JP2012043046A - Conference minute creation system and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a conference minute creation system capable of easily attaching item information to conference minutes; and to provide an art related thereto.SOLUTION: This conference minute creation system extracts a static image PS1 related to a display scene of conference materials from image data of the conference, and generates a conference minute file PR based on the static image PS1. This conference minute creation system generates a bookmark item BM1 (for example, "a report of decrease in productivity") related to the static image PS1 in the conference minute file PR. Alternatively, an item name of the bookmark item BM1 is preferably determined based on a headline item HD1 extracted from the static image PS1 and character data CH1 generated by recognizing voice data in a specific period corresponding to the static image PS1.

Description

  The present invention relates to a minutes creation system for creating minutes of a meeting and a technology related thereto.

  There is a system for generating meeting minutes (see Patent Document 1). For example, Patent Document 1 describes a system in which a minutes maker inputs the progress of a meeting as text information and the text information is held in association with the date and time.

JP 2006-268800 A

  By the way, when grasping the whole picture of the meeting, or grasping the contents of each person's remarks while referring to the relationship with the progress of the meeting, the item level (heading level) is used as information about the minutes. It is preferable to use this information. That is, the item level information (item information) is desirably added to the minutes.

  In addition, when such item information is generated by an operation input by the operator, the labor of the operator is great. The labor is preferably reduced.

  Therefore, an object of the present invention is to provide a minutes creation system that can easily add item information to minutes, and a technology related thereto.

  In order to solve the above-mentioned problem, the invention of claim 1 is a minutes creation system for creating a minutes of a meeting, which extracts a still image related to a display scene of a meeting material from video data of the meeting. Means for generating a minutes file based on the still image, and bookmark generation means for generating a bookmark item related to the still image in the minutes file.

  The invention of claim 2 is the minutes creation system according to the invention of claim 1, wherein the audio data of the specific period corresponding to the still image is extracted from the audio data of the meeting, and the audio data of the specific period Voice data analyzing means for generating character data by voice recognition, wherein the bookmark generating means extracts a headline item from the still image, and the headline item extraction means extracts the headline item extraction means. And bookmark item generating means for determining an item name of the bookmark item based on the heading item and the character data generated based on the voice data.

  A third aspect of the present invention is the minutes creation system according to the second aspect of the invention, wherein the bookmark item generating means analyzes a sentence relating to the character data and extracts a word related to the heading item. The bookmark item is generated by changing the item name of the heading item using the extracted word.

  A fourth aspect of the present invention is the minutes creation system according to any one of the first to third aspects, wherein the bookmark generation means converts the link data with the still image as a link destination to the bookmark item. It is characterized by giving to.

  A fifth aspect of the present invention is the minutes creation system according to any one of the first to fourth aspects, wherein the bookmark generation means arranges the bookmark item in a bookmark display field of the minutes file. It is characterized by that.

  A sixth aspect of the present invention is the minutes creation system according to the second or third aspect of the invention, wherein the file generation unit corresponds to a character string of the character data in the minutes file corresponding to the still image. It is characterized by arranging them.

  A seventh aspect of the present invention is the minutes creation system according to any one of the second to sixth aspects, wherein the file generation means places the voice data of the specific period at a position corresponding to the bookmark item. It is characterized by linking.

  The invention of claim 8 is the minutes creation system according to the invention of claim 1, wherein the audio data of the meeting is analyzed to extract audio data of a specific period corresponding to the still image, and the specific period Voice data analyzing means for generating voice data by voice recognition of the voice data, wherein the file generating means arranges the character data in correspondence to the still image in the minutes file. And

  According to the ninth aspect of the present invention, a) a procedure for extracting a still image related to a display scene of a conference material from video data of a conference; b) a procedure for generating bookmark items associated with the still image; c ) A program for executing a procedure for generating a minutes file having the still image and the bookmark item.

  The invention of claim 10 includes: a) a procedure for extracting a still image related to a display scene of a conference material from video data of a conference; and b) a specific period corresponding to the still image from audio data of the conference. A procedure for extracting voice data and generating voice data by voice recognition of the voice data of the specific period; c) a procedure for extracting a heading item from the still image; and d) the heading item and the character data. And e) a procedure for generating a bookmark item having the item name determined based on e) and a procedure for generating a minutes file having the still image and the bookmark item associated with the still image. It is a program.

  According to the first to tenth aspects of the present invention, a still image related to the display scene of the conference material is extracted from the video data of the conference, and a bookmark item related to the still image is generated in the minutes file. Therefore, compared with the case where the items in the minutes are manually added, it is possible to reduce the time and easily add item information to the minutes.

  In particular, according to the second aspect of the present invention, the item name of the bookmark item is determined based on the heading item extracted from the scene image and the character data generated based on the partial voice data. Therefore, the item name of the bookmark item includes more information than the original heading item. Therefore, the user can more easily grasp the contents of the minutes file by using the bookmark item.

It is a conceptual diagram which shows the outline | summary of a conference system. It is a block diagram which shows schematic structure of a meeting management apparatus (MFP). It is a flowchart which shows the creation operation of the minutes file. 4 is a flowchart showing in detail a part of the operation of FIG. 3. It is a figure which shows the minutes file. It is a figure which shows the minutes file. It is a figure which shows video data, audio | voice data, etc. It is a figure which shows a scene image. It is a figure which shows another scene image. It is a figure which shows the character data extracted from the partial audio | voice data. It is a conceptual diagram which shows a mode that a heading item is extracted from a scene image. It is a conceptual diagram which shows the analysis operation | movement of character data. It is a conceptual diagram which shows the determination operation | movement of a bookmark item.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<1. System configuration>
FIG. 1 is a conceptual diagram showing an outline of the conference system 100. The conference system 100 has a function of creating a minutes, and is also expressed as a minutes creation system for creating the minutes of a meeting.

  The conference system 100 includes a conference management device 10, a display device 30, a computer 50, a video camera 71, and a microphone 72.

  The conference management device 10, the display device 30, the computer 50, the video camera 71, and the microphone 72 are connected to each other via a network NW, and can perform network communication. Here, the network NW is configured by a LAN, a WAN, the Internet, or the like. A connection form of each device to the network NW may be a wired connection or a wireless connection.

  The microphone 72 collects audio in the conference, generates an audio signal, and transmits the audio signal to the conference management device 10. The conference management apparatus 10 generates audio data AD (see FIG. 7) based on the audio signal recorded by the microphone 72 and stores it in the storage unit 5.

  The video camera 71 shoots a person video or the like in the conference, generates a video signal, and transmits the video signal to the conference management apparatus 10.

  The conference management apparatus 10 generates video data (moving image data) PM (see FIG. 7) based on the video signal captured by the video camera 71 and stores it in the storage unit 5. Furthermore, the conference management apparatus 10 also receives a signal from the computer 50, for example, display data related to the conference material DM displayed on the computer 50. Then, the conference management apparatus 10 generates video data PM based on data from the computer 50 and the like. Specifically, the video data PM is composed of images taken by the video camera 71 at a certain point in time, and the video data PM is composed based on data (such as display data) from the computer 50 at another point in time. Alternatively, the video data PM may be configured by combining an image captured by the video camera 71 and data from the computer 50. As described above, the video data PM is generated based on the image taken by the video camera 71 and the video from the computer 50. The video data PM is output from the conference management device 10 to the display device 30 and displayed on the display device 30 in real time.

  Here, the case where the video of the conference material DM is transmitted from the computer 50 is illustrated, but the present invention is not limited to this. For example, the conference management device 10 itself may store and manage material data related to the conference and output the data to the display device 30 during the conference.

  The display device 30 is a device that displays various information and the like transmitted from the conference management device 10. The display device 30 is configured as, for example, a large liquid crystal display device, and the display device 30 is connected to the conference management apparatus 10 via the network NW. However, the present invention is not limited to this, and the display device 30 may be a projector device (projection device) or the like. Further, the display device 30 may be indirectly connected to the conference management device via the computer 50. More specifically, the display device 30 may display and output a video signal from the conference management device 10 via a video signal output unit (specifically, an external output terminal) of the computer 50 or the like.

  The attendees of the meeting can visually recognize various information (material data and the like) related to the meeting by the display image by the display device 30.

  The conference management apparatus 10 can also create a conference minutes file PR (see FIGS. 5 and 6). Therefore, the conference management device 10 is also expressed as a minutes creation device. The minutes creation operation will be described in detail later.

  In this embodiment, the conference management apparatus 10 is configured as an MFP (Multi Function Peripheral).

  FIG. 2 is a block diagram illustrating a schematic configuration of the MFP 10.

  As shown in the functional block diagram of FIG. 2, the MFP 10 includes an image reading unit 2, a print output unit 3, a communication unit 4, a storage unit 5, an input / output unit 6, a controller 9, and the like. Various functions can be realized by operating them in an automated manner.

  The image reading unit 2 is a processing unit that optically reads a document placed at a predetermined position of the MFP 10 and generates image data (also referred to as a document image) of the document.

  The print output unit 3 is an output unit that prints out an image on various media such as paper based on data related to a print target.

  The communication unit 4 is a processing unit capable of performing facsimile communication via a public line or the like. Furthermore, the communication unit 4 can perform network communication via the communication network NW. In this network communication, various protocols such as TCP / IP (Transmission Control Protocol / Internet Protocol) and FTP (File Transfer Protocol) are used. By using the network communication, the MFP 10 can exchange various data with a desired destination.

  The storage unit 5 includes a storage device such as a hard disk drive (HDD). The storage unit 5 stores a document image (image data) generated by the image reading unit 2 or the like. The storage unit 5 also stores video data PM, audio data AD, and the like.

  The input / output unit 6 includes an operation input unit 6a that receives input to the MFP 10 and a display unit 6b that displays and outputs various types of information.

  The controller 9 is a control device that is built in the MFP 10 and controls the MFP 10 in an integrated manner. The controller 9 is configured as a computer system including a CPU and various semiconductor memories (such as a RAM and a ROM). The controller 9 implements various processing units by executing predetermined software programs (hereinafter also simply referred to as programs) stored in a ROM (for example, an EEPROM) in the CPU.

  Specifically, as shown in FIG. 2, the controller 9 includes a video data analysis unit 31, an audio data analysis unit 33, a minutes file generation unit 35, a bookmark generation unit 37, and the like.

  The video data analysis unit 31 is a processing unit that analyzes video data (moving image data) PM (see FIG. 7) of a meeting. The video data analysis unit 31 is a processing unit that extracts a plurality of scene images (still image data) PSn related to the display scene of the conference material from the conference video data PM, and is also expressed as a still image extraction unit.

  The audio data analysis unit 33 analyzes the audio data AD of the conference and extracts audio data SAn for a specific period corresponding to each scene image PSn, and also recognizes the audio data SAn for the specific period and recognizes the character data CHn. Is a processing unit for generating

  The minutes file generation unit 35 is a processing unit that generates a minutes file PR (see FIGS. 5 and 6) based on a plurality of scene images PSn.

  The bookmark generation unit 37 is a processing unit that generates a bookmark item BMn to which link data (hyperlink data) with the scene image PSn as a link destination is assigned. In addition, the bookmark generation unit 37 cooperates with the minutes file generation unit 35 to arrange a plurality of bookmark items BMn in the bookmark display field RB of the minutes file PR.

  The bookmark generation unit 37 includes a heading item extraction unit 38 and a bookmark item generation unit 39.

  The heading item extraction unit 38 is a processing unit that extracts the heading item HDn from the scene image PSn.

  The bookmark item generation unit 39 is a processing unit that generates an item name of the bookmark item BMn. The bookmark item generation unit 39 generates the item name of “bookmark item” BMn based on the heading item HDn extracted by the heading item extraction unit 38 and the character data CHn generated based on the audio data SAn of a specific period. To do. More specifically, the bookmark item generation unit 39 analyzes a sentence related to the character data CHn, extracts a word related to the heading item HDn, and changes the item name of the heading item HDn using the extracted word. Bookmark item BMn is generated.

<2. Operation>
Next, the operation of creating the minutes file PR by the system 100 will be described in detail.

  Here, at the beginning of the conference, a person image of the presenter (presenter) at the conference is taken, and then the conference material DM displayed on the display device 30 from the computer 50 via the conference management device 10 is used. Assume that the presenter explains. Further, the video data PM (see FIG. 7) acquired for the conference has a person video of the presenter at the beginning of the video data PM, and subsequently has video regarding a plurality of pages of the conference material DM. To do. Similarly, it is assumed that the voice data AD has the voice of the person of the presenter at the beginning thereof, and subsequently has the voice related to the explanation regarding a plurality of pages of the conference material DM. In the following, a case where a meeting minutes file PR is created based on the video data PM and the like after the meeting is completed will be exemplified.

  FIG. 3 is a flowchart showing the operation of creating the minutes file PR, and FIG. 4 is a flowchart showing in detail the operation of a part (step S5) of FIG. 5 and 6 are diagrams showing the minutes file PR, and FIG. 7 is a diagram showing the video data PM, the audio data AD, and the like. FIG. 8 is a diagram showing a scene image PS1, and FIG. 9 is a diagram showing another scene image PS2. FIG. 10 is a diagram showing character data CH1 extracted from the partial voice data SA1. FIG. 11 is a conceptual diagram showing how a headline item is extracted from the scene image PS1, and FIG. 12 is a conceptual diagram showing an analysis operation of character data CH1. FIG. 13 is a conceptual diagram showing an operation for determining the bookmark item BM1.

  As shown in FIG. 3, first, in step S1, the conference management apparatus 10 (specifically, the video data analysis unit 31) executes video data analysis processing, detects a scene change in the video data PM, and receives video data. PM is divided into a plurality of partial images PTn (PT0, PT1, PT2,...) (See FIG. 7).

  Specifically, first, a switching portion (shooting time T1) between a scene relating to the first person and a scene displaying an image of the first page of the conference material DM is detected. Next, a switching portion (shooting time T2) between the scene in which the image of the first page of the conference material DM is displayed and the scene in which the image of the second page of the conference material DM is displayed is detected. Thereafter, similarly, a switching portion between the scene in which the image of the nth page of the conference material DM is displayed and the scene in which the image of the (n + 1) th page of the conference material DM is displayed (shooting time T (n + 1)) ) Is detected.

  Then, the video data analysis unit 31 converts the partial video PTn (moving image having a plurality of frame images) related to the period PDn from the shooting time Tn to the shooting time T (n + 1) into a scene video (moving image) related to the image on the nth page. Image). For example, the video data analysis unit 31 extracts the partial video PT1 over the period PD1 from the shooting time T1 to the shooting time T2 as a scene video related to the first page image.

  In addition, the video data analysis unit 31 displays a representative image (still image) in the partial video PTn, for example, the first frame image of the partial video PTn, and a scene image related to the nth (eg, nth page) display scene of the conference material DM. (Still image) Extracted as PSn. Specifically, the video data analysis unit 31 displays a representative image (still image) in the partial video PT1, for example, the first frame image of the partial video PT1, and the first display scene (first page is displayed on the conference material DM). Is extracted as a scene image PS1 (see FIG. 8). In addition, the video data analysis unit 31 converts a representative image (still image) in the partial video PT2, for example, the first frame image of the partial video PT2, into the scene image PS2 (FIG. 9) related to the second display scene of the conference material DM. Reference).

  Note that the video data analysis unit 31 analyzes the partial video PT0 and determines that the beginning scene (scene relating to the person) from time T0 to time T1 is not the scene relating to the conference material DM. Therefore, here, the scene image related to the display scene of the conference material DM is not extracted from the scene video (partial video) PT0 related to the opening person.

  In step S <b> 2, the audio data analysis unit 33 performs an analysis process on the audio data AD of the conference.

  Specifically, the audio data analysis unit 33 first extracts each partial audio data SAn corresponding to each scene image PSn from the audio data AD of the meeting. Specifically, the audio data of the corresponding period PDn of each scene image PSn is extracted as partial audio data SAn. In other words, the audio data analysis unit 33 divides the conference audio data AD into a plurality of partial audio data SAn. Note that the process of extracting the partial audio data SAn in the period PDn is performed using, for example, time data (shooting time data) in the video data PW and time data (recording time data) in the audio data AD.

  The voice data analysis unit 33 recognizes each partial voice data SAn and generates character data. Specifically, voice recognition processing is performed on each partial voice data SAn, and a voice signal in the partial voice data SAn is converted into character data (text data or the like). For example, character data CH1 as shown in FIG. 10 is extracted from the partial voice data SA1 related to the period PD1.

  Next, in step S3, the heading item extraction unit 38 performs a heading item HDn extraction process for each scene image PSn. For example, as shown in FIG. 11, the headline item extraction unit 38 performs image analysis processing (including OCR processing (optical character reading processing)) on the scene image PS1 (FIG. 8), and headline item HD1 (“ Productivity report ") is extracted. Similarly, the extraction process of the heading item HDn is executed for the other scene images PS2 and the like.

  In the next step S4, the bookmark item generating unit 39 generates “bookmark item” to be displayed in the bookmark display column RB in the minutes file PR. The item name of the “bookmark item” is determined based on the heading item HDn extracted by the extraction process for each scene image PSn and the character data CHn generated based on the partial audio data SAn. Note that the “bookmark item” of the minutes file PR also has the function of “headline” and can also be referred to as “headline item”. However, here, the headline item BMn of the minutes file PR is referred to as a “bookmark item” in order to distinguish it from the headline item HDn of the scene image PSn.

  Specifically, the bookmark item generation unit 39 first performs part-of-speech decomposition processing for each heading item HDn. For example, the part-of-speech decomposition process is performed on the heading item HD1 (“productivity report”), and the heading item HD1 is decomposed into “productivity” (noun) and “report” (noun). And the bookmark item production | generation part 39 detects that the heading item HD1 is comprised with two nouns (refer FIG. 11).

  Here, the bookmark item generation unit 39 is the number of nouns in the heading item HD1 (here, “2”) plus one (ie, (N + 1)) (here, “3”) nouns. It is determined that the item name of the bookmark item BM1 is configured. In other words, a name in which one noun is added to the heading item HD1 is determined as the item name of the bookmark item BM1. In addition, each (that is, (N + 1)) nouns are connected to each other by an appropriate particle or auxiliary verb. For example, in principle, (N + 1) nouns may be connected to each other using “no” (particle).

  Next, the bookmark item generation unit 39 analyzes the text related to each character data CHn extracted in step S2, and extracts words related to the heading item HDn (see FIG. 12).

  For example, first, as shown in the upper half area of FIG. 12, a plurality of sentences in the sentence related to the character data CH1 are each decomposed into parts of speech. In the top row of FIG. 12, the sentence “I will report on productivity” includes “productivity” (noun), “ni” (particle), “t” (verb), “te” (particle), “ It is broken down into each word (part of speech) of “report” (noun), “shi” (auxiliary verb), “mas” (auxiliary verb), and “.” (Special). A similar part-of-speech decomposition process is performed on other sentences in the sentence relating to the character data CH1.

  Each word is classified according to the appearance frequency. In the lower half area (left side) of FIG. 12, the words in the character data CH1 are shown sorted in the order of their appearance frequency (number of appearances). Specifically, words such as “productivity” (noun), “yoru” (verb), and “decrease” (noun) are shown in order of their appearance frequency.

  In addition, the bookmark item generation unit 39 extracts “nouns” from the part-of-speech decomposed words, and among the “nouns”, “productivity” (nouns) and “report” in the original heading item HD1. (Noun) is excluded from candidates for addition to eliminate duplication. Here, the noun “productivity” having the highest appearance frequency is also included in the original heading item HD1, and thus is excluded from the candidates for addition. Then, the next-ranked noun “decrease” is determined as a candidate for addition.

  Then, the bookmark item generating unit 39 detects the relationship between the noun “decrease” determined as the addition target candidate and the nouns (“productivity” and “report”) in the original heading item HD1. Here, since the term “decrease in productivity” exists at the beginning of the third sentence of the character data CH1, it is determined that “productivity” and “decrease” are highly related to each other. . It is also determined that “productivity” and “decrease” are preferably combined in this order (order in the third sentence) via “no” (particle).

  As a result, as shown in FIG. 13, the bookmark item generation unit 39 converts the three words “productivity”, “decrease”, and “report” in this order, and through this “no” (particle). The term “report of reduction in productivity” combined in the above is determined as the item name of the bookmark item BM1. That is, based on the heading item HD1 (“productivity report”), a bookmark item BM1 (“productivity reduction report”) is generated.

  Similarly, the item names are generated for the other bookmark items BM2 and the like.

  As described above, the bookmark item generating unit 39 analyzes the sentence related to the character data CHn, extracts the word related to the heading item HDn, and uses the extracted word to determine the item name of the heading item HDn. Change the bookmark item BMn. The word extracted from the character data CHn (target to be added in the bookmark item BMn) is a word related to the heading item HDn, and is a word other than the word included in the heading item HDn (in other words, the heading item HDn). It is preferable that it is a word that does not overlap with any other word.

  In the next step S5, the minutes file generator 35 performs a file generation process for the minutes file PR. As shown in FIG. 5 and FIG. 6 and the like, the generated minutes file PR includes a plurality of scene images PSn (PS1, PS2,...) And a plurality of bookmark items BMn (BM1) associated with each scene image PSn. , BM2, ...). The minutes file PR further includes a plurality of character data CHn and a plurality of partial voice data SAn.

  FIG. 4 is a flowchart showing the detailed operation of step S5.

  As shown in FIG. 4, first, in step S11, processing for arranging each scene image PSn in the minutes file PR is executed. Specifically, each scene image PSn is arranged in the upper area in the nth page PGn in the text display field RA of the minutes file PR. For example, as shown in FIG. 5, in the text display field RA, the scene image PS1 is arranged in the upper area in the first page PG1. Similarly, as shown in FIG. 6, the scene image PS2 is arranged in the upper region in the second page PG2.

  Further, in step S12, processing for arranging each character data CHn in correspondence with each scene image PSn in the minutes file PR is executed. Specifically, each character data CHn is arranged in a lower area in the nth page PGn (the same page as the scene image PSn) in the text display field RA of the minutes file PR. More specifically, as shown in FIG. 5, the character data CH1 is arranged in a lower area in the first page PG1 in the text display column RA. Similarly, as shown in FIG. 6, each character data CH2 is arranged in a lower area in the second page PG2.

  Further, in step S13, a process for arranging the bookmark item BMn in the minutes file PR is executed. Specifically, each bookmark item BMn is arranged in the bookmark display field RB of the minutes file PR. More specifically, as shown in FIG. 5, the bookmark item BM1 is arranged as the first item in the bookmark display field RB. The bookmark item BM2 is arranged as the next item in the bookmark display field RB. Note that the name generated in step S4 is used as the item name of each bookmark item BMn.

  Further, link data with the scene image PSn as a link destination is given to the bookmark item BMn. When an operation (click operation, etc.) is performed on the bookmark item BMn in the completed minutes file PR, the link data is referred to according to the operation, and the scene image PSn linked to the bookmark item BMn is displayed in the text. It is updated and displayed in the column RA.

  In step S14, a process of linking the partial audio data SAn to the position corresponding to the bookmark item BMn in the minutes file PR is executed. Specifically, each partial audio data SAn is linked to the proximity position of the corresponding bookmark item in the bookmark display field RB of the minutes file PR. More specifically, as shown in FIG. 5, an icon CN1 that schematically represents audio data is displayed at a position directly below the first bookmark item BM1 in the bookmark display field RB, and a part of the icon CN1 is partially displayed. The audio data SA1 is linked. Similarly, an icon CN2 schematically representing audio data is displayed at a position immediately below the second bookmark item BM2 in the bookmark display field RB, and the partial audio data SA2 is linked to the icon CN2. In this way, each partial audio data SAn is linked to the icon CNn associated with the corresponding bookmark item BMn in the bookmark display field RB.

  As described above, the minutes file PR is generated.

  According to the above embodiment, the scene image (still image) PSn related to the display scene of the conference material is extracted from the video data PM of the conference, and the bookmark item BMn related to the scene image PSn is automatically generated in the minutes file PR. Is done. Therefore, compared with the case of manually assigning items in the minutes, labor is reduced, and item information can be easily given to the minutes.

  Further, in the minutes file PR, a bookmark item BMn to which link data with the scene image PSn as a link destination is added is generated, and the bookmark item BMn is arranged in the bookmark display field of the minutes file PR. Therefore, by clicking the bookmark item BMn, the scene image PSn corresponding to the bookmark item BMn and the character data CHn corresponding to the bookmark item BMn can be displayed in the text display field RA. Therefore, it is possible to easily refer to the scene image PSn and the like using each bookmark item BMn.

  By the way, in the said embodiment, although bookmark item BMn is produced | generated based also on the character data CHn, this invention is not limited to this. For example, the heading item HDn extracted from the scene image PSn may be used as it is as the bookmark item BMn.

  However, the heading item HDn of the scene image PSn related to the conference material is often described very simply on the assumption that it will be supplementarily explained at the conference. Therefore, it is not always easy to grasp the contents of the minutes from only the term. For example, also in the above, the heading item HD1 of the scene image PS1 is “productivity report”, and there is not necessarily much information obtained only from this heading item HD1.

  On the other hand, in the above-described embodiment, the bookmark generation unit 37 is based not only on the heading item HDn extracted from the scene image PSn but also on the character data CHn generated based on the partial audio data SAn. The item name of the item BMn is determined. In other words, the item name of the bookmark item BMn is determined based on the heading item HDn extracted from the scene image PSn and the character data CHn generated based on the partial audio data SAn. Specifically, the bookmark generation unit 37 analyzes the sentence related to the character data CHn, extracts words related to the heading item HDn, and changes the item name of the heading item HDn using the extracted word. The name of the bookmark item BMn is generated. More specifically, one new word extracted from the character data CHn is added to the heading item HDn to generate a bookmark item BMn.

  For this reason, the amount of information related to the new bookmark item BMn is increased as compared with the amount of information of the heading item HDn. For example, the bookmark item BM1 is a “report of productivity decline”, and by additionally having the word “decrease”, the bookmark item BM1 is more than information obtained only from the heading item HD1 (“productivity report”). Have information. As described above, the item name of the bookmark item BMn includes more information than the original heading item HDn, and is an easily understood name.

  Therefore, the user can more easily grasp the contents of the minutes file PR by using the bookmark item BMn.

<3. Modified example>
Although the embodiments of the present invention have been described above, the present invention is not limited to the contents described above.

  For example, in the above-described embodiment, the case where the meeting minutes file PR is created after the meeting is illustrated, but the present invention is not limited to this. (In real time, it may be created).

  Moreover, in the said embodiment, the case where the new one word extracted from character data CHn was added to the heading item HDn, and the information amount regarding the bookmark item BMn was increased compared with the information amount of the heading item HDn was illustrated. However, it is not limited to this. For example, two or more new words extracted from the character data CHn may be added to the heading item HDn to increase the information amount regarding the bookmark item BMn compared to the information amount of the heading item HDn.

  In the above embodiment, the bookmark item BMn is generated by connecting the words in the heading item HDn and the predetermined number of words extracted from the character data CHn to each other using the particle “NO”. However, the present invention is not limited to this. For example, terms relating to the connection between words may be determined based on a database relating to the connection between words, and the accuracy as a natural language may be improved.

  In the above embodiment, the case where the bookmark item BMn is generated by changing the character string of the heading item HDn extracted from the scene image PSn (specifically, the order of the character string) is illustrated. It is not limited to. For example, in addition to the heading item HDn extracted from the scene image PSn, a character string including a character string different from the heading item HDn may be generated as the bookmark item BMn. Specifically, the character string “productivity report (decrease in productivity)” in which the character string “productivity report” generated from the character data CHn is added to the character string “productivity report” in the heading item HD1 in parentheses. May be generated as a bookmark item BMn.

  In the above embodiment (step S2 and the like), the case where the audio data of the corresponding period PDn of each scene image PSn is extracted as the partial audio data SAn is illustrated, but the present invention is not limited to this. For example, the voice data AD is independently subjected to voice analysis processing, the voice data is divided into a plurality of division data finer than the partial voice data SAn, and some of the division data is collected, thereby obtaining the partial voice. Data SAn may be generated. Further, the partial audio data SAn is not linked to the position corresponding to the bookmark item BMn, but a plurality of segment data finer than the partial audio data SAn are linked to the corresponding positions in the bookmark display field RB. You may do it.

10 Conference management device (MFP)
30 Display device 50 Computer 71 Video camera 72 Microphone 100 Conference system AD Audio data BM1, BM2, BMn Bookmark item CH1, CH2, CHn Character data HD1, HD2, HDn Headline item PM Video data PR Minutes file PS1, PS2, PSn Scene Image RA Text display field RB Bookmark display field SA1, SA2, SAn Partial audio data Tn Time

Claims (10)

A minutes creation system for creating meeting minutes,
Still image extraction means for extracting a still image related to the display scene of the conference material from the video data of the conference,
File generation means for generating a minutes file based on the still image;
Bookmark generating means for generating a bookmark item related to the still image in the minutes file;
A minutes creation system characterized by comprising: A minutes creation system according to claim 1,
Voice data analysis means for extracting voice data of a specific period corresponding to the still image from voice data of the meeting and generating character data by voice recognition of the voice data of the specific period;
Further comprising
The bookmark generating means includes:
A headline item extracting means for extracting a headline item from the still image;
Bookmark item generating means for determining an item name of the bookmark item based on the header item extracted by the header item extracting means and the character data generated based on the voice data;
Minutes creation system characterized by having a. A minutes creation system according to claim 2,
The bookmark item generation means analyzes a sentence related to the character data, extracts a word related to the heading item, changes the item name of the heading item using the extracted word, and stores the bookmark item. Minutes creation system characterized by generating. A minutes creation system according to any one of claims 1 to 3,
The minutes generation system, wherein the bookmark generation means assigns link data with the still image as a link destination to the bookmark item. A minutes creation system according to any one of claims 1 to 4,
The minutes creation system, wherein the bookmark generation means arranges the bookmark items in a bookmark display field of the minutes file. A minutes creation system according to claim 2 or claim 3,
The said file production | generation means arrange | positions the character string of the said character data corresponding to the said still image in the said minutes file, The minutes production system characterized by the above-mentioned. A minutes creation system according to any one of claims 2 to 6,
The minutes generation system characterized in that the file generation means links the audio data of the specific period to a position corresponding to the bookmark item. A minutes creation system according to claim 1,
Analyzing voice data of the conference, extracting voice data of a specific period corresponding to the still image, and voice data analysis means for generating voice data by voice recognition of the voice data of the specific period;
Further comprising
In the minutes file, the file generation means arranges the character data in correspondence with the still image. On the computer,
a) a procedure for extracting a still image related to the display scene of the conference material from the video data of the conference;
b) generating a bookmark item associated with the still image;
c) generating a minutes file having the still image and the bookmark item;
A program for running On the computer,
a) a procedure for extracting a still image related to the display scene of the conference material from the video data of the conference;
b) a procedure for extracting voice data of a specific period corresponding to the still image from the audio data of the meeting and generating character data by voice recognition of the voice data of the specific period;
c) a procedure for extracting a heading item from the still image;
d) generating a bookmark item having an item name determined based on the heading item and the character data;
e) generating a minutes file having the still image and the bookmark item associated with the still image;
A program for running

Images