JP2005197883A - Writing analysis apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a writing analysis apparatus for analyzing items written on a whiteboard. <P>SOLUTION: A flowchart indicating setting processes for annotation includes: a step S10 of receiving animation data resulting from photographing writing processes to the white board; a step S14 of acquiring information about staying time of a person; a step S18 of recognizing written items on the basis of a temporal change in data not displaying people; a step S20 of acquiring the written time; a step S22 of comparing the information of the staying time with the information of the write time to determine a person who writes the items; a step S24 of identifying the determined person from other people through pattern recognition applied to the animation data; and a step S26 of thereby setting the annotation to the animation data, the annotation relating to the items, the person for writing the items, and the writing time. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a technique for recording contents written on a writing board such as a white board or a blackboard.

  A system in which content written on a whiteboard is imaged by a scanner device attached to the whiteboard is widely used. In this system, when writing is performed by a plurality of persons, there is a drawback that the written contents cannot be associated with the written person.

  On the other hand, Patent Document 1 describes a technology of an electronic bulletin board having a special pen and a function of recognizing writing by the pen. In this technique, for example, each person having a different pen can record the identification of the person performing the writing and the written content in association with each other.

JP 09-171448 A

  A technique is known in which annotations such as sticky notes and memos are set on image data and displayed in an overlapping manner with the images.

  The technique of Patent Document 1 has an advantage that it is easy to specify a person who performs writing, but there is a problem that the installation cost increases due to special equipment used to identify the person who performs writing. is there.

  An object of the present invention is to associate written contents with a person who has written using highly versatile equipment.

  Another object of the present invention is to display the associated results in an easily recognizable form.

  The writing analysis apparatus according to the present invention includes input means for inputting time-series image data created by photographing a writing board on which a plurality of items are written by a plurality of persons during the writing period; Based on the time change of the writing board in the sequence image data, each written item is identified, and item analysis means for acquiring information about the writing timing of each identified item, Identification means for identifying the person to perform, acquisition of information about the stop timing, and specifying means for specifying the person who wrote each item based on the correspondence between the write timing and the stop timing, Prepare.

  A writing board is a white board, a blackboard, a pasted imitation paper, etc. that can be written using a writing instrument, and that the written contents are provided so that multiple people can view the written contents simultaneously. Point to. The writing board is used to share writing items when a plurality of persons meet such as a meeting. A person who writes on the writing board moves in front of the writing board and stops, writes items such as characters, symbols, and drawings, and leaves the place. Stopping means staying in the vicinity of the writing board and includes not only staying in one place but also moving back and forth in front of the writing board. Therefore, even when the user leaves the writing board for a short period, it can be regarded as a continuous stop. Other persons view the written items during or after writing. The writing of items is generally performed by a plurality of persons, and a plurality of items are cumulatively written on the writing board. The item refers to a group of written contents. Here, the term “group” refers to, for example, a description written by a person continuously, a description of each part when writing continuously in multiple places, a description written during a certain period, and the like. Refers to the unit of description written in the original.

  The input means inputs time-series image data obtained by photographing the writing board in this series of writing processes. The time-series image data may be temporally continuous moving image data, a set of still image data captured many times during the period, or a combination thereof. The item analysis means removes the image information about the person who appears in front of the writing board and analyzes the time change only for the writing board to identify the written item or the deleted item. In addition, information on the writing timing about when the writing was performed is also acquired. The person analyzing means identifies the person staying near the writing board using the time-series image data or other information. That is, it is determined who of the plurality of persons who performed the writing. In addition, using the time-series image data or other information, information on the stop timing about when the person stopped is also acquired. The specifying means associates which person performed which writing from the writing timing and the stop timing.

  The execution order of each means is not particularly limited. That is, it may be performed in the order of the item analysis unit, the person analysis unit, and the identification unit, or may be performed in the order of the person analysis unit, the item analysis unit, and the identification unit. Further, the person identification and the acquisition of the stop timing information in the person analysis means may be performed separately. For example, the person may be identified after the stop timing information is acquired and the specifying means is performed. Good.

  According to this configuration, by analyzing time-series image data created by a camera, a video camera, or the like, it is possible to associate the written item with the person who has written. Cameras and video cameras are generally widely used, and the introduction cost can be suppressed. In addition, since the user of the writing board does not need to do anything special in writing such as having a special pen, the writing board can be used smoothly. Note that not only time-series image data but also corresponding audio data may be acquired. Further, when an image taken separately by a scanner attached to the whiteboard or the like can be taken in, such an image may be used as the final image, for example. Note that the analysis target may be extended not only to a writing operation but also to a posting operation such as fixing a sheet on a board with a magnet.

  Preferably, in the writing analysis apparatus according to the present invention, the item analysis means includes means for acquiring information on the position of the identified item, and the person analysis means is configured to stop a person who stops near the writing board. Means for acquiring information about the position, and the specifying means writes at least one item written based on the correspondence between the position of the item and the position where the person stopped Identify people. According to this configuration, it is possible to easily determine which writing is performed by which person, particularly when a plurality of persons are simultaneously stopped near the writing board.

  Preferably, in the writing analysis apparatus according to the present invention, the person analysis means identifies a person staying near the writing board by performing pattern recognition analysis using the time-series image data. Various algorithms are known for pattern recognition, and a person can be identified by using them. In order to surely identify by pattern recognition, a recognition pattern may be learned by separately photographing a person who performs writing in advance or afterwards. In addition, the person who performs writing may wear, for example, different colors of tusk or bibs (simple clothes).

  Desirably, in the writing analysis apparatus of the present invention, the person analyzing means identifies a person who stops near the writing board by acquiring position information of an RFID tag possessed by each person. An RFID tag is an IC chip that can communicate without contact. By inputting the identification information in the chip, it is possible to calculate information about which person has stopped near the writing board.

  Preferably, the writing analysis apparatus of the present invention includes means for inputting infrared image data created by infrared imaging each person staying in the vicinity of the writing board, and the person analyzing means includes the red A person who stops near the writing board is identified by performing pattern recognition analysis using the outside image data. The infrared image data may be time-series data, or may be image data taken only when a person's stop is detected, for example. The infrared image data holds the temperature of the person to be photographed as luminance information. For this reason, when there is a body temperature difference between the persons who perform writing or a temperature difference due to a difference in clothes, each person can be easily recognized as a pattern.

  Preferably, the writing analysis apparatus of the present invention includes means for detecting the end of the writing period or the start of the next writing period based on the time change of the time-series image data. A series of writing is performed on the time-series image data by erasing or turning the board over by the writing board user when there is no new space on the writing board or when the topic is changed. Disappears from. Therefore, by detecting such an event, when a series of writing is repeated a plurality of times, each writing period is separated.

  Preferably, the writing analysis apparatus of the present invention is configured to write a person who writes each item and its writing timing based on a result specified by the specifying unit with respect to the time-series image data or processed image data obtained by processing the time-series image data. The annotation is provided with annotation setting means for setting the annotation to which the information is added. The annotation is additional information given to the time-series image data which is main information or processed image data obtained by processing the time-series image data. Specific examples of the additional information include emphasizing, leaving a note, and associating with other parts of a specific part or region of the main information. An annotation can be displayed in the form of, for example, a tag, a memo, a frame, an underline, coloring, or shading. The annotation setting means sets the annotation for each item so that the writing person of the item can recognize. As the processed image data of the time-series image data, for example, the time-series image data in which the display range is simply adjusted, the person who stops before the writing board or the person who crosses the front of the writing board is erased and written. Examples include time-series image data in which only the board is displayed, still image data obtained by cutting out an image at a specific time, and the like.

  Preferably, in the writing analysis apparatus of the present invention, the annotation setting means includes means for performing additional editing of annotation based on a user input. That is, the user can add a new annotation or write a memo in an already set annotation.

  Preferably, the writing analysis apparatus of the present invention includes display means for displaying a moving image or a still image based on the time-series image data or the processed image data, and the display means includes the still image and the still image. The writing timing display for displaying the writing person of each item in the drawing and the annotation indicating the writing timing information may be time information or writing order information.

  Preferably, in the writing analysis apparatus of the present invention, when the displayed annotation is selected by the user, the display means displays time-stage image data in which items corresponding to the annotation are written. That is, relying on the annotation, the image data of the time when the corresponding item is written or slightly before is displayed. As a result, it is possible to easily grasp the circumstances in which the matter of interest is written. When time-series display (animation display) is specified along with user selection of an annotation, time-series image data in which items corresponding to the annotation are written can be displayed in time-series.

  The writing analysis method of the present invention is a method executed by a computer, and is time-series image data created by photographing a writing board on which a plurality of items are written by a plurality of persons during the writing period. Detecting the time change of the writing board based on the item analysis step for identifying each written item, obtaining information about the writing timing of each identified item, and in the vicinity of the writing board Identifying a person to be stopped, a person analyzing step for obtaining information about the stop timing, and a specifying step for specifying a person who wrote each item based on the correspondence between the write timing and the stop timing; including.

  The writing analysis program of the present invention is an input procedure for inputting time-series image data created by photographing a writing board in which a plurality of items are written by a plurality of persons to the computer during the writing period. And a matter analysis procedure for identifying each written matter based on a time change of the writing board in the time-series image data, and obtaining information on a writing timing of each identified matter, and the writing Identification of the person who has written each item based on the identification of the person who stops near the board, the person analysis procedure for acquiring information about the stop timing, and the correspondence between the writing timing and the stopping timing And execute the procedure.

  Hereinafter, typical embodiments of the present invention will be described.

  FIG. 1 is a perspective view schematically showing how time-series image data used in the present embodiment is created. A white board 10 as a writing board is supported and fixed by a leg 12. Items 14 and 16 are already written on the whiteboard 10. At the time shown in the figure, the writer 18 standing in front of the whiteboard 10 writes the matter 22 with the pen 20. There are a plurality of people who participate in the conference in a range not shown in the figure, and the discussion is performed while looking at the whiteboard 10.

  Each item 14, 16, 22,. . . Is recorded by the video camera 24 installed on the front side of the whiteboard 10. The video camera 24 plays a role as a photographing unit that forms a part of the writing analysis system, and records the state of the whiteboard 10 and its peripheral part in the writing period in moving image data as time-series image data. The range 26 captured by the video camera 24 is slightly wider than the outer frame of the whiteboard 10, so that even when writing is performed on the periphery of the whiteboard 10, at least a part of the writer is captured. can do. Of course, priority may be given to increasing the shooting resolution of the whiteboard 10, and the shooting area may be matched to the frame of the whiteboard. The moving image data generated in this way is input to a PC (personal computer) 28 as a writing analysis device. As shown below, the PC 28 specifies the writer of each item based on the input moving image data, and performs processing for presenting the specified information to the user in a form that is easy to use. Note that the PC 28 can also input audio data generated together with the moving image data.

  Next, a configuration example of the writing analysis apparatus will be described with reference to FIG. FIG. 2 is a schematic functional block diagram of the writing analysis apparatus 30. The writing analysis device 30 is configured by a device that performs processing by defining the operation of hardware having a calculation function by software (program), such as the PC 28 in FIG. The writing analysis apparatus 30 includes an input / output unit 32, a user input unit 34, a display unit 36, a display processing unit 40, a storage unit 50, a control unit 60, an image analysis unit 70, an association processing unit 110, an annotation, as main components. A setting unit 120 is provided.

  The input / output unit 32 inputs / outputs data and command signals to / from an external device. Specifically, moving image data is input from a photographing device 150 such as a video camera 24 that generates moving image data. In addition, it is connected to the network 160 and communicates various data with an external device on the network 160, or receives a command signal for controlling the writing analysis device 30 from the external device. The user input unit 34 includes an input device such as a keyboard and a mouse, and accepts user commands from here. The display unit 36 includes a display device such as a display, and displays a moving image and a still image based on the input moving image data, in addition to general display for control.

  The display processing unit 40 performs data processing for display on the display unit 36, and includes an annotation processing unit 42 and a jump processing unit 44 that perform processing related to moving image data as a characteristic configuration. The annotation processing unit 42 performs processing for displaying the annotation set in the moving image data on the display unit 36 in the annotation setting unit 120 described later. Specifically, preprocessing for displaying a memo, a frame, or the like on the displayed moving image or image is performed. Further, when an annotation displayed together with a moving image or a still image is selected by the user, the jump processing unit 44 performs a process of displaying a moving image or a still image near the time when items corresponding to the annotation are written.

  The storage unit 50 includes a storage device such as a semiconductor memory or a hard disk, and stores various data, programs, and the like. The storage unit 50 is provided with a moving image database 52. In addition to the moving image data input through the input / output unit 32, the moving image database 52 stores moving image data and still image data obtained by processing the moving image data, annotation data set by the annotation setting unit 120, and the like. . The control unit 60 is configured using an arithmetic processing function of an MPU (ultra-small processing device), and controls the operation of each component of the writing analysis device 30.

  The image analysis unit 70 performs image processing on the moving image data stored in the moving image database 52. The image analysis unit 70 includes an item analysis unit 80, a person analysis unit 90, and a writing period setting unit 100. The item analysis unit 80 is a component that identifies each item written based on the moving image data, and includes a timing analysis unit 82 and a position analysis unit 84 as main components. The item analysis unit 80 compares the items described on the whiteboard 10 with the previous time, and detects that the corresponding writing (or erasing) has been performed when there is a change. The timing analysis unit 82 obtains write timing information indicating at which time each item was written by detecting the time before change detection and the time after change detection. Further, the position analysis unit 84 acquires information on which position on the whiteboard 10 each item has been written.

  The person analysis unit 90 is a component that identifies a person who has written based on moving image data, and includes a timing analysis unit 92, a position analysis unit 94, and a pattern recognition unit 96 as main components. The person analysis unit 90 detects a person image that appears in the moving image data. And the timing analysis part 92 acquires the stop timing information which the person stops in front of the whiteboard 10 by detecting the appearance time and exit time of the person. In addition, the position analysis unit 94 acquires information on the stop position indicating which position the person was at. The pattern recognition unit 96 identifies a person who appears in the moving image data by pattern recognition. The pattern recognition algorithm is not particularly limited. Further, pattern recognition may be facilitated by registering the characteristics of a person to be recognized in advance. Furthermore, by providing information on proper nouns of those persons in advance or after identification, it becomes possible to specifically specify the person who wrote them.

  The writing period setting unit 100 monitors a time change of the whiteboard 10 in the moving image data and detects a writing period in which a series of writing is performed. In general, in the whiteboard 10, the page is updated by deleting items described in the whiteboard 10. Therefore, this erasure is monitored to detect that a new writing period has been entered. The processing of the item analysis unit 80 and the person analysis unit 90 is performed for each writing period.

  The association processing unit 110 identifies the person who wrote the item by associating the item with the person based on the analysis results of the item analysis unit 80 and the person analysis unit 90. The association processing unit 110 includes a timing correspondence unit 112 and a position correspondence unit 114. The timing correspondence unit 112 detects which person has performed which writing by comparing the writing timing obtained in the item analyzing unit 80 with the stationary timing obtained in the person analyzing unit 90. That is, items written during a period when the person is stopped are treated as items written by the person. Further, the position correspondence unit 114 compares the item position information obtained by the item analysis unit 80 with the stop position information obtained by the person analysis unit 90. As a result, when two persons are in front of the whiteboard 10 at the same time, the person who is in a position close to the written item is considered to have written the item, and the person corresponding to the written item is specified. can do. If there are a plurality of candidates for a person who has written an item using the timing correspondence unit 112 and the position correspondence unit 114, the association processing unit 110 associates the item with the person with the longest residence time. Exceptional measures such as treating the writing person as unknown can be taken.

  The annotation setting unit 120 sets writing person information and writing timing information as annotations for the writing items associated by the association processing unit 110. An annotation is additionally given to moving image data as main data or processed moving image data obtained by processing the moving image data. Annotations can be set for specific locations or specific areas of the main data, and can be expressed in the form of sticky notes, notes, box, underline, coloring, shading, etc. . The annotation setting unit 120 includes an additional editing unit 122. The additional editing unit 122 is configured to add and edit annotations based on a user command, such as adding a memo to a set annotation or adding a new annotation.

  Next, how the writing analysis apparatus 30 is used will be described with reference to the flowcharts of FIGS.

  FIG. 3 is a flowchart for explaining the process from the input of moving image data until the annotation is set. First, moving image data is generated by the video camera 24 as the photographing apparatus 150, and the moving image data is input via the input / output unit 32 (S10). The input moving image data is stored in the moving image database 52 of the storage unit 50. The person analysis unit 90 analyzes whether or not a person is included in the moving image data. The analysis may be performed at regular intervals (for example, 1 second or 5 seconds) without being performed on all images. Then, person non-display data, which is time-series data obtained by collecting image data not including a person, is created and stored in the moving image database 52 (S12). In general, no change is seen in the whiteboard 10 during a period in which no person is included, so that the capacity of the person non-display data can be very small. The timing analysis unit 92 of the person analysis unit 90 acquires information on the stop timing at which the person was in front of the whiteboard 10, that is, the time of appearance and the time of departure (S14). Further, the image analysis unit 70 creates a still image showing the whiteboard 10 at the end of a series of writing periods set by the writing period setting unit 100 (S16), and the created still image is the moving image database 52. Stored in

  Next, the item analysis unit 80 determines whether there is a time difference between the non-display data of the person stored in the moving image database 52, and recognizes the written item (S18). At the same time, the timing analysis unit 82 of the item analysis unit 80 acquires the writing start time and the writing end time as the writing timing of this item (S20).

  Subsequently, the association processing unit 110 compares the stationary timing information obtained in step S14 with the writing timing information obtained in step S20 to determine the person who performed each writing ( S22). The determined person is recognized by the pattern recognition unit 96 of the person analysis unit 90 (S24). Each identified person is distinguished by a proper noun based on information given in advance. The annotation setting unit 120 gives, as an annotation, information on the proper noun of the person who wrote it and information on the writing timing for each item in the original moving image data or person non-display data (S26). The set annotation data is stored in the moving image database 52. Thereby, each user can access the moving image data (the processed data) on which the annotation is set via the network 160.

  FIG. 4 is a flowchart for explaining an annotation display process. The user who wants to grasp the outline of the content of the meeting normally instructs the display of the final image of the whiteboard 10 and the corresponding still image is displayed on the display unit 36 (S30). However, if the contents cannot be understood well, the user instructs the control unit 60 to perform annotation display through the user input unit 34. The control unit 60 instructs the annotation processing unit 42 of the display processing unit 40 to display the corresponding annotation, and thereby the annotation is displayed on the display unit 36 so as to be superimposed on the final still image ( S32). For example, the annotation may be displayed stepwise in order of writing time. That is, each time the mouse or the like is clicked, the annotation for the first writing is displayed, and the annotation for the second writing is displayed by the next click. In the annotation corresponding to each item, the proper noun of the person who wrote it is written. The user can deepen the understanding of the writing items by relying on the information of the writing person.

  Furthermore, the user who wants to grasp the detailed situation instructs to display a moving image and selects an annotation of a matter of interest (S34). As a result, the image data jumps to the time when this item is written, and the corresponding image data is displayed. Then, the moving image is displayed from that time (S38). The moving image display may be performed along with the reproduction of the audio recording in real time, or may be performed in a frame advance state (so-called flip-flop cartoon state) or the like.

  In the above description, the person analysis unit 90 has been described as performing person analysis based on moving image data. However, other than this, a method for analyzing a person can be cited. For example, each person can have an RFID tag and communicate with the RFID tag to grasp the location of the RFID tag. Then, by associating the identification information of the RFID tag with the person, it is possible to identify the person who wrote the item without performing analysis by the pattern recognition unit 96. Further, for example, when a person who performs writing is separately photographed with an infrared camera, identification using the temperature difference of each person becomes possible. As described above, when using the captured image of the RFID tag or the infrared camera, the writing timing and the writing position may be acquired using such information, or the timing analysis unit 92 and the position analysis may be performed. The writing timing and the writing position may be acquired based on the unit 94.

  Next, an example in which annotation display is performed will be described with reference to FIGS.

  FIG. 5 is a diagram illustrating an example of the still image 170 in the final state of the whiteboard 10 that has been written in a certain meeting. Various items such as characters, diagrams, and tables are written in the still image. However, in this state, the order of writing and the person who performed writing may not be known, and the contents may not be fully understood.

  6 to 8 are diagrams illustrating an example in which annotations are displayed one after another on the same still image 170 as in FIG. 5 based on a user instruction. First, in FIG. 6, the portion where writing is performed first is clearly indicated by the annotation in the surrounding frame 172. In the vicinity of the surrounding frame 172, an annotation 172a indicating the proper noun “Mr. Miyazaki” of the person who wrote the matter is displayed. The annotation 172a is numbered (1), indicating that it was written first.

  FIG. 7 is a diagram showing the next stage of FIG. The display of FIG. 7 is performed when the user clicks the mouse in the stage of FIG. Here, in addition to the enclosure frame 172 and the annotation 172a shown in FIG. 6, an enclosure frame 174 and an annotation 174a are displayed. As a result, the user can understand that “Mr. Terao” described the items enclosed by the enclosure frame 174 in the second stage.

  FIG. 8 is a diagram showing a display in the final stage. Up to this point, annotations are displayed one after another based on the user's instructions, and the frame 172, 174, 176, 178, 180 in which each item is divided, and the proper noun of the person who wrote each item and Annotations 172a, 174a, 176a, 178a, 180a displaying the sequence numbers are sequentially displayed. In the last stage, a memo 182 that is an annotation pasted by the user is displayed. This memo 182 includes a character description column 184 comprising six pages, a page display column 186, and a traveling direction instruction column 188. The user can view the contents described in other pages by operating the pointer 190 with a mouse or the like and clicking the arrow in the traveling direction instruction column 188. In this way, it is effective to take contents edited later by the user as annotations and display them together with annotations indicating areas and the like in organizing and browsing the contents of the conference.

  In this final stage of FIG. 8, there may be cases where it is desired to know further details about each item. Therefore, for example, when it is desired to know in detail the item enclosed by the enclosure frame 178, the user can select the enclosure frame 178 to jump to the time at which this item is described. That is, the image at the time in the original moving image data or time-series data obtained by removing a person from the original moving image data is displayed. Furthermore, if the moving image display is performed at the same time, it becomes possible to grasp the details of writing near this time. In this case, it is also effective to simultaneously reproduce audio data.

  In addition to the display examples shown here, annotations can be displayed in various forms. For example, instead of displaying an area using a surrounding frame, it is also effective to display the area by color or shading. In this case, descriptions by the same person may be displayed in the same color, or may be displayed so that the color is gradually changed according to the writing order. Further, in the moving image display, it is not always necessary to sequentially display the written items, and the display time can be shortened by displaying them collectively for every predetermined time. Such a display instruction may be implemented by providing a small window for performing the instruction.

It is a figure which shows the example which image | photographs the writing process to a white board. It is a functional block diagram which shows the example of schematic structure of a writing analysis apparatus. It is a flowchart which shows the setting process of annotation. It is a flowchart which shows the display process of an annotation. It is a figure of the still picture which shows the white board of the final state. It is a figure which shows the state which performed a part of annotation display to the still image of FIG. FIG. 7 is a continuation of FIG. 6. It is a figure which shows the state which performed all the annotation display to the still image of FIG.

Explanation of symbols

  10 Whiteboard, 12 Legs, 14, 16, 22 Items, 18 Writers, 20 Pen, 24 Video Camera, 26 Range, 28 PC, 30 Writing Analyzer, 32 Input / Output Unit, 34 User Input Unit, 36 Display unit, 40 display processing unit, 42 annotation processing unit, 44 jump processing unit, 50 storage unit, 52 moving image database, 60 control unit, 70 image analysis unit, 80 item analysis unit, 82 timing analysis unit, 84 position analysis unit, 90 person analysis unit, 92 timing analysis unit, 94 position analysis unit, 96 pattern recognition unit, 100 writing period setting unit, 110 association processing unit, 112 timing correspondence unit, 114 position correspondence unit, 120 annotation setting unit, 122 addition Editing unit, 150 photographing device, 160 network.

Claims (12)

Input means for inputting time-series image data created by photographing a writing board in which a plurality of items are written by a plurality of persons during the writing period;
Item analysis means for identifying each written item based on a time change of the writing board in the time-series image data, and acquiring information about a writing timing of each identified item;
Identification of a person who stops near the writing board, and person analysis means for acquiring information about the stop timing;
Based on the correspondence between the writing timing and the stop timing, a specifying means for specifying the person who wrote each item;
A writing analysis apparatus comprising: The writing analysis apparatus according to claim 1,
The matter analysis means includes means for obtaining information about the position of the identified matter,
The person analysis means includes means for acquiring information about a position where a person who stops near the writing board has stopped,
For the at least one item written, the specifying means specifies the person who wrote based on the correspondence between the position of the item and the position where the person stopped. apparatus. The writing analysis apparatus according to claim 1,
The writing analysis apparatus characterized in that the person analysis means identifies a person staying near the writing board by performing pattern recognition analysis using the time-series image data. The writing analysis apparatus according to claim 1,
The writing analysis apparatus characterized in that the person analysis means identifies a person staying in the vicinity of the writing board by acquiring position information of an RFID tag possessed by each person. The writing analysis apparatus according to claim 1,
Means for inputting infrared image data created by infrared imaging each person staying near the writing board;
The writing analysis apparatus characterized in that the person analysis means identifies a person staying near the writing board by performing pattern recognition analysis using the infrared image data. The writing analysis apparatus according to claim 1,
A writing analysis apparatus comprising: means for detecting the end of the writing period or the start of the next writing period based on a time change of the time-series image data. The writing analysis apparatus according to claim 1,
Annotation setting means for setting an annotation to which information on the writing person of each item and the writing timing is set based on the result specified by the specifying means for the time-series image data or processed image data obtained by processing the time-series image data A writing analysis apparatus comprising: The writing analysis apparatus according to claim 7,
The writing analysis apparatus, wherein the annotation setting means includes means for adding or editing an annotation based on a user input. The writing analysis apparatus according to claim 7,
Based on the time-series image data or the processed image data, comprising a display means for displaying a moving image or a still image,
The writing analysis apparatus characterized in that the display means displays, together with a still image, an annotation indicating information on a writing person of each item in the still image and writing timing thereof. The writing analysis apparatus according to claim 9,
The writing analysis apparatus, wherein when the displayed annotation is selected by a user, the display means displays time-stage image data in which a matter corresponding to the annotation is written. A method performed by a computer,
Writing is performed by detecting time changes of the writing board based on time-series image data created by photographing a writing board on which a plurality of items are written by a plurality of persons during the writing period. An item analysis step for identifying each item and obtaining information about the writing timing of each identified item;
A person analysis step for identifying a person who stops near the writing board and acquiring information about the stop timing;
Based on the correspondence between the writing timing and the stop timing, a specific step of specifying the person who wrote each item;
A writing analysis method comprising: Against the computer
An input procedure for inputting time-series image data created by photographing a writing board on which a plurality of items are written by a plurality of persons during the writing period;
An item analysis procedure for identifying each written item based on the time change of the writing board in the time-series image data, and acquiring information about the writing timing of each identified item;
A person analysis procedure for identifying a person who stops near the writing board and obtaining information about the stop timing;
Based on the correspondence between the writing timing and the stop timing, a specific procedure for identifying the person who wrote each item,
A writing analysis program characterized in that

Images