ending term from the term dictionary 9. Then, the system 1 determines whether the term (conversational term) stored in the storage unit is consistent with the read-out ending term. Then, when the term (conversational term) stored in the storage unit is consistent with any one of the ending terms, the time when the conversational term is input into the system 1 is read out from the storage unit. Then, the read-out time is stored in the storage unit as the end time. In this way, the system 1 can calculate and store the end time based on the time when the speech recognition unit recognizes that the pre-stored end term is input.

[0029] In another example of (3), for example, when the user leaves the meeting place, the system 1 is powered on and an ending phrase (e.g., "recording is over") is spoken, so that the system 1 recognizes that the meeting is over and receives the time when the ending phrase is input to the system 1 from the timing unit and stores the time as the ending time in the storage unit. At this time, the system 1 may be automatically powered off.

[0030] The speech recognition unit 3 stores the speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores utterance keywords from the terms included in the recorded speech, wherein the utterance keywords are terms included in the term dictionary 9 .

[0031] A plurality of discourse keywords are stored in the term dictionary 9. The system 1 reads the discourse keywords from the term dictionary 9 and the conversation terms stored in the storage unit, and determines whether the conversation terms match any of the discourse keywords. Then, if the conversation terms match any of the discourse keywords, the discourse keyword is stored in the storage unit.

[0032] Next, the system generates data for reporting. The automatic report generation system 1 reads out information related to the meeting place, start time, end time, and speech keywords stored in the storage unit. In addition, the system reads out data related to the recorded voice of the conversation stored in the storage unit. Then, the system 1 uses the read information related to the meeting place, start time, end time, and speech keywords and the data related to the recorded voice to generate a file, which includes the meeting place, start time, end time, and speech keywords and is associated with the recorded voice. In this way, the system automatically generates a report.

[0033] Figure 3 : is a conceptual diagram showing an example of the format of a report. Display data is generated in such a way that the information related to the interview location stored in the storage unit and read out is displayed at the location-related position 31. For example, the data related to the interview location read out can be placed next to the location-related label. When such data is prepared, the information related to the interview location is displayed at the location-related position 31 on the screen that displays the report to the user. The date and time can also be displayed at the date and time-related position 33 using the information related to the start time and the end time stored in the storage unit. The date and time can be displayed by displaying the start time and the end time, by displaying only the start time, or by calculating the difference between the end time and the start time and displaying the interview time. At the content position 37, one or more speech keywords stored in the storage unit are displayed. The speech keywords are, for example, terms that need to be explained for a specific drug. Thus, it is easy to check whether all the terms that need to be explained, such as the informed consent, have been said. Furthermore, when not all utterance keywords are displayed at the position 37 of the content, the content of the conversation may be checked using a voice file reading icon 39 described later.

[0034] In addition, the speech keywords that need to be explained can be pre-stored in the storage unit. Then, the system 1 can determine whether the speech keywords included in the conversation are consistent with all the necessary keywords. On this basis, the system 1 can display information to call attention in the report when the speech keywords included in the conversation do not include all the necessary keywords. The terms and marks used to call attention can be pre-stored in the storage unit, and when the speech keywords included in the conversation do not include all the necessary keywords, the terms and marks used to call attention stored in the storage unit are processed so that they are displayed in the report.

[0035] Furthermore, when the participants have installed the same application, the parties participating in the same meeting can be identified by using the location information confirmation unit 5. In this case, the storage unit stores information related to the user of the terminal that has installed each system, reads the stored information related to the user, and controls to display the information at the location 35 where the participants are displayed.

[0036] The position information confirmation unit 5 can display the candidate meeting places based on the position information of the automatic report generation system 1 and store the place input as the meeting place as the meeting place.

[0037] For example, there may be multiple companies settled in a building. In this case, the storage unit stores information related to the companies, such as multiple company names, in association with the location of the building. The location information confirmation unit 5 confirms the location of the system 1 and reads out information related to the multiple companies settled in the building. Then, the system 1 reads out information related to the multiple companies from the storage unit, for example, and displays the multiple company names at the location 31 related to the reported location. The multiple company names can be displayed in a pull-down manner. Figure 4 This is a conceptual diagram showing a state where a plurality of company names are read out from a storage unit and the read company names (AAAAAA, BBBBBB, and CCCCCC) are displayed as candidate locations (meeting locations). Furthermore, when a user selects a company name, the system 1 receives input related to the company name. Then, the system 1 stores the input company name in the storage unit as a meeting location. In this way, candidate meeting locations can be displayed, and a meeting location can be easily determined without manual input.

[0038] The configuration may be such that: it is linked with a schedule management tool or a conference room reservation system, receives meeting information stored in the schedule management tool or the conference room reservation system, and when the location information confirmed by the location information confirmation unit 5 is associated with the information contained in the meeting information, reads out either or both of the meeting location information and attendee information contained in the meeting information.

[0039] Figure 54 is a conceptual diagram showing an example of a display screen of a schedule management tool. In this example, when a date on the calendar is specified by an arrow icon 41, a schedule scheduled for that date is displayed on a schedule display screen 43. The system stores the schedule for each date in a storage unit in association with the date. In addition to the topic and time of the meeting, the schedule also includes meeting information such as the meeting place, scheduled attendees, and a meeting summary. Furthermore, the system 1 that cooperates with the schedule management tool reads out information related to the meeting place from the storage unit. Then, when the location information confirmed by the location information confirmation unit 5 is consistent with the meeting place, the system 1 can determine that the meeting place is correct. In addition, the system 1 can also read out information about the scheduled attendees stored for the meeting from the storage unit. Furthermore, the read information related to the scheduled attendees can be recorded in the attendee column in the report described above.

[0040] A different form of the automatic report generation system from the above is a system for automatically generating reports having a voice recognition unit 3, a timing unit 7, and a term dictionary 9. Preferably, the system is a system that automatically processes by a computer. In addition, the system may be a system that can process the input from the user as one of the information when there is an input from the user. The system also basically performs the same processing as the system described above.

[0041] For example, a user speaks a starting word to the system. Then, the speech is input into the system. Then, the speech recognition unit 3 recognizes that a pre-stored starting word has been input. For example, the word dictionary 9 stores one or more starting words. The speech recognition unit 3 stores the speech in the storage unit. The system 1 reads out the word stored as the starting word from the word dictionary 9. Then, the system 1 determines whether the word (conversational word) stored in the storage unit is consistent with the read starting word. Then, when the word (conversational word) stored in the storage unit is consistent with any one of the starting words, the time when the conversational word is input into the system 1 is read out from the storage unit. At this time, the timing unit 7 calculates the start time based on the time when the speech recognition unit 3 recognizes that the starting word has been input.

[0042] For example, the user speaks an ending phrase to the system. Then, the voice is input into the system. The voice recognition unit 3 recognizes that a pre-stored ending phrase has been input. For example, the term dictionary 9 stores one or more ending phrases. The voice recognition unit 3 stores the voice in the storage unit. The system 1 reads out the term stored as the ending phrase from the term dictionary 9. Then, the system 1 determines whether the term (conversational term) stored in the storage unit is consistent with the read ending phrase. Then, when the term (conversational term) stored in the storage unit is consistent with any one of the ending phrases, the time when the conversational term is input into the system 1 is read out from the storage unit. The timing unit 7 can calculate the ending time based on the time when the voice recognition unit 3 recognizes that the ending phrase is input.

[0043] The speech recognition unit 3 stores the speech from the start time to the end time as the recorded speech. Then, the system 1 generates a recorded speech file. The recorded speech is analyzed, and speech keywords in the terms contained in the recorded speech are extracted and stored, wherein the speech keywords are terms contained in the term dictionary 9. Then, a file is generated that contains information related to the start time, the end time and the speech keywords and is associated with the recorded speech.

[0044] The automatic report generation system 1 described above can be said to be a configuration in which the automatic report generation system of the above-described embodiment further includes a position information confirmation unit 5 . [Industrial Applicability]

[0045] The present invention can be used in information communication related fields. [Explanation of Reference Numerals]

[0046] 1: Automatic report generation system; 3: Voice recognition unit; 5: Position information confirmation unit; 7: Timing unit; 9: Dictionary of terms.

Claims

1. A report automatic generation system, characterized in that: It has a voice recognition unit, a position information confirmation unit, a timing unit and a term dictionary, wherein: The location information confirmation unit confirms the location of the automatic report generation system, and when the automatic report generation system stays within a specified geographical range for a certain period of time or longer, the location information confirmation unit obtains a meeting place associated with the location of the automatic report generation system. The timing unit calculates the start time based on the time when the automatic report generation system starts to stay within the specified geographical range, the time when speech is input to the speech recognition unit after the automatic report generation system starts to stay within the specified geographical range, or the time when the speech recognition unit recognizes that a pre-stored start word is input. The timing unit calculates the end time based on the time when the automatic report generation system leaves the specified geographical range, the time when the voice recognition unit last inputs the voice before the automatic report generation system leaves the specified geographical range, and the time when the voice recognition unit recognizes that a pre-stored end phrase is input. The speech recognition unit stores the speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores speech keywords from the terms contained in the recorded speech, wherein the speech keywords are terms contained in the term dictionary, The automatic report generation system generates a file, which contains information related to the meeting place, the start time, the end time and the speech keywords, and is associated with the recorded voice.

2. The automatic report generation system according to claim 1, characterized in that: The position information confirmation unit displays candidate meeting places based on the position information of the automatic report generation system, and stores the place input as the meeting place as the meeting place.

3. The automatic report generation system according to claim 1, characterized in that: Linked with schedule management tools or conference room reservation systems, receiving the meeting information stored in the schedule management tool or the conference room reservation system, When the position information confirmed by the position information confirming unit is associated with information included in the meeting information, one or both of the meeting place information and the attendee information included in the meeting information are read out.

4. A report automatic generation system, characterized in that: It has a voice recognition unit, a timing unit and a term dictionary. The speech recognition unit recognizes that a pre-stored start phrase is input, The timing unit calculates the start time based on the time when the speech recognition unit recognizes that the start word is input. The speech recognition unit recognizes that a pre-stored ending phrase is input, The timing unit calculates the end time based on the time when the voice recognition unit recognizes that the end word is input, The speech recognition unit stores the speech from the start time to the end time as recorded speech, analyzes the recorded speech, extracts and stores speech keywords from the terms contained in the recorded speech, wherein the speech keywords are terms contained in the term dictionary, The automatic report generation system generates a file, which contains information related to the start time, the end time and the utterance keyword, and is associated with the recorded voice.

Images