JP5316248B2 - Video conference device, video conference method, and program thereof - Google Patents

Description

  The present invention relates to, for example, a video conference apparatus, a video conference method, and a program thereof used in a video conference performed between a plurality of points separated from each other.

  A video conference apparatus used in a conventional video conference will be described. The generation unit generates a feature amount (for example, a face feature amount) necessary for person image identification (face image identification) from a person image (face image) of a conference participant photographed from a TV camera in advance. The generated feature value and the profile of the conference participant are associated with each other and registered in the database storage unit. When the profile is displayed for the conference participant displayed on the display of the partner terminal through the TV camera of the own terminal, the generating unit generates the feature amount again. Then, the video conference that collates the feature quantity that matches the feature quantity, sends the profile corresponding to the feature quantity to the requesting counterpart terminal and displays it on the display of the counterpart terminal when there is a matching feature quantity An apparatus is already known (see, for example, Patent Documents 1 and 2).

However, in the case of a conventional video conference apparatus, face image identification must be performed in order to generate a feature amount again. In general, face image identification has the highest identification rate when the displayed person faces the camera, and has a property that the identification rate decreases as the face direction changes from diagonal to horizontal. In a video conference in which a plurality of persons participate, it is difficult for all participants to face the camera, and there may be participants whose faces are photographed obliquely or sideways. Even when the face faces the camera, the identification rate similarly decreases when the display size of the face portion is small or the room is dark. In these cases, there is a problem that all the participants cannot be correctly identified by the face image identification technique alone and an appropriate profile cannot be displayed.
In view of the above-described problems, an object of the present invention is to provide a video conference apparatus, a video conference method, and a program for appropriately transmitting a profile to a communication terminal of a partner while reducing the frequency of using a participant's feature amount. For the purpose.

  In order to solve the above problem, the video conference apparatus according to the present embodiment includes a database storage unit in which a feature amount for identifying a prospective participant and attribute information indicating an attribute of the prospective participant are registered in association with each other. The feature amount generation unit that generates the feature amount of the participant photographed by the photographing device, and the position information of the photographed participant are detected and associated with the position information and the attribute information of the photographed participant. Participants photographed at the second and subsequent times by identifying the participant based on the position information detection unit registered in the database storage unit, the feature amount of the photographed participant and the feature amount in the database storage unit The identification unit for identifying the participant based on the location information of the participant and the location information in the database storage unit, the attribute information of the participant identified by the feature amount, and the video of the photographed participant Send And a transmission unit for transmitting in association with the video of participants captured in the second or subsequent attribute information of the participants identified in the information.

  If the video conference apparatus, video conference method, and program thereof according to the present invention are used, the frequency of using feature amounts is reduced, so that the accuracy of participant identification can be improved and the profile can be appropriately transmitted to the other communication terminal.

The figure which showed the function structural example of the video conference system of a present Example. The perspective view of an example of the imaging device used with the video conference system of a present Example. The figure which showed an example of the form of the meeting suitable for the video conference apparatus of a present Example. The figure which showed the function structural example of the video conference apparatus of a present Example. The flowchart figure which showed the flow of the main processes of the video conference apparatus of a present Example. The figure which shows an example of the data table of a present Example. The figure which showed an example of the image after a participant's cutout. The figure which showed an example of the image when the cut out participant is represented by two persons. The figure which shows an example of the image on which attribute information was superimposed. The figure which shows an example of the data table with which positional information was matched. The figure which shows an example of the image at the time of making it display on a display part. The flowchart figure which showed a part of flow of the main processes of the video conference apparatus of another Example. The figure which shows an example of the image which matched error information.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In addition, the same number is attached | subjected to the process part which performs the same process with the same function, and duplication description is abbreviate | omitted.

Hereinafter, the video conference apparatus according to the first embodiment will be described. In the video conference performed between a plurality of points distant from each other, the video conference apparatus is a video of a photographing device (for example, a TV camera) at each point. Are used to be sent to other points via the network.
FIG. 1 shows a functional configuration example of a video conference system 1000 including the video conference apparatus 100 according to the first embodiment. In the following description, it is assumed that a video conference is performed at an L point and an M point that are separated from each other, and the video conference system 1000 is disposed at each point. Then, the L point is set as the self-side point, the M point is set as the partner-side point, the L-point side video conference device is the first video conference device 100, and the M-point side video conference device is the second video conference device 200. And In addition, the video conference apparatus of a present Example can perform a video conference also in three or more points.

  As shown in FIG. 1, the video conference system 1000 includes an operation unit 102, a video output unit 103 (for example, a display), an audio output unit 104 (for example, a speaker), a photographing device 106 to be described later, a hard disk 108, a memory 110, The video conferencing apparatus 100 and the communication control unit 112 are configured.

  In the video conference system 1000, components other than the photographing device 106 and the video conference device 100 are components that are generally used for a PC or the like. The operation unit 102 is, for example, a keyboard or a mouse. The video output unit 103 outputs the other party's video conference scenery and the participants of the video conference (that is, if it is the L point, the M point video conference scenery and the participant), and the audio output unit 104 is the other party's television conference. The audio of the conference participant (the audio of the video conference participant) is output.

FIG. 2 shows a perspective view of an example of the photographing apparatus 106. The imaging apparatus 106 of the present embodiment includes a first cylindrical portion 106a and a second cylindrical portion 106d having different diameters. On the outer peripheral surface of the first cylindrical portion 106a, photographing means 106c are arranged at equal intervals. The photographing means 106c is, for example, a television camera, and photographs the participants in the conference. On the outer peripheral surface of the second cylindrical portion 106b, sound collecting means 106d are arranged at equal intervals in the circumferential direction. The sound collecting means 106d is, for example, a microphone, and picks up the voices of conference participants. Thus, in this embodiment, as shown in FIG. 2, it is preferable to use a photographing apparatus capable of photographing all directions and collecting sound. The photographing apparatus 106 is “a system that automatically creates conference minutes contents from video / audio data of a conference” [online] October 14, 2008, National Institute of Advanced Industrial Science and Technology [May 21, 2009]. Date search], Internet <URL: http://www.aist.go.jp/aist_j/press_release/pr2008/pr20081014_2/pr20081014_2.html>.
Further, the photographing device 106 may not be capable of photographing in all directions as described above, but may be a photographing device that automatically zooms and focuses on a speaker.

  FIG. 3 shows an optimum video conference mode in which the video conference device 100 according to the first embodiment is used. As shown in FIG. 3, the video conference is preferably configured such that the participants are positioned so as to surround a circle and the photographing device 106 is positioned in the center. This is because the photographing apparatus shown in FIG. 2 can always photograph the faces of all the participants facing each other.

FIG. 4 illustrates a functional configuration example of the video conference apparatus 100 according to the first embodiment. FIG. 5 shows a main processing flow of the video conference apparatus 100. The video conference apparatus 100 according to the first embodiment includes a cutout unit 2, a feature amount generation unit 6, a position information detection unit 8, an identification unit 10, a synthesis unit 12, an encoding unit 14, and a database storage unit 16. Has been.
First, before starting a video conference, a data table as shown in FIG. 6 is registered in the database storage unit 16 in advance. As shown in FIG. 6, the data table associates the attribute information and the feature amount of the prospective participant in the video conference. In the example of FIG. 6, a data table for nine prospective participants whose names are A, B, C, D, E, F, G, H, and I is shown. A prospective participant is a person who is planning to participate, and indicates, for example, an actual participant and a person who has been scheduled to participate but can no longer participate.
Here, the attribute information is information indicating the attribute of the prospective participant in the video conference. In the example of FIG. 6, the attribute information includes name information indicating the name and profile information. In the example of FIG. 6, the profile information is the title of the prospective participant. For example, the profile information (title) of a prospective participant whose name is A is “development manager”. Further, as shown in FIG. 6, past achievements may be registered in addition to the titles for the prospective participants of names B and C. For example, for the prospective participant of name B, the past performance “Successfully completed XX project” is registered as profile information in addition to “Planning department manager”.
The feature amount is information for identifying a prospective participant. The feature amount includes, for example, a face feature amount, a voice feature amount, a fingerprint feature amount, and a retinal feature amount. From the viewpoint of ease of processing, the feature quantity is preferably a face feature quantity or a voice feature quantity. For example, the facial feature value is a numerical value of a facial feature of a user who can identify and identify the person and others. For example, the shape of the face components (eyes, nose, mouth, eyebrows, etc.) and the features such as the mutual arrangement are expressed numerically. Further, for example, a digitized face image of a registrant acquired in advance may be used as face data, for example, JPEG image data with a prescribed size.
The speech feature amount is a feature amount vector that can express speech features with a small amount of information, and is a feature amount vector composed of cepstrum and physical amounts of dynamic features of the cepstrum, for example.
As a data table creation process in the case of using a face feature amount as the feature amount, first, a face is photographed by the photographing means 106 with respect to a prospective participant of the video conference. And the process of the following flows by the clipping part 2 demonstrated below is performed.
(1) Detection of face area from input video, specification of position of face area (2) Cutout process of face area (see FIG. 7) by cutout unit 2 (3) Size, brightness, etc. of cutout face area Normalization processing of variation (4) Extraction processing of face feature amount from normalized face area When the extraction processing of face feature amount is completed, for example, the attribute information (name) of the prospective participant using the keyboard of the operation unit 102 Information and profile information) are input and associated with the facial feature amount and registered in the database storage unit 16. In the following description, the feature amount registered in the database storage unit 16 is referred to as a registered feature amount.

  The process of the video conference apparatus 100 after registering the data tables of all prospective participants in the database storage unit 16 will be described. Moreover, the form of a video conference is an example shown in FIG. 3, and it is assumed that participants in the conference are A to H and I is absent.

Assume that participant B speaks during the meeting. Then, the participant B is photographed by the photographing device 106, the uttered voice is collected, and the position information detection unit 8 detects position information (described later) (step S1). In the following description, the first shooting is the first shooting, and the second and subsequent shootings are the second shooting.
The audio signal a of the participant B is input to the position information detection unit 8. The video signal b of the participant B is input to the cutout unit 2. As described above, the photographing apparatus shoots in all directions (360 degrees), but the clipping unit 2 clips the image so as to obtain an image of the speaker B that is substantially directly facing. The cut-out process is as described in (1) to (4) above. The cutout unit 2 outputs the cut out face image information c and the face area location information d (hereinafter referred to as “face location information”). The face image information c is input to the feature amount generation unit 6, and the face location information d is input to the position information generation unit 8. An example of the face image information c is shown in FIG.
Next, the control unit 18 determines whether or not the shooting is the second or later shooting (step S2). Here, since it is the first shooting (No in step S2), the process proceeds to step S3. A method for determining whether or not the second or subsequent shooting is performed by the control unit 18 will be described later.

  The feature amount generation unit 6 generates the feature amount of the first photographed participant B using the face image information c (step S3). The feature amount is, for example, a face feature amount or a voice feature amount as described above. When the audio feature amount is used, the audio signal a from the position information detection unit 8 is used.

The identification unit 10 performs identification based on the registered feature amount in the database storage unit 16 and the feature amount from the feature amount generation unit 6 (step S4). In this example, the identification unit 10 includes a face identification unit 152, a voice identification unit 154, and a position information identification unit 156. Either the face identification unit 152 or the voice identification unit 154 may be used.
The identification unit 10 identifies a participant based on the registered feature amount in the database storage unit 16 and the feature amount generated by the feature amount generation unit 6. In the following description, the first identification (identification using the feature amount) is referred to as “first identification”, and the second and subsequent identifications (identification using position information described later) are referred to as “second identification”. Specifically, the similarity between the feature quantity and the registered feature quantity is calculated. When the feature amount is a numerical value, for example, the reciprocal of the absolute value of the difference between the feature amount and the registered feature amount is calculated as the similarity. Also, a value obtained by adding minus to the absolute value of the difference between the feature quantity and the registered feature quantity is calculated. In addition, any degree of similarity may be used as long as it indicates the degree of similarity between the feature quantity and the registered feature quantity.
Then, a participant having a feature amount with respect to a similarity whose calculated similarity is larger than a predetermined first threshold is identified (hereinafter referred to as “similarity using method”).
Another example is the mutual subspace method. In the mutual subspace method, when there are two subspaces to be examined, the angle between the subspaces (canonical angle (N-dimensional subspace exists in N)) is calculated and the obtained positive This is a technique in which the smallest angle among the quasi-angles is used as the similarity.
In addition, as an identification method using the face feature amount of the face identification unit 152, there is a method of dividing a participant's face image into mosaics, extracting feature points for each of those mosaics, and collating them. In addition, there is a method of performing collation based on isodensity distribution extracted from the participant face image.

  Further, the identification method using the voice feature amount of the voice identification unit 15 is, for example, a database storage unit using, as a feature amount, voiceprint data obtained by inputting a voice to be authenticated and analyzing the input voice. 16 is stored.

  In addition, as shown in FIG. 8, the accuracy of specifying the direction of the position information detection unit 8 is insufficient, and there may be a case where a person between two persons (center line W in FIG. 8) is regarded as a speaker. In this case, identification processing by the identification unit 10 is performed for each of the two or more participants.

  Further, if the types of feature quantities used for identification (in the above example, face feature quantities and voice feature quantities) are increased, the identification accuracy of the identification unit 10 can be increased.

  Next, the synthesizing unit 12 extracts attribute information corresponding to the feature amount of the participant identified by the identifying unit 10 from the database storage unit 16. Then, the synthesizing unit 12 associates the identified participant video imaged by the imaging device 106 with the extracted attribute information. Here, associating, for example, attribute information is superimposed on the video of participant B as shown in FIG. In addition to superposition, other methods may be used as long as the participants of the meeting at the point B can understand that the video of the participant B corresponds to the attribute information. The associated participant's video and attribute information are input to the encoding unit 14.

On the other hand, an audio signal from which sounds other than the participants have been removed by the position information generation unit 8 is also input to the encoding unit 14. The encoding unit 14 encodes the video and attribute information of the associated participant and the sound from the sound source localization unit 4, and transmits the video conference device on the M point side to the second video conference device 200 (step) S6).
Next, position information will be described. In step S <b> 1, as described above, the position information detection unit 8 obtains the position information e for the face images (participants) cut out from the cut-out unit 2 during the first shooting (first shooting). The position information e is, for example, an angle θ (radian) at which the speaking participant B is located when the participant A is used as a reference. This criterion may be another location. In addition, it is preferable that the cutout by the cutout unit 2 is performed so that the direction θ of the speaker matches the center position of the cutout image in the horizontal direction.
However, as shown in FIG. 8, there is a case where the position information generating unit 8 has insufficient accuracy for specifying the direction, and the direction between the two persons (center line W) is regarded as the direction of the speaker. In this case, it is preferable that the position information e is represented by pixels. In the example of FIG. 8, the angle of the center line W when the participant A is the reference is θ. The position information in this case is the resolution (number of all 360 degrees in the horizontal direction) obtained from the angle θ and the direction (angle from the reference direction) corresponding to the relative coordinate x in the horizontal direction on the clipped image. Can be uniquely determined. For example, when 360 degrees are displayed by N pixels, the position information can be (θ · N / 2π) + x (pixels).
Here, the relative coordinate x corresponds to the face location information d obtained by the cutout unit 2. As described above, even when the position information generation unit 8 uses the pixel as the position information between the two persons (center line) with the direction of the speaker as the sound source direction, the speaker is uniquely identified. Position information can be detected. In addition, the position information e may be anything as long as it uniquely determines the participant. Further, as shown in FIG. 7, even when the speaker direction θ and the horizontal center position of the cut-out image are cut out, the position information is expressed in pixels, that is, θ · N / 2π. It may appear.
After the first identification of the participant (after completion of step S4), the position information detection unit 8 associates the position information e with the attribute information of the photographed participant (participant B in this example), and stores the database. Register in the unit 16 (step S10). That is, as shown in FIG. 10, the position information β ₂ is additionally updated for the participant B. In this way, the participant utters and, after the first shooting, generates the feature amount of the feature amount generation unit 6 and detects the position information e of the participant uttered by the position information detection unit 8, and the database The storage unit 16 is additionally updated as needed in association with the attribute information of the participant who spoke. Hereinafter, the position information e in the database storage unit 16 is referred to as registered position information e. When the processes of step S6 and step S10 are finished, all the processes for the first shooting of the video conference apparatus are finished.

Next, it is assumed that the participant B speaks again after the participant B finishes speaking. In this case, it starts again from the start in the flowchart of FIG. In this case, the imaging device 106 images the participant B who is the speaker and detects the position information e ′ (step S1). And the control part 18 judges whether the said imaging | photography is the imaging | photography after the 2nd time (step S2). The control unit 18 may determine whether or not the second and subsequent shootings are performed by determining whether or not the position information about the direction of the sound source estimated by the position information detection unit 8 is registered in the database storage unit 16. . That is, the fact that the second and subsequent shootings have been performed means that the location information of the participant who has performed the shooting is registered in the database storage unit 16.
Since the participant B is in the second and subsequent shooting (Yes in step S2), the position information identification unit 156 performs identification based on the position information e in the database storage unit and the detected position information e ′. . The identification method using the position information by the position information identification unit 156 is the same as the identification method using the feature amount, and for example, the similarity may be used.
If the second identification is made, the synthesizing unit 12 refers to the database storage unit 16, and the attribute information corresponding to the position information e in the database storage unit 16 having a high similarity with the detected position information e ′, for example, 2 The transmission unit transmits the video image of the participant (participant B in this example) photographed after the first time, and transmits it (step S14). When the process of step S14 ends, the process for the second and subsequent shootings of the video conference apparatus ends.

  When a participant who has never been photographed (for example, participant C) is photographed, No is determined in step S2, and the feature value generation unit 6 generates a feature value of the participant C in step S3. . Then, the video conference apparatus performs steps S4, S6, and S10.

As described above, the video conference apparatus 100 according to the first embodiment detects the position information of the participant who is photographed for the first time by speaking and registers the position information in the database storage unit 16 in association with the attribute information. For the participants photographed for the second time and thereafter, the position information is used to identify the participants, and the attribute information is extracted and transmitted without generating the feature amount. Therefore, the frequency of feature quantity generation can be reduced, and the identification rate of participants can be reduced even when the face is not directly facing the camera, the display size of the face portion is small, or the room is dark. Appropriate attribute information can be transmitted.
In addition, the photographing apparatus 106 may use a camera array that can omnidirectionally photograph a plurality of cameras with all photographing surfaces facing outward. Alternatively, a photographing device that automatically photographs the speaker.

  The video conference apparatus 200 according to the second embodiment is different from the video conference apparatus 100 according to the first embodiment in that the display section 22 is included. As a result of the first identification using the above-described similarity use method by the identification unit 10 described in the first embodiment, there may be a plurality of similarities greater than the first threshold. In that case, the attribute information of a plurality of candidates of the first identified participant having the feature quantities regarding these similarities is displayed on the display unit 22.

  An example displayed on the display unit 22 is shown in FIG. The example of FIG. 11 shows a case where two participants are displayed. In the example of FIG. 11, for the left participant, the name (A, B, or C) is given to the user (usually a person who is usually a conference participant at the point L and knows about the participant). Display to be selected. The user inputs (clicks) using the operation unit 102 (for example, a mouse). When the correct participant does not correspond to all the displayed candidate participants, the correct name is entered in the lower input space Y by the operation unit 102 (for example, a keyboard). In addition, for the participant on the right side of FIG. 11, the participant with the name D and the input space Y are displayed. Also, the left person is about to be entered as participant B, and the right person is about to be entered as participant D.

  The input by the user is the same for the second identification (identification using position information) process.

  FIG. 12 shows a part of the main processing flow of the video conference apparatus 200. In the processing flow of the video conference apparatus 200 according to the second embodiment, the flowchart shown in FIG. 12 is inserted between step S4 and step S6 shown in FIG. 3, and step S10 is moved to the position shown in FIG. Is.

  In step S102, when there is a first identification candidate person (Yes in step S102), the control unit 18 displays the candidate person names (participants A, B, and C in the above example) on the display unit 22. It is displayed (step S106). Then, the user is allowed to select a person name from candidate persons, or the person name is input to the input space Y (step S108).

  On the other hand, if no candidate person exists as a result of the first identification in step S102 (No in step S102), the control unit 18 displays an input space on the display unit 22 and inputs a person name (step S104). . Then, the attribute information of the participant person name selected or input and the video and audio of the participant are transmitted to the second video conference device 200 (step S6). Further, the position information and the selected or input person name are associated with each other and stored in the database storage unit 16 (step S10). When step S6 and step S10 are completed, the processing for the first shooting of the video conference apparatus according to the second embodiment is completed.

  Moreover, what is necessary is just to make the video output part 103 output the screen (for example, FIG. 11) on which a participant is selected. In this case, the meeting scenery at B and the participant selection screen may be displayed on two screens or may be switched automatically. Moreover, you may install the video output part for a participant selection screen.

  In the video conference apparatus 200 according to the second embodiment, even if there are a plurality of candidate participants identified by the first identification and the second identification, a correct name or the like is selected or input to the conference participant. As a result, even when the identification accuracy of the identification unit 10 is low or incorrect, the attribute information can be appropriately transmitted without imposing a burden on the user.

  The correspondence between the attribute information and the position information may change depending on whether the participant changes the position of the seat during the video conference or changes the position when entering or leaving the room. In such a case, if the database storage unit 16 is used as it is, incorrect attribute information is transmitted. In the third embodiment, a video conference apparatus capable of transmitting appropriate attribute information even when a participant changes the position of a seat during a video conference or changes the room in an entrance / exit will be described. The identification unit 20 of the video conference apparatus 300 according to the third embodiment includes a determination unit 20. In the following description, a case where the participant B who has once uttered is replaced with a new participant I will be described.

  When Participant I sitting in Participant B's seat speaks, photographing device 106 photographs Participant I, and position information detection unit 8 obtains position information of Participant I (Participant B's seat). To detect. And the feature-value production | generation part 6 produces | generates the participant's I feature-value. Hereinafter, the shooting of the participant I, the detection of the position information, and the generation of the feature amount are referred to as the current shooting, the detection of the current position information, and the generation of the feature amount, respectively.

  Here, the determination unit 20 obtains the similarity between the registered feature value corresponding to the position information detected this time and the feature value generated this time. The determination unit 20 determines whether or not the similarity is smaller than a predetermined value g (second threshold). When the determination unit 20 determines that the similarity is smaller than the predetermined value g, the feature amount when the position information is detected and registered (that is, the feature amount generated at the first shooting) and the current generation This means that the feature amount is greatly different and the participants have been replaced. In this case, the determination unit 20 extracts attribute information corresponding to a feature amount close to the feature amount generated this time (a feature amount having a large similarity with the feature amount generated this time). Then, in addition to the video imaged this time (that is, the video image of participant I) and attribute information, error information f is also transmitted in association with it.

  Here, the error information f is information transmitted when the similarity between the feature amount of the uttered participant and the registered feature amount corresponding to the location information of the uttered participant is greater than the predetermined value g. is there. FIG. 13 shows an example of a video associated with the error information f. The error information in the example of FIG. 13 is “New participant”. The error information in the example of FIG. 13 is error information when the participant B changes to a new participant I. However, the participant B, the participant F who originally participated in the video conference, and the seat If changed, “participant F has changed seat with participant B” may be transmitted as error information f. The error information is not limited to these.

In addition, when a facial feature amount and a speech feature amount are used as the feature amount, the facial feature amount or the speech feature amount of the participant who spoke, the facial feature amount and the speech feature registered in the database storage unit 16 The error information may be transmitted when the difference in amount is larger than either of the predetermined values g, and the error information may be transmitted when both are larger than the predetermined value g.
In the case of the video conference apparatus 300 according to the third embodiment, even if a participant is switched during a video conference or a new participant has entered, it has been switched or newly entered without transmitting attribute information by mistake. By transmitting the error information indicating “”, it is possible to notify the other party of the video conference appropriately that the participant has been replaced or newly entered together with the attribute information.
The video conference apparatus 300 according to the third embodiment is preferably used in a video conference in which it is known in advance that the replacement of participants and the participation of new participants are frequent.
In addition, the feature value generation unit 6 of the video conference apparatus 300 generates a feature value each time the participant speaks. Accordingly, in order to reduce the frequency of feature quantity generation, when a participant is changed in the middle or a new participant is about to enter, another participant at the L point side can start from the video conference apparatus 100 (or 200). It is preferable to switch to the mode of the video conference apparatus 300. This switching may be input from an input unit (not shown).
In the above example, one video conference device is provided at each of the L point and the M point. However, as another example, only one TV conference device is provided at one of the points L and M (or another point connected to the points L and M via the network), and the TV The conference apparatus may register the face feature data and title / name data for the participants scheduled at both the L point and the M point in the database storage unit, and execute the video conference for both participants.
In the above example, the database storage unit 16 is held in the video conference apparatus. However, the database storage unit 16 may be integrated with the hard disk 108 or the memory 110.

In addition, as attribute information, not only the title and name, but also the data that summarizes the claims of the prospective participants at the past meeting (such as views of approval or disagreement for a certain project) are registered in this database storage unit, It may be displayed on the video output unit 103.
Moreover, in the above example, this invention is applied to the video conference system which connects two points, L point and M point. However, the present invention is not limited to this, and the present invention may be applied to a video conference system that connects three or more points, and an appropriate two-way communication system other than the video conference system.

In addition, when the present invention is applied to an entertainment interactive communication system, for example, image data of a favorite animation of a prospective participant is registered in the database storage unit 16 as attribute information to be displayed on the video output unit 103. The animation image may be displayed near the participant's face, or the animation image may be displayed on the participant's face displayed on the video output unit 103.
Further, when the present invention is applied to an interactive communication system in which there is a person who is not desirable to display a face on the video output unit 103 among prospective participants, a mosaic is applied as attribute information about the person. May be registered in the database storage unit 16 so that the face of the person displayed on the video output unit 103 is mosaiced.
In addition, the present embodiment is not limited to the above example, and it is needless to say that various other configurations can be taken without departing from the gist of the present invention.

  The video conference apparatus according to the present embodiment described above can be realized by causing a computer to decode the video conference program. The video conference program proposed in this embodiment is written in a computer-readable program language, recorded on a recording medium such as a magnetic disk or a CD-ROM, and installed in the computer from these recording media, or a communication line. Installed in the computer, and decrypted by a CPU provided in the computer to function as a video conference device. Specifically, the video conference apparatus 100 shown in FIG. 1 may be replaced with a CPU, and the CPU may decode the video conference program.

1000 video conference system 100 video conference device 102 operation unit 103 video output unit 104 audio output unit 106 imaging device 108 hard disk 110 memory 112 communication control unit 100 video conference device 2 clipping unit 6 feature quantity generation unit 8 position information detection unit 10 identification unit 12 synthesizing unit 14 encoding unit 16 database storage unit 18 control unit 20 determination unit 22 display unit 152 face identification unit 154 voice identification unit 156 position information identification unit

JP-A-6-121310 Japanese Patent No. 4055539

Claims (7)

A database storage unit in which a feature amount for identifying a prospective participant and attribute information indicating an attribute of the prospective participant are registered in association with each other;
A feature value generating unit that generates a feature value of a participant photographed by the photographing device;
A position information detection unit that detects the position information of the photographed participant and registers the position information in association with the attribute information of the photographed participant in the database storage unit;
The participant is identified based on the feature amount of the photographed participant and the feature amount in the database storage unit, and the location information of the participant photographed for the second time and later and the location information in the database storage unit An identification unit for identifying the participant based on
The attribute information of the participant identified by the feature amount and the captured video of the participant are transmitted in association with each other,
A video conference device comprising: a transmission unit that associates and transmits the attribute information of the participant identified by the position information and the video of the participant captured after the second time.   If the similarity between the feature quantity in the database storage unit corresponding to the location information of the participant photographed by the photographing device and the feature quantity of the photographed participant is smaller than a predetermined value, the photographing The video conference apparatus according to claim 1, further comprising: a determination unit that causes the transmission unit to transmit the video of the participant and the error information.   3. The display device according to claim 1, further comprising: a display unit that displays all the candidate candidates when there are a plurality of candidate candidates to be identified as a result of identification by the identification unit. Video conferencing equipment.   The video conference apparatus according to claim 1, wherein the feature amount is at least one of a face feature amount and an audio feature amount. The photographing device is capable of omnidirectional photographing,
The video conference apparatus according to claim 1, wherein the participant is located around the photographing apparatus. A generation step of generating a feature amount of the participant photographed by the photographing device;
The feature amount for identifying the prospective participant and the attribute information indicating the attribute of the prospective participant are registered in association with each other, and the feature amount is identified based on the generated feature amount. A first transmission step of transmitting the captured video of the participant and the attribute information corresponding to the feature amount in the database storage unit in association with each other;
A registration step of registering in the database storage unit in association with the position information of the photographed participant and the attribute information of the photographed participant;
A detection step of detecting the location information of the participants taken after the second time;
After the detection step, the position information in the database storage unit, the video shot after the second time of the participant identified based on the detected position information, and the position information in the database storage unit, And a second transmission step of transmitting the corresponding attribute information in association with each other.   The program for functioning a computer as a video conference apparatus in any one of Claims 1-5.

Images