WO2019139101A1 - 情報処理装置、情報処理方法およびプログラム - Google Patents
情報処理装置、情報処理方法およびプログラム Download PDFInfo
- Publication number
- WO2019139101A1 WO2019139101A1 PCT/JP2019/000597 JP2019000597W WO2019139101A1 WO 2019139101 A1 WO2019139101 A1 WO 2019139101A1 JP 2019000597 W JP2019000597 W JP 2019000597W WO 2019139101 A1 WO2019139101 A1 WO 2019139101A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- participant
- conversation
- participants
- image
- information processing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/24—Speech recognition using non-acoustical features
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/35—Categorising the entire scene, e.g. birthday party or wedding scene
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/19—Sensors therefor
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/10—Speech classification or search using distance or distortion measures between unknown speech and reference templates
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/20—Speech recognition techniques specially adapted for robustness in adverse environments, e.g. in noise, of stress induced speech
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/226—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics
- G10L2015/227—Procedures used during a speech recognition process, e.g. man-machine dialogue using non-speech characteristics of the speaker; Human-factor methodology
Definitions
- the present technology relates to an information processing apparatus, an information processing method, and a program, and more specifically, an information processing apparatus for grasping the status of a conversation (a discussion, a meeting, a group work, an active learning, etc.) configured by a plurality of participants. Etc.
- Patent Document 1 proposes a technique for recording and displaying the utterance time of a participant of a conversation.
- Patent Document 2 proposes a technique for presenting a correlation diagram by grasping the state of conversation from voice data or the like obtained by name tag type sensors worn by a plurality of participants. These techniques were able to take action such as nodding by voice and acceleration sensor but were insufficient.
- an information processing apparatus includes: a determination unit that determines a conversation state based on the utterances of the plurality of participants and the states of the plurality of participants.
- the speech processing unit detects the speech of a plurality of participants in a speech from the speech signal. Further, the states of a plurality of participants in a conversation are detected from the image signal. Then, the determination unit determines the state of the conversation based on the utterances of the plurality of participants and the states of the plurality of participants.
- the state of the conversation may be made to be a conversational relationship between participants in the conversation.
- the state of the conversation may include at least any of the content of the speech, the number of speeches, or the speech time.
- the state of the conversation may be made to be the establishment of the conversation. In this case, for example, when the participant looks at the other party of the conversation and speaks, it may be determined that the conversation is established.
- the image processing unit may be configured to detect the direction viewed as the state of the participant. Then, in this case, for example, the image processing unit may be configured to detect the direction viewed from the direction or line of sight of the participant's face.
- the second participant may be determined to have a conversation between the second participant and the second participant.
- the state of conversation is determined based on the states of a plurality of participants in a conversation detected from an audio signal and the plurality of participants in a conversation detected from an image signal. Therefore, it becomes possible to determine the state of conversation more accurately than in the case of determining the state of conversation only by the participant's speech.
- an image generation unit that generates a planar map image of a space used for conversation is provided, and in the planar map image, a participant indicating each participant participating in the conversation There may be a display, and in this planar map image, there may be a conversation line connecting the corresponding two participant displays based on the detected conversation. This makes it possible, for example, to easily visually recognize who has a conversation with who.
- the image processing unit processes the image signal to further detect the position of each participant participating in the conversation, and the image generation unit displays the position of the participant display indicating each participant in the planar map image May be changed according to the position of each of the detected participants. This makes it easy to visually recognize at which position each participant participating in the conversation is located.
- the image generation unit may change the position of the participant display indicating each participant in the planar map image according to the number of conversations of each participant.
- the image generation unit may arrange the positions of the participant displays indicating the respective participants closer to the center of the plane map image as the number of conversations increases. This makes it easy to visually recognize who has a large number of conversations.
- the image generation unit may set the interval between the display of the participants of the two participants closer as the number of conversations between the two participants increases. This makes it easy to visually recognize who has a large number of conversations with each other.
- the participant display may be sized according to the number of utterances of the corresponding participant.
- the participant display may be accompanied by a character indicating the number of utterances of the corresponding participant. This makes it possible to visually and easily recognize the number of utterances of each participant.
- the conversation line may be made to have a thickness corresponding to the number of conversations between corresponding participants.
- the conversation line may be accompanied by characters indicating the number of conversations between corresponding participants. As a result, it is possible to visually and easily recognize the number of conversations between each two participants.
- the display unit may further include a display unit that displays the planar map image generated by the image generation unit. This makes it possible to check the planar map image on the display unit.
- the server may further include a server that holds the planar map image generated by the image generation unit. As a result, it is possible to reproduce and check the past plan map image from the server.
- the communication device may further include a communication unit that transmits the planar map image generated by the image generation unit to the mobile terminal. This makes it possible to check the planar map image in the mobile terminal.
- the conversation situation can be grasped in more detail.
- the effect described here is not necessarily limited, and may be any effect described in the present disclosure.
- FIG. 1 shows an example of the configuration of an information processing apparatus 100 according to the first embodiment.
- the information processing apparatus 100 includes an audio processing unit 101, an image processing unit 102, and a determination unit 103.
- the audio processing unit 101 is supplied with an audio signal in which the microphone detects the audio of each participant participating in a conversation (discussion, meeting, group work, active learning, etc.).
- voice detection may be performed using a microphone attached to each participant, or voice detection may be performed using a microphone array configured of a plurality of microphones.
- the audio processing unit 101 detects the utterances of a plurality of participants participating in a conversation from the audio signal.
- the volume is measured from the audio signal for each participant, and when the state above the threshold continues for a predetermined time or more, it is determined to be a speech. If the state above the threshold value does not satisfy the predetermined time, it is not determined that the speech is an utterance relating to consent, denial, false alarm or the like.
- the image processing unit 102 is supplied with an image signal obtained by capturing an image of a participant participating in a conversation with a camera.
- a camera capable of capturing the entire circumference such as a 360-degree camera may be disposed at the center of all participants and each participant may be imaged, or a plurality of cameras may be disposed so as to cover the entire circumference. And each participant may be imaged.
- the image processing unit 102 detects the states of a plurality of participants participating in the conversation from the image signal.
- the state of the participant detected in this manner is the state required by the participant's speech to determine the state and presence of the conversation.
- the state of the participant includes the direction in which the participant is looking.
- the image processing unit 102 detects the direction of the face of the participant or detects the line of sight direction of the participant by image processing in order to detect the direction in which the participant is looking.
- the determination unit 103 determines the state and the presence of the conversation based on the utterances of the plurality of participants detected by the audio processing unit 101 and the states of the plurality of participants detected by the image processing unit 102.
- the state of the conversation is the conversational relationship between the participants in the conversation.
- the state of the conversation includes at least either the content of the speech, the number of speeches or the speech time.
- the state of conversation is establishment of conversation.
- the determination unit determines that the conversation is established when the participant looks at the other party of the conversation and speaks.
- the determination unit has a second participant's speech and the second participant sees the first direction. It is determined that there is a conversation between the first participant and the second participant when in the direction of the participant. Note that even when the direction in which the second participant is looking is not the direction of the first participant, it may be determined that there is a conversation. For example, when the second participant speaks after the first participant speaks, or after the first participant speaks, the second participant's face moves up and down (Aichichi) It is possible to determine that there is a conversation even when, for example.
- a voice signal formed by detecting the voice of each participant participating in the conversation by the microphone is supplied to the voice processing unit 101.
- the audio processing unit 101 processes the audio signal to detect the speech of each participant participating in the conversation.
- an image signal obtained by capturing an image of each participant participating in a conversation with a camera is supplied to the image processing unit 102.
- the image processing unit 102 processes the image signal to detect the state of each participant participating in the conversation, for example, the direction in which it is viewed.
- the speech of each participant participating in the conversation detected by the voice processing unit 101 and the state of each participant participating in the conversation detected by the image processing unit 102 are supplied to the determination unit 103.
- the determination unit 103 determines the state and presence of the conversation based on the speech of each participant and the state of each participant, and obtains determination information.
- the state and the presence of the conversation are determined based on the speech of each participant participating in the conversation and the state detected by the image processing. Therefore, the state and the presence of the conversation can be determined more accurately than in the case where the state and the presence of the conversation are determined only by the participant's speech.
- FIG. 2 shows an example of the configuration of an information processing apparatus 200 according to the second embodiment.
- the information processing apparatus 200 includes N (N is an integer) cameras 201-1 to 201 -N, M (M is an integer) microphones 202-1 to 202 -M, and an image / sound control system 203.
- a monitor 204, a server 205, and a mobile terminal 206 are included.
- the cameras 201-1 to 201-N are for taking images of participants participating in a conversation.
- the cameras 201-1 to 201-N are installed to image participants participating in a conversation, particularly faces.
- a camera capable of capturing the entire circumference such as a 360 ° camera, may be installed at the center of the participant, or a plurality of cameras may be installed to capture the entire circumference.
- FIG. 3 shows an example of a plane map of a room in which a conversation takes place.
- the circled display with “A” to “D” indicates the participant display showing the participants participating in the conversation, and each participant is located in the center of the room. I'm around.
- a 360 ° camera capable of imaging the entire circumference is mounted on a desk. In this case, the orientation and position of the camera in the planar map are adjusted to be constant.
- FIG. 4 shows another example of a plane map of a room in which a conversation takes place.
- the circular display with “A” to “D” indicates the participant display of the conversation, and each participant is around the desk placed at the center of the room.
- a camera 1 which is a 360 ° camera capable of imaging the entire circumference is installed on a desk, and a camera 2 is installed so as to capture a whiteboard present around the room.
- a whiteboard can be imaged, and when the participant who has written on the whiteboard explains the content to be written in front of the whiteboard, the face of the participant can also be imaged.
- the camera 1 and the camera 2 are installed, the position and the orientation of the camera in the plane map are adjusted.
- the microphones 202-1 to 202-M are for detecting the voices of the participants participating in the conversation.
- the microphones 202-1 to 202-M may be microphones attached to each of the participants, or may be a plurality of microphones constituting a microphone array for detecting the voice of each participant .
- the image / sound control system 203 processes the image signals obtained by the cameras 201-1 to 201 -N and the sound signals obtained by the microphones 202-1 to 202 -M, and a plane map image of a room in which a conversation is conducted. Generate The image voice control system 203 supplies the generated planar map image to the monitor 204, and displays the planar map image on the monitor 204.
- the image voice control system 203 supplies the generated planar map image to the mobile terminal 206 instead of the monitor 204 or together with the monitor 204 according to the user operation, and the planar map image to the mobile terminal 206. indicate. Also, the image voice control system 203 supplies the generated planar map image to the server 205 and records it, and also reproduces the planar map image from the server 205 according to the user's operation to monitor 204 or the mobile terminal 206. To display the reproduced planar map image.
- the image / sound control system 203 includes a control unit 231, a user operation unit 232, an audio processing unit 233, an image processing unit 234, a determination unit 235, an image generation unit 236, and a communication unit 237.
- the control unit 231 controls the operation of each unit of the image and sound control system 203.
- the user operation 232 is, for example, a key, a button, a touch panel, or a remote control for the user to perform various operations.
- the face of each participant may be registered in the image processing unit 234 at the preliminary preparation stage.
- the face of each participant registered may be associated with the number of the microphone attached to each participant. This allows voice detection associated with each participant.
- the face of each participant is captured by a camera.
- the size of the participant's face and the distance from the camera to the participant are associated and measured. This makes it possible to estimate the distance from the camera to the participant based on the size of the face of the participant captured by the camera.
- the image processing unit 234 performs face detection processing on the image signal (captured image) to determine the position and state of the face of each participant, for example, the direction in which it is viewed.
- the direction in which the user is looking is determined by determining the direction of the face or the viewing direction.
- the technique for determining the position and the line of sight of the face is described in, for example, "Japanese Patent Application Laid-Open No. 2012-226665", but it goes without saying that other methods may be used.
- the image processing unit 234 performs personal verification by comparing the face detected in the face detection process with the face registered in advance, and each detected face is the face of any participant. Identify if it is.
- the image processing unit 234 determines the position (direction, distance) on the planar map of the predetermined participant based on the position and size of the participant's face within the imaging angle of view.
- the direction of the participant viewed from the camera on the planar map can be determined from the position of the participant's face in the imaging angle of view.
- the ratio of the size of the face at this registration to the size of the current face You can determine the distance from the participant to the
- the participant does not necessarily face the camera.
- the user may write on a whiteboard during conversation, and the participant's face may not exist in the captured image of the camera. If the face of the participant can not be detected, the place where the face was detected last is fixed as the position (face position) of the participant.
- an object tracking technique may be used to perform object tracking from a place where the face disappears, and the position (face position) of the face participant on the planar map may be updated.
- the audio processing unit 233 processes an audio signal corresponding to each participant, and detects an utterance of each participant participating in the conversation.
- voice detection is performed by the microphone attached to each participant
- the voice processing unit 233 can easily grasp which participant each detected voice belongs to. Thereby, matching between the voice and the participant is performed.
- the voice processing unit 233 can grasp which participant each detected voice belongs to as follows. That is, the microphone array is installed, for example, at the same position as or around a 360 ° camera capable of imaging the entire circumference, and detects a sound and also detects the direction in which the sound is emitted. Therefore, the voice processing unit 233 refers to the position information of the face of each participant detected by the image processing unit 234 as described above, and which participant the voice detected by the microphone array belongs to Understand Thereby, matching between the voice and the participant is performed.
- the technology of detecting the sound direction by the microphone array is described, for example, in "Japanese Patent Application Laid-Open No. 9-251299", but it goes without saying that other methods may be used.
- the audio processing unit 233 processes an audio signal corresponding to each participant, and detects an utterance of each participant participating in the conversation.
- the volume is measured from the audio signal for each participant, and when the state above the threshold continues for a predetermined time or more, it is determined to be a speech. If the state above the threshold value does not satisfy the predetermined time, it is not determined that the speech is an utterance relating to consent, denial, false alarm or the like.
- the determination unit 235 detects a conversation based on the speech of each participant obtained by the audio processing unit 233 and the state of each participant obtained by the image processing unit 234. In this case, for example, after or immediately before the end of the first participant, there is a second participant's utterance and the direction in which the second participant looks is the first participant. It is determined that there is a conversation between the first participant and the second participant.
- the image generation unit 236 generates a plane map image of a space such as a room used for conversation.
- the image generation unit 236 includes in the planar map image, a participant display indicating each participant participating in the conversation detected by the image processing unit 234. Then, the image generation unit 236 dynamically changes the position of the participant display according to the position (face position) of the participant determined by the image processing unit 234. Further, the image generation unit 236 includes a conversation line connecting the two corresponding participant displays in the plane map image based on the conversation determination information of the determination unit 235.
- FIG. 5 shows an example of a planar map image generated by the image generation unit 236.
- the circle display with “A” to “D” indicates a participant display indicating the participants participating in the conversation.
- the display format of the participant display may be a name, a registered face, a color, or a combination of plural symbols other than the symbols as illustrated. Further, the display format may be fixed, or may be set or switched to the user.
- the participant display of “A” to “D” in this example indicates the current position of the participants of “A” to “D”. Also, in this example, there is a conversation line connecting the display of the participants of "B” and “C”, which indicates that there is a conversation between the participants of "B” and “C”. The display position of the conversation line moves along with the movement of the position of the corresponding participant display.
- the speech lines displayed on the planar map image in this manner may be deleted after being displayed for a predetermined time after the speech is detected.
- FIG. 6 shows another example of the planar map image generated by the image generation unit 236.
- the circle display with “A” to “D” indicates the participant display indicating the participants participating in the conversation.
- the number of conversations is accumulated, and each conversation line is accompanied by a letter indicating the number of conversations.
- the count of the number of conversations between participants is incremented each time it is determined that a conversation exists between the participants as described above.
- there are five conversations between the "B" and “C” participants and one conversation between the "A” and “B” participants, and "A” and "B” Two conversations between participants are shown.
- the cumulative number of conversations may be all from the beginning of the conversation, or may be the number of conversations that have been performed within a certain period of time, for example, the past five minutes.
- the number of conversations may be indirectly indicated as the thickness of the conversation line corresponds to the number of conversations. Also, it is conceivable to use a combination of adding a letter indicating the number of conversations to the conversation line and making the thickness of the conversation line according to the number of conversations.
- FIG. 8 shows another example of the planar map image generated by the image generation unit 236.
- characters indicating the number of utterances are attached to the participant display corresponding to each participant.
- participant "A” has 3 utterances
- participant "B” has 6 utterances
- participant "C” has 5 utterances
- participant "A" It is shown that the number of statements of "D" is two.
- the size (size) of the participant display is indirectly indicated as the one corresponding to the number of utterances. May be It is also conceivable to use a combination of a letter indicating the number of utterances in the participant display and a setting of the size of the participant display according to the number of utterances. It is also conceivable to put a letter indicating the speaking time instead of the number of speakings on the participant display.
- the image generation unit 236 supplies the generated planar map image to the monitor 204 and causes the monitor 204 to display the planar map image.
- the image generation unit 236 supplies the generated planar map image to the server 205 to hold it.
- the image generation unit 236 supplies the generated planar map image to the mobile terminal 206 through the communication unit 237 instead of the monitor 204 or together with the monitor 204 according to the user's operation, and the plane is transmitted to the mobile terminal 206. Display map image.
- the image generation unit 236 reproduces the planar map image from the server 205 and supplies the planar map image to the monitor 204 or the mobile terminal 206 according to the user's operation, and displays the reproduced planar map image. It is also conceivable to instruct the reproduction of the planar map image from the server 205 from the mobile terminal 206 via the communication unit 237.
- the cameras 201-1 to 201-N are installed so that all the people participating in the conversation, in particular the face, are imaged.
- a camera capable of imaging the entire circumference such as a 360 ° camera may be installed at the center of the participant, or a plurality of cameras may be installed so as to enable imaging of the entire circumference.
- the orientation and position of the camera in the planar map are adjusted to be constant.
- microphones 202-1 to 202-M are installed as preparation in advance.
- a microphone may be attached to all people participating in a conversation, or a microphone array may be placed in the center of a participant.
- the microphones are associated with the participants.
- the face of each participant is taken by a camera and registered as preparation in advance.
- the distance between the participant and the camera is imaged at a constant state, and the size is also measured and held together with the face.
- the information on the distance in this case may be held together.
- the participant is registered for the first time, it becomes a new registration, but when there are already registered participants, it is sufficient to additionally register the remaining participants.
- step ST1 captured images including the participants participating in the conversation are obtained using the cameras 201-1 to 201-N. Then, in step ST2, the image processing unit 234 processes the captured image signal to determine the position and state of each participant's face, for example, the direction in which it is viewed (the direction or line of sight).
- step ST3 the microphones 202-1 to 202-M are used to detect the voices of the participants participating in the conversation.
- step ST4 the audio processing unit 233 processes the audio signal to detect the speech of each participant participating in the conversation.
- step ST5 the determination unit 235 determines based on the utterances of all or part of the participants obtained by the audio processing unit 233 and the state of all or part of the participants obtained by the image processing unit 234. Detect conversations. In this case, for example, after or immediately before the end of the first participant, there is a second participant's utterance and the direction in which the second participant looks is the first participant. It is determined that there is a conversation between the first participant and the second participant. Also, for example, when the second participant speaks after the first participant speaks, or after the first participant speaks, the face of the second participant moves up and down (aichichi) ", It is determined that there is a conversation between the first participant and the second participant.
- the image generation unit 236 generates a plane map image of a space such as a room used for conversation.
- the planar map image includes a participant display indicating each participant participating in the conversation detected by the image processing unit 234, and the position of the participant display is determined by the image processing unit 234. It is dynamically changed according to the position (face position) of the participant.
- the planar map image includes a conversation line connecting the corresponding two participant displays based on the conversation determination information of the determination unit 235, and the number of utterances of each participant and the conversation between the two participants Indications such as numbers are also included.
- step ST7 the image generation unit 236 outputs the created planar map image.
- the planar map image is supplied to the monitor 204, and the planar image is displayed on the monitor 204.
- the planar map image is supplied to the server 205, and the planar map image sequentially generated by the server 205 is accumulated and held.
- the planar map image is transmitted to the mobile terminal 206 through the communication unit 237, and this planar display is displayed on the mobile terminal 206.
- the determination unit 235 determines the state and presence of conversation based on the speech of each participant participating in the conversation and the state detected by the image processing. It is a thing. Therefore, it becomes possible to determine the state and the presence of the conversation more accurately than in the case where the state and the presence of the conversation are determined only by the speech of the participants.
- the image processing unit 234 processes the image signal to further detect the position of each participant participating in the conversation
- the image generation unit 236 is a plane map image The position of the participant display indicating each participant in the is changed according to the position of each detected participant. Therefore, it is possible to visually recognize easily where each participant participating in the conversation is located.
- the participant display present in the planar map image generated by the image generation unit 236 has a size corresponding to the number of utterances of the corresponding participant, or the participant display A letter accompanied by the number of utterances of the corresponding participant is attached to. Therefore, it is possible to easily visually recognize the number of utterances of each participant.
- a conversation line present in the planar map image generated by the image generation unit 236 has a thickness corresponding to the number of conversations between corresponding participants, or Is accompanied by a letter indicating the number of conversations between corresponding participants. Therefore, it is possible to easily visually recognize the number of conversations between each two participants.
- the planar map image generated by the image generation unit 236 is supplied to the monitor 204 as a display unit, and the planar map image is displayed on the monitor 204. Therefore, for example, a facilitator of a conversation can confirm a planar map image having various information on speech and conversation on the monitor 204.
- the planar map image generated by the image generation unit 236 is supplied to the server 205 and held. Therefore, the planar map image of the past can be reproduced from the server 205, supplied to, for example, the monitor 204 and displayed, and the planar map image of the past can be easily confirmed.
- the planar map image read by the image generation unit 236 or the past planar map image read from the server 205 is supplied to the mobile terminal 206 through the communication unit 237 as necessary.
- the current or past planar map image is displayed on the mobile terminal 206. Therefore, for example, a facilitator of a conversation can confirm a planar map image having various information on speech and conversation on the mobile terminal 206.
- the face of the participant is registered in advance, and the personal verification is performed by comparing the face detected from the image and the registered face.
- this can not provide an appropriate display when an unregistered person participates. Therefore, when individual verification fails, it is also conceivable to newly register the face as a participant.
- the accurate distance can not be known unless a means capable of measuring the distance such as a stereo camera is used. Therefore, the distance from the camera of the new participant is fixed, or the distance to the camera is simply calculated using an average face size.
- the position, the number of utterances, and the number of conversations of the participants are displayed for each group.
- the determination of whether or not the members of the group have been replaced may be replaced at the moment when a face appears on the camera for capturing an image of another group, or after a set time has elapsed since the replacement. You may replace it later. In this case, the time when the face appears first on the camera is also recorded, and it is determined that the time has been set in advance and it has been replaced after a lapse of a predetermined time, and the number of conversations and the like are displayed.
- the conversation line is displayed in the same manner as described above, but the conversation line regarding the conversation between participants in another space is displayed across the plane maps.
- the example of FIG. 12 has shown the example of 2 rooms, the same may be said of the case of 3 or more rooms.
- the speech contents may be converted into sentences and displayed by speech recognition from the acquired speech. In this case, it may be considered to extract and display only the keyword, or to display the full text of the utterance content.
- the participant was displayed on the position on the plane map observed by the camera image of the participant of conversation.
- the participant display which shows each participant according to the number of conversations of each participant. For example, a person with a large number of conversations is placed closer to the center of the screen, and a person with a smaller number of conversations is placed outside the screen.
- when there are many conversations between two participants they are arranged near each other, and when the number of conversations between two participants is small, they are arranged far.
- the coulomb force F is expressed by the following equation (1).
- k is a proportionality constant
- m1 and m2 are magnetic amounts of the objects 1 and 2 respectively
- r is a distance between the two objects.
- F km 1 m 2 / r 2 (1)
- the virtual object 0 is arranged as the magnetic amount m0 at the center of the plane map, and when there are four participants, the magnetic amounts m1 to m4 proportional to the number of conversations are given to the objects 1 to 4, respectively.
- the amount of magnetism increases as the number of conversations increases.
- m0 and m1 to m4 attract each other by the Coulomb force calculated by equation (1).
- the calculated Coulomb forces repel each other with a force inversely proportional to the number of conversations.
- the attractive force on m0 tends to center all participants in position. On the other hand, they try to leave each other by opposing each other with a force that is inversely proportional to the number of conversations. The position of the participant is stopped and determined, with a balance between attractive and repulsive forces being maintained. When the number of conversations changes, the display position on the planar map of the participants also changes.
- FIG. 14A shows a state in which the number of conversations of the participant of “A” is small and the number of conversations of the remaining 3 people is large, and it can be easily visually recognized from this state.
- FIG. 14 (b) shows a state in which the number of conversations of the participants of “A” is large and plays a hub role of conversation, and it is easily recognized visually that this state is from now on it can.
- each participant display indicating each participant is determined according to the number of conversations of each participant as described above, adding a character indicating the number of conversations to the conversation line, or May be made according to the number of conversations, and each participant display may be accompanied by a character indicating the number of utterances of each participant, or the size of each participant display may be the number of utterances of each participant And the size according to the speech time (see FIGS. 6 to 9).
- FIGS. 15 (a) and 15 (b) show the case where the size of the participant display of each participant is changed in the speech time as shown in FIG. 14 (b).
- FIG. 15 (a) it is possible to confirm that the speaking time of the participant of "A” is long, for example, a state of being silently speaking like a German concert.
- FIG. 15 (b) the participant of "A” has many conversations, but the speaking time is not so long, that is, it seems to be playing a role as a chairperson or facilitator of the conversation group Can be confirmed.
- the state of conversation can be objectively grasped by arranging the participant display indicating each participant at a position according to the number of conversations.
- a viewer such as a facilitator may switch the display method of a plurality of plane maps such as a plane map based on a position observed by a camera or a plane map based on a position according to the number of conversations.
- the evaluation rank of the conversation is displayed for each conversation group.
- the evaluation ranks are displayed in three stages of A, B, and C from the top. For example, in the group displayed in the upper right, it is evaluated as "overall evaluation A".
- the evaluation for each group is evaluated by, for example, the number of utterances or the number of conversations.
- the evaluation value Et of each group is obtained using the following formula (2).
- Ngs is the number of utterances
- Ngc is the number of conversations
- ⁇ g and ⁇ g are coefficients.
- An evaluation rank can be obtained by processing the evaluation value calculated by Equation (2) with a predetermined threshold value. In addition, you may display an evaluation value as it is, without converting into an evaluation rank.
- Et ⁇ g ⁇ Ngs + ⁇ g ⁇ Ngc (2)
- the evaluation value Em of each participant can be similarly obtained by the number of utterances and the number of conversations using the following equation (3).
- Nms is the number of utterances
- Nmc is the number of conversations
- ⁇ m and ⁇ m are coefficients.
- An evaluation rank can be obtained by processing the evaluation value calculated by Equation (3) with a predetermined threshold value. In addition, you may display an evaluation value as it is, without converting into an evaluation rank.
- Em ⁇ m ⁇ N ms + ⁇ m ⁇ N mc (3)
- the evaluation value of each participant since the evaluation is low if only some people are excited, the evaluation value of each participant may be added to the evaluation.
- the evaluation value Et ′ of each group is determined based on the following formula (4).
- min (Em) represents the lowest evaluation value of all the participants participating in the conversation group, and ⁇ g is a coefficient.
- Et ′ ⁇ g ⁇ Ngs + ⁇ g ⁇ Ngc + ⁇ g ⁇ min (Em) (4)
- the number of utterances is used in the equations (2) to (4), the number of utterances may be changed to the utterance time, or both of them may be used.
- information of scores or grades automatically assigned to each student or each group can be considered.
- the score or the score for example, the cumulative or instantaneous one is appropriately displayed according to the operation of the teacher.
- information to assist students can be considered. For example, a summary of the previous conversations is displayed in order to participate in the next conversation. Also, as information presented to students, information for warning students can be considered. For example, the number of utterances and the amount of utterances contributing to the overall flow are displayed.
- information such as speech or conversation is used as information for planning the situation of conversation, but information on utterances related to consent, denial, false alarm etc. is also used for conversation situation It is also conceivable to use it as information for planning. For example, the number of occurrences of consent, denial, false alarm etc. may be counted, and the number of utterances as well as the number of conversations and the number of utterances may be displayed on the planar map image.
- a microphone and a camera were separate bodies
- these are united.
- a microphone is integrated in a 360-degree camera
- a microphone is integrated in each of a plurality of cameras covering the entire periphery
- the state of the detected participant is not limited to this.
- the state of the participant includes facial expressions of the participant (joy, anger, hate, pleasure, surprise, etc.), gestures (movement of fingers, hands, arms), movement of the face (upper or lower nodding (agreement, It may be made to detect a swing in the horizontal direction (negative), biological information (temperature change or pulse change), and the like.
- people who are nodding are likely to be actively participating in the conversation, and such information is counted and displayed along with the number of conversations and the number of utterances. It may be done.
- FIGS. 18 to 23 show examples of screen display on the monitor 204 or the mobile terminal 206.
- the state of conversation in pods (conversation group) A to D is shown in the screen display example of FIG.
- the pod identifiers of POD A to POD D are displayed corresponding to the display of each pod. Further, in each pod display, as shown by an arrow P1, the participant is shown by an image including a face.
- the timeline of the discussion is displayed in the area indicated by the arrow P2.
- a thumbnail image is disposed at the beginning of each utterance as indicated by an arrow P3 so as to indicate who the utterance is, and a name character is also disposed as indicated by an arrow P4. It may be either one of the thumbnail image and the name character.
- the display of the area where the timeline of this discussion is displayed is automatically updated so that a new message is displayed at the latest position every time there is a message, but the user performs a scroll operation on the screen. By doing this, it is possible to confirm past remarks.
- the screen display example of FIG. 19 shows the state of conversation in pods (conversation groups) A to D, as in the display example of FIG.
- the pod identifiers of POD A to POD D are displayed corresponding to the display of each pod. Further, in each pod display, as shown by an arrow P1, the participant is shown by an image including a face. Further, in each pod display, the timeline of the discussion is displayed in the area indicated by the arrow P2.
- the color of the image of each participant dynamically changes according to the number of utterances in the participant display indicated by the arrow P1. For example, a participant with a large number of utterances is yellow, a participant with a medium number of utterances is red, and a participant with a small number of utterances is green.
- the difference in color is represented by the difference in pattern superimposed on the image.
- the pattern to be superimposed on the image of each participant may be dynamically changed according to the number of utterances.
- not only the number of utterances but also the number of consenting and nodding utterances and the number of conversations can be similarly expressed by colors and patterns. As a result, for example, the teacher who is the user can easily visually confirm the number of utterances of each participant in each pod, etc., and it becomes easy to intervene in the pod which has felt that the discussion is stuck.
- the screen display example of FIG. 20 shows the state of conversation in pods (conversation groups) A to D, as in the display example of FIG.
- the pod identifiers of POD A to POD D are displayed corresponding to the display of each pod.
- the timeline of the discussion is displayed in the area indicated by the arrow P2.
- the area indicated by the arrow P1 can be switched to the shared screen in the pod in addition to the state used for the participant display by the face image or the like.
- the pods A and B are switched to the shared screen. For example, a teacher who is a user can operate such screen switching to know what material or description is being discussed in each pod.
- the speech of the conversation participant displayed in the area indicated by the arrow P2 is accumulated in the server 205, and for example, the timeline is sent to the mobile terminal 206 of each student Come on.
- the time of each utterance is associated with the time axis of the camera image and held.
- the screen display example of FIG. 21 shows the state of conversation by four participants.
- the participant is indicated by an image including a face.
- a timeline of the discussion is displayed in the area indicated by the arrow Q2.
- name characters are arranged at the beginning of each statement as indicated by an arrow Q4 so that it can be understood who is speaking. It may be a thumbnail image instead of the name letter.
- the slide and the board used in the discussion are displayed. In the illustrated example, among the four participants, the participant corresponding to the image displayed in the upper right of the area indicated by the arrow Q1 is facing backward for the board at this point in time.
- the image can not be captured, and tracking of the person is performed and the image is cropped. It is also possible to associate the participant with the speech by making the frame color surrounding the image of each participant and the color of the character part of the first name of each speech the same color. In the illustrated example, the line type is changed instead of the frame color.
- the image display example of FIG. 22 shows the state of conversation by four participants.
- the face image of the four participants is displayed at the four corners of the rectangular frame.
- circles corresponding to each participant are displayed, and these circles are made to move diagonally.
- the circle and the color of the frame of the image of the corresponding participant are made the same color.
- the line type is changed instead of the frame color.
- the circle of each participant increases as the number of utterances increases, and the distance from the center of the rectangular frame decreases as the number of conversations increases.
- the participant displayed in the upper right image has a large number of utterances and many conversations
- the participant displayed in the upper left image has a small number of utterances and a small number of conversations I understand.
- the image display example of FIG. 23 shows the state of conversation by four participants.
- a face image showing four participants is displayed.
- the face image of the participant having a large number of utterances is displayed larger.
- a conversation line connecting the participants is displayed, and the larger the number of conversations, the thicker the display.
- the participant shown in the upper left image has a large number of utterances, particularly the number of conversations with the participant shown in the lower right image.
- the participant shown in the upper right image has a medium number of utterances, but the number of conversations with the participant particularly shown in the lower right image is large, and the participant shown in the lower left image It turns out that there is no conversation between them.
- the present technology can also be configured as follows.
- An audio processing unit that detects the speech of a plurality of participants in a conversation from an audio signal;
- An image processing unit that detects states of a plurality of participants in the conversation from an image signal;
- An information processing apparatus comprising: a determination unit that determines a state of conversation based on the utterances of the plurality of participants and the states of the plurality of participants.
- the information processing apparatus according to (1) wherein the state of the conversation is a relationship between the participants of the conversation based on the conversation.
- the state of the conversation at least includes any of the content of the utterance, the number of utterances, and the utterance time.
- the information processing apparatus When it is determined that there is a conversation between the first participant and the second participant, the information processing apparatus according to (6) or (7).
- the voice processing unit measures the volume from the voice signal, and determines that it is a speech when the state above the threshold continues for a certain period of time or more, described in any of (1) to (8). Information processing equipment.
- the information processing apparatus according to (9), wherein the voice processing unit determines that the utterance relates to consent, denial, or false alarm when the state of the threshold or more does not reach a predetermined time.
- the image processing unit processes the image signal to further detect the position of each participant participating in the conversation,
- the image generation unit changes a position of a participant display indicating the respective participant in the planar map image according to the position of the detected participant.
- the information processing apparatus according to (11) .
- the image generation unit changes the position of the participant display indicating the respective participants in the planar map image according to the number of conversations of the respective participants. apparatus.
- the information processing apparatus according to any one of (11) to (17), wherein the conversation line has a thickness corresponding to the number of conversations between corresponding participants. (19) The information processing apparatus according to any one of (11) to (18), wherein a character indicating the number of conversations between corresponding participants is attached to the conversation line. (20) The information processing apparatus according to any one of (11) to (19), further including: a display unit configured to display the planar map image generated by the image generation unit. (21) The information processing apparatus according to any one of (11) to (20), further including: a server configured to hold the planar map image generated by the image generation unit.
- the information processing apparatus according to any one of (11) to (21), further including: a communication unit that transmits the planar map image generated by the image generation unit to the mobile terminal.
- a communication unit that transmits the planar map image generated by the image generation unit to the mobile terminal.
- a voice processing step in which the voice processing unit detects the speech of a plurality of participants in a conversation from the voice signal;
- An image processing step in which the image processing unit detects the states of the plurality of participants in the conversation from the image signal;
- An information processing method comprising a detecting step of determining a state of conversation based on the speech of the plurality of participants and the state of the plurality of participants.
- (24) computer Voice processing means for detecting the utterances of a plurality of participants in a conversation from an audio signal; Image processing means for detecting the states of the plurality of participants in the conversation from the image signal; A program that functions as a determination unit that determines the utterances of the plurality of participants and the states of the plurality of participants.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Acoustics & Sound (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Ophthalmology & Optometry (AREA)
- Telephonic Communication Services (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
- Information Transfer Between Computers (AREA)
Abstract
Description
音声信号から会話の複数の参加者の発言を検出する音声処理部と、
画像信号から上記会話の複数の参加者の状態を検出する画像処理部と、
上記複数の参加者の発言と上記複数の参加者の状態に基づいて会話の状態を判定する判定部を備える
情報処理装置にある。
1.第1の実施の形態
2.第2の実施の形態
3.変形例
[情報処理装置]
図1は、第1の実施の形態としての情報処理装置100の構成例を示している。この情報処理装置100は、音声処理部101と、画像処理部102と、判定部103を有している。
[情報処理装置]
図2は、第2の実施の形態としての情報処理装置200の構成例を示している。この情報処理装置200は、N個(Nは整数)のカメラ201-1~201-Nと、M個(Mは整数)のマイク202-1~202-Mと、画像音声制御システム203と、モニタ204と、サーバ205と、モバイル端末206を有している。
なお、上述実施の形態においては、予め参加者の顔を登録して、画像から検出した顔と登録している顔を比較することで個人照合を行っている。しかしながら、これでは登録していない人が参加した場合に適切な表示をすることができない。そこで、個人照合に失敗した場合は、その顔を新たに参加者として登録することも考えられる。
F = km1m2/r2 ・・・(1)
Et = αg・Ngs + βg・Ngc ・・・(2)
Em = αm・Nms + βm・Nmc ・・・(3)
Et’ = αg・Ngs + βg・Ngc + γg・min(Em) ・・・(4)
(1)音声信号から会話の複数の参加者の発言を検出する音声処理部と、
画像信号から上記会話の複数の参加者の状態を検出する画像処理部と、
上記複数の参加者の発言と上記複数の参加者の状態に基づいて会話の状態を判定する判定部を備える
情報処理装置。
(2)上記会話の状態は、上記会話の参加者間の会話による関係性である
前記(1)に記載の情報処理装置。
(3)上記会話の状態は、少なくとも、発言内容、発言回数または発言時間のいずれかを含む
前記(1)に記載の情報処理装置。
(4)上記会話の状態は、会話の成立である
前記(1)に記載の情報処理装置。
(5)上記判定部は、上記参加者が会話の相手を見て話をしているとき、上記会話が成立したと判定する
前記(4)に記載の情報処理装置。
(6)上記画像処理部は、上記参加者の状態として、見ている方向を検出する
前記(1)から(5)のいずれか記載の情報処理装置。
(7)上記画像処理部は、上記参加者の顔の向きまたは視線方向から該参加者の見ている方向を検出する
前記(6)に記載の情報処理装置。
(8)上記判定部は、第1の参加者の発言の後に第2の参加者の発言があり、かつ上記第2の参加者の見ている方向が上記第1の参加者の方向であるとき、上記第1の参加者と上記第2の参加者との間の会話があると判定する
前記(6)または(7)に記載の情報処理装置。
(9)上記音声処理部は、上記音声信号から音量を計測し、閾値以上の状態が一定時間以上継続した場合には発言であると判定する
前記(1)から(8)のいずれかに記載の情報処理装置。
(10)上記音声処理部は、閾値以上の状態が一定時間に満たない場合には、同意、否定またはあいづちに係る発声であると判定する
前記(9)に記載お情報処理装置。
(11)上記会話に使用されている空間の平面マップ画像を生成する画像生成部をさらに備え、
上記平面マップ画像には、上記会話に参加しているそれぞれの参加者を示す参加者表示が存在し、
上記平面マップ画像には、上記検出された会話に基づいて、対応する2つの参加者表示をつなぐ会話線が存在する
前記(1)から(10)のいずれかに記載の情報処理装置。
(12)上記画像処理部は、上記画像信号を処理して上記会話に参加しているそれぞれの参加者の位置をさらに検出し、
上記画像生成部は、上記平面マップ画像における上記それぞれの参加者を示す参加者表示の位置を、上記検出されたそれぞれの参加者の位置に応じて変更する
前記(11)に記載の情報処理装置。
(13)上記画像生成部は、上記平面マップ画像における上記それぞれの参加者を示す参加者表示の位置を、上記それぞれの参加者の会話数に応じて変更する
前記(11)に記載の情報処理装置。
(14)上記画像生成部は、上記それぞれの参加者を示す参加者表示の位置を、会話数が多いほど上記平面マップ画像の中央寄りに配置する
前記(13)に記載の情報処理装置。
(15)上記画像生成部は、2参加者の参加者表示の間隔を、該2参加者間の会話数が多いほど近くする
前記(13)または(14)に記載の情報処理装置。
(16)上記参加者表示は、対応する参加者の発言数に応じたサイズとされる
前記(11)から(15)のいずれかに記載の情報処理装置。
(17)上記参加者表示には、対応する参加者の発言数を示す文字が添えられる
前記(11)から(16)のいずれかに記載の情報処理装置。
(18)上記会話線は、対応する参加者間の会話数に応じた太さとされる
前記(11)から(17)のいずれかに記載の情報処理装置。
(19)上記会話線には、対応する参加者間の会話数を示す文字が添えられる
前記(11)から(18)のいずれかに記載の情報処理装置。
(20)上記画像生成部で生成された平面マップ画像を表示する表示部をさらに備える
前記(11)から(19)のいずれかに記載の情報処理装置。
(21)上記画像生成部で生成された平面マップ画像を保持するサーバをさらに備える
前記(11)から(20)のいずれかに記載の情報処理装置。
(22)上記画像生成部で生成された平面マップ画像をモバイル端末に送信する通信部をさらに備える
前記(11)から(21)のいずれかに記載の情報処理装置。
(23)音声処理部が、音声信号から会話の複数の参加者の発言を検出する音声処理ステップと、
画像処理部が、画像信号から上記会話の複数の参加者の状態を検出する画像処理ステップと、
会話検出部が、上記複数の参加者の発言と上記複数の参加者の状態に基づいて会話の状態を判定する検出ステップを有する
情報処理方法。
(24)コンピュータを、
音声信号から会話の複数の参加者の発言を検出する音声処理手段と、
画像信号から上記会話の複数の参加者の状態を検出する画像処理手段と、
上記複数の参加者の発言と上記複数の参加者の状態を判定する判定手段として機能させる
プログラム。
101・・・音声処理部
102・・・画像処理部
103・・・判定部
200・・・情報処理装置
201-1~201-N・・・カメラ
202-1~202-M・・・マイク
203・・・画像音声制御システム
204・・・モニタ
205・・・サーバ
206・・・モバイル端末
231・・・制御部
232・・・ユーザ操作部
233・・・音声処理部
234・・・画像処理部
235・・・判定部
236・・・画像生成部
237・・・通信部
Claims (24)
- 音声信号から会話の複数の参加者の発言を検出する音声処理部と、
画像信号から上記会話の複数の参加者の状態を検出する画像処理部と、
上記複数の参加者の発言と上記複数の参加者の状態に基づいて会話の状態を判定する判定部を備える
情報処理装置。 - 上記会話の状態は、上記会話の参加者間の会話による関係性である
請求項1に記載の情報処理装置。 - 上記会話の状態は、少なくとも、発言内容、発言回数または発言時間のいずれかを含む
請求項1に記載の情報処理装置。 - 上記会話の状態は、会話の成立である
請求項1に記載の情報処理装置。 - 上記判定部は、上記参加者が会話の相手を見て話をしているとき、上記会話が成立したと判定する
請求項4に記載の情報処理装置。 - 上記画像処理部は、上記参加者の状態として、見ている方向を検出する
請求項1に記載の情報処理装置。 - 上記画像処理部は、上記参加者の顔の向きまたは視線方向から該参加者の見ている方向を検出する
請求項6に記載の情報処理装置。 - 上記判定部は、第1の参加者の発言の後に第2の参加者の発言があり、かつ上記第2の参加者の見ている方向が上記第1の参加者の方向であるとき、上記第1の参加者と上記第2の参加者との間の会話があると判定する
請求項6に記載の情報処理装置。 - 上記音声処理部は、上記音声信号から音量を計測し、閾値以上の状態が一定時間以上継続した場合には発言であると判定する
請求項1に記載の情報処理装置。 - 上記音声処理部は、閾値以上の状態が一定時間に満たない場合には、同意、否定またはあいづちに係る発声であると判定する
請求項9に記載お情報処理装置。 - 上記会話に使用されている空間の平面マップ画像を生成する画像生成部をさらに備え、
上記平面マップ画像には、上記会話に参加しているそれぞれの参加者を示す参加者表示が存在し、
上記平面マップ画像には、上記検出された会話に基づいて、対応する2つの参加者表示をつなぐ会話線が存在する
請求項1に記載の情報処理装置。 - 上記画像処理部は、上記画像信号を処理して上記会話に参加しているそれぞれの参加者の位置をさらに検出し、
上記画像生成部は、上記平面マップ画像における上記それぞれの参加者を示す参加者表示の位置を、上記検出されたそれぞれの参加者の位置に応じて変更する
請求項11に記載の情報処理装置。 - 上記画像生成部は、上記平面マップ画像における上記それぞれの参加者を示す参加者表示の位置を、上記それぞれの参加者の会話数に応じて変更する
請求項11に記載の情報処理装置。 - 上記画像生成部は、上記それぞれの参加者を示す参加者表示の位置を、会話数が多いほど上記平面マップ画像の中央寄りに配置する
請求項13に記載の情報処理装置。 - 上記画像生成部は、2参加者の参加者表示の間隔を、該2参加者間の会話数が多いほど近くする
請求項13に記載の情報処理装置。 - 上記参加者表示は、対応する参加者の発言数に応じたサイズとされる
請求項11に記載の情報処理装置。 - 上記参加者表示には、対応する参加者の発言数を示す文字が添えられる
請求項11に記載の情報処理装置。 - 上記会話線は、対応する参加者間の会話数に応じた太さとされる
請求項11に記載の情報処理装置。 - 上記会話線には、対応する参加者間の会話数を示す文字が添えられる
請求項11に記載の情報処理装置。 - 上記画像生成部で生成された平面マップ画像を表示する表示部をさらに備える
請求項11に記載の情報処理装置。 - 上記画像生成部で生成された平面マップ画像を保持するサーバをさらに備える
請求項11に記載の情報処理装置。 - 上記画像生成部で生成された平面マップ画像をモバイル端末に送信する通信部をさらに備える
請求項11に記載の情報処理装置。 - 音声処理部が、音声信号から会話の複数の参加者の発言を検出する音声処理ステップと、
画像処理部が、画像信号から上記会話の複数の参加者の状態を検出する画像処理ステップと、
会話検出部が、上記複数の参加者の発言と上記複数の参加者の状態に基づいて会話の状態を判定する検出ステップを有する
情報処理方法。 - コンピュータを、
音声信号から会話の複数の参加者の発言を検出する音声処理手段と、
画像信号から上記会話の複数の参加者の状態を検出する画像処理手段と、
上記複数の参加者の発言と上記複数の参加者の状態を判定する判定手段として機能させる
プログラム。
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2019564747A JP7276158B2 (ja) | 2018-01-12 | 2019-01-10 | 情報処理装置、情報処理方法およびプログラム |
| EP19738367.2A EP3739573B1 (en) | 2018-01-12 | 2019-01-10 | Information processing device, information processing method, and program |
| CN201980007295.6A CN111542876B (zh) | 2018-01-12 | 2019-01-10 | 信息处理装置、信息处理方法以及程序产品 |
| US16/959,884 US11837233B2 (en) | 2018-01-12 | 2019-01-10 | Information processing device to automatically detect a conversation |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2018-003778 | 2018-01-12 | ||
| JP2018003778 | 2018-01-12 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2019139101A1 true WO2019139101A1 (ja) | 2019-07-18 |
Family
ID=67219559
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2019/000597 Ceased WO2019139101A1 (ja) | 2018-01-12 | 2019-01-10 | 情報処理装置、情報処理方法およびプログラム |
Country Status (5)
| Country | Link |
|---|---|
| US (1) | US11837233B2 (ja) |
| EP (1) | EP3739573B1 (ja) |
| JP (1) | JP7276158B2 (ja) |
| CN (1) | CN111542876B (ja) |
| WO (1) | WO2019139101A1 (ja) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021182089A (ja) * | 2020-05-20 | 2021-11-25 | 清水建設株式会社 | 人物関係抽出システム |
| WO2021245759A1 (ja) * | 2020-06-01 | 2021-12-09 | ハイラブル株式会社 | 音声会議装置、音声会議システム及び音声会議方法 |
| JP2022072455A (ja) * | 2020-10-29 | 2022-05-17 | 京セラ株式会社 | 仮想イベントシステム、情報処理装置、出力制御方法及びプログラム |
| JPWO2024100703A1 (ja) * | 2022-11-07 | 2024-05-16 | ||
| KR102772721B1 (ko) * | 2024-04-02 | 2025-02-26 | 케이에듀테크 주식회사 | 온라인 및 오프라인 하이브리드 러닝 참여자 분석 시스템 |
| JP7847733B1 (ja) * | 2025-07-15 | 2026-04-17 | 三菱電機株式会社 | 発言データ提供装置、発言推定システム、発言データ提供システム、発言データ提供方法及びプログラム |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7177348B2 (ja) * | 2019-02-06 | 2022-11-24 | 日本電信電話株式会社 | 音声認識装置、音声認識方法およびプログラム |
| JP2021099670A (ja) * | 2019-12-23 | 2021-07-01 | 富士フイルムビジネスイノベーション株式会社 | 情報処理装置及びプログラム |
| JP7528638B2 (ja) | 2020-08-26 | 2024-08-06 | トヨタ自動車株式会社 | コミュニケーションシステム |
| JP7354992B2 (ja) * | 2020-11-19 | 2023-10-03 | トヨタ自動車株式会社 | 発言評価システム、発言評価方法、及び、プログラム |
| US20250343987A1 (en) * | 2024-01-26 | 2025-11-06 | Huddly As | Systems and methods for framing meeting environments and participants using camera spatial positioning information |
| FR3163196A1 (fr) * | 2024-06-11 | 2025-12-12 | Orange | Traitement de données d’image, notamment de visioconférence, pour déterminer l’interlocuteur d’un locuteur courant |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6280768A (ja) | 1985-10-04 | 1987-04-14 | Hitachi Ltd | ステレオ画像処理方法 |
| JPH04323689A (ja) | 1991-04-24 | 1992-11-12 | Toshiba Corp | 会議進行支援装置 |
| JPH06169458A (ja) | 1992-11-30 | 1994-06-14 | Matsushita Electric Ind Co Ltd | 移動体の自動追尾装置 |
| JPH09251299A (ja) | 1996-03-15 | 1997-09-22 | Toshiba Corp | マイクロホンアレイ入力型音声認識装置及び方法 |
| JP2000184345A (ja) * | 1998-12-14 | 2000-06-30 | Nec Corp | マルチモーダルコミュニケーション支援装置 |
| WO2008139717A1 (ja) * | 2007-05-09 | 2008-11-20 | Panasonic Corporation | 表示装置、表示方法、表示プログラム |
| WO2009004731A1 (ja) * | 2007-07-05 | 2009-01-08 | Panasonic Corporation | 画像音声通信装置及び画像表示方法 |
| JP2010266522A (ja) * | 2009-05-12 | 2010-11-25 | Nippon Telegr & Teleph Corp <Ntt> | 対話状態分割装置とその方法、そのプログラムと記録媒体 |
| JP2011237965A (ja) * | 2010-05-10 | 2011-11-24 | Seiko Epson Corp | 会話比率算出装置、接客データ記録装置、会話比率算出方法およびプログラム |
| JP2012226665A (ja) | 2011-04-21 | 2012-11-15 | Canon Inc | 情報処理装置、その処理方法、プログラム及び撮像装置 |
| JP2013058221A (ja) | 2012-10-18 | 2013-03-28 | Hitachi Ltd | 会議分析システム |
| JP2017009825A (ja) * | 2015-06-23 | 2017-01-12 | トヨタ自動車株式会社 | 会話状況分析装置および会話状況分析方法 |
Family Cites Families (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB2351628B (en) * | 1999-04-14 | 2003-10-01 | Canon Kk | Image and sound processing apparatus |
| US6894714B2 (en) * | 2000-12-05 | 2005-05-17 | Koninklijke Philips Electronics N.V. | Method and apparatus for predicting events in video conferencing and other applications |
| US6782363B2 (en) * | 2001-05-04 | 2004-08-24 | Lucent Technologies Inc. | Method and apparatus for performing real-time endpoint detection in automatic speech recognition |
| US7386799B1 (en) * | 2002-11-21 | 2008-06-10 | Forterra Systems, Inc. | Cinematic techniques in avatar-centric communication during a multi-user online simulation |
| KR100580619B1 (ko) * | 2002-12-11 | 2006-05-16 | 삼성전자주식회사 | 사용자와 에이전트 간의 대화 관리방법 및 장치 |
| WO2007044806A2 (en) * | 2005-10-11 | 2007-04-19 | Aol Llc | Ordering of conversations based on monitored recipient user interaction with corresponding electronic messages |
| US9250703B2 (en) * | 2006-03-06 | 2016-02-02 | Sony Computer Entertainment Inc. | Interface with gaze detection and voice input |
| EP2012304B1 (en) * | 2007-07-06 | 2012-09-05 | Zero To One Technology | Methods for electronically analysing a dialogue and corresponding systems |
| WO2009104564A1 (ja) * | 2008-02-20 | 2009-08-27 | インターナショナル・ビジネス・マシーンズ・コーポレーション | 仮想空間における会話サーバ、会話のための方法及びコンピュータ・プログラム |
| US8289288B2 (en) * | 2009-01-15 | 2012-10-16 | Microsoft Corporation | Virtual object adjustment via physical object detection |
| US8416715B2 (en) * | 2009-06-15 | 2013-04-09 | Microsoft Corporation | Interest determination for auditory enhancement |
| EP2541543B1 (en) | 2010-02-25 | 2016-11-30 | Panasonic Intellectual Property Management Co., Ltd. | Signal processing apparatus and signal processing method |
| US10198775B2 (en) * | 2010-06-23 | 2019-02-05 | Microsoft Technology Licensing, Llc | Acceleration of social interactions |
| US20150189233A1 (en) * | 2012-04-30 | 2015-07-02 | Goggle Inc. | Facilitating user interaction in a video conference |
| JP6385150B2 (ja) * | 2014-06-13 | 2018-09-05 | 株式会社Nttドコモ | 管理装置、会話システム、会話管理方法及びプログラム |
| JP6400445B2 (ja) | 2014-11-27 | 2018-10-03 | Kddi株式会社 | 会話分析装置、会話分析システム、会話分析方法及び会話分析プログラム |
| WO2016158014A1 (ja) * | 2015-03-31 | 2016-10-06 | ソニー株式会社 | 情報処理装置、通信システム、情報処理方法及びプログラム |
| US20160308920A1 (en) * | 2015-04-16 | 2016-10-20 | Microsoft Technology Licensing, Llc | Visual Configuration for Communication Session Participants |
| JP6651989B2 (ja) * | 2015-08-03 | 2020-02-19 | 株式会社リコー | 映像処理装置、映像処理方法、及び映像処理システム |
| CN106453030B (zh) * | 2015-08-12 | 2019-10-11 | 大连民族学院 | 一种获取社交关系链的方法及装置 |
| JP6772023B2 (ja) * | 2015-10-30 | 2020-10-21 | コニカ ミノルタ ラボラトリー ユー.エス.エー.,インコーポレイテッド | ユーザー状態検出による集団相互作用の方法及びシステム |
| US20170301037A1 (en) * | 2016-04-19 | 2017-10-19 | International Business Machines Corporation | Group discourse architecture |
| CN106291744A (zh) * | 2016-07-29 | 2017-01-04 | 深圳朝伟达科技有限公司 | 岩质边坡估计交互式稳定性显示系统 |
-
2019
- 2019-01-10 WO PCT/JP2019/000597 patent/WO2019139101A1/ja not_active Ceased
- 2019-01-10 US US16/959,884 patent/US11837233B2/en active Active
- 2019-01-10 JP JP2019564747A patent/JP7276158B2/ja active Active
- 2019-01-10 CN CN201980007295.6A patent/CN111542876B/zh active Active
- 2019-01-10 EP EP19738367.2A patent/EP3739573B1/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6280768A (ja) | 1985-10-04 | 1987-04-14 | Hitachi Ltd | ステレオ画像処理方法 |
| JPH04323689A (ja) | 1991-04-24 | 1992-11-12 | Toshiba Corp | 会議進行支援装置 |
| JPH06169458A (ja) | 1992-11-30 | 1994-06-14 | Matsushita Electric Ind Co Ltd | 移動体の自動追尾装置 |
| JPH09251299A (ja) | 1996-03-15 | 1997-09-22 | Toshiba Corp | マイクロホンアレイ入力型音声認識装置及び方法 |
| JP2000184345A (ja) * | 1998-12-14 | 2000-06-30 | Nec Corp | マルチモーダルコミュニケーション支援装置 |
| WO2008139717A1 (ja) * | 2007-05-09 | 2008-11-20 | Panasonic Corporation | 表示装置、表示方法、表示プログラム |
| WO2009004731A1 (ja) * | 2007-07-05 | 2009-01-08 | Panasonic Corporation | 画像音声通信装置及び画像表示方法 |
| JP2010266522A (ja) * | 2009-05-12 | 2010-11-25 | Nippon Telegr & Teleph Corp <Ntt> | 対話状態分割装置とその方法、そのプログラムと記録媒体 |
| JP2011237965A (ja) * | 2010-05-10 | 2011-11-24 | Seiko Epson Corp | 会話比率算出装置、接客データ記録装置、会話比率算出方法およびプログラム |
| JP2012226665A (ja) | 2011-04-21 | 2012-11-15 | Canon Inc | 情報処理装置、その処理方法、プログラム及び撮像装置 |
| JP2013058221A (ja) | 2012-10-18 | 2013-03-28 | Hitachi Ltd | 会議分析システム |
| JP2017009825A (ja) * | 2015-06-23 | 2017-01-12 | トヨタ自動車株式会社 | 会話状況分析装置および会話状況分析方法 |
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP3739573A4 |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2021182089A (ja) * | 2020-05-20 | 2021-11-25 | 清水建設株式会社 | 人物関係抽出システム |
| JP7471911B2 (ja) | 2020-05-20 | 2024-04-22 | 清水建設株式会社 | 人物関係抽出システム |
| JP2024147690A (ja) * | 2020-06-01 | 2024-10-16 | ハイラブル株式会社 | 音声会議装置、音声会議システム及び音声会議方法 |
| WO2021245759A1 (ja) * | 2020-06-01 | 2021-12-09 | ハイラブル株式会社 | 音声会議装置、音声会議システム及び音声会議方法 |
| JPWO2021245759A1 (ja) * | 2020-06-01 | 2021-12-09 | ||
| JP7766887B2 (ja) | 2020-06-01 | 2025-11-11 | ハイラブル株式会社 | 音声会議装置、音声会議システム及び音声会議方法 |
| US12260876B2 (en) | 2020-06-01 | 2025-03-25 | Hylable Inc. | Voice conference apparatus, voice conference system and voice conference method |
| JP7530070B2 (ja) | 2020-06-01 | 2024-08-07 | ハイラブル株式会社 | 音声会議装置、音声会議システム及び音声会議方法 |
| JP2022072455A (ja) * | 2020-10-29 | 2022-05-17 | 京セラ株式会社 | 仮想イベントシステム、情報処理装置、出力制御方法及びプログラム |
| JP7560321B2 (ja) | 2020-10-29 | 2024-10-02 | 京セラ株式会社 | 仮想イベントシステム、情報処理装置、出力制御方法及びプログラム |
| WO2024100703A1 (ja) * | 2022-11-07 | 2024-05-16 | マクセル株式会社 | 映像表示装置、映像表示システム及び映像表示装置の制御方法 |
| JPWO2024100703A1 (ja) * | 2022-11-07 | 2024-05-16 | ||
| JP7792533B2 (ja) | 2022-11-07 | 2025-12-25 | マクセル株式会社 | 映像表示装置、映像表示システム及び映像表示装置の制御方法 |
| KR102772721B1 (ko) * | 2024-04-02 | 2025-02-26 | 케이에듀테크 주식회사 | 온라인 및 오프라인 하이브리드 러닝 참여자 분석 시스템 |
| JP7847733B1 (ja) * | 2025-07-15 | 2026-04-17 | 三菱電機株式会社 | 発言データ提供装置、発言推定システム、発言データ提供システム、発言データ提供方法及びプログラム |
Also Published As
| Publication number | Publication date |
|---|---|
| JPWO2019139101A1 (ja) | 2021-01-28 |
| CN111542876A (zh) | 2020-08-14 |
| US20200335105A1 (en) | 2020-10-22 |
| US11837233B2 (en) | 2023-12-05 |
| EP3739573B1 (en) | 2023-06-28 |
| EP3739573A4 (en) | 2021-01-13 |
| JP7276158B2 (ja) | 2023-05-18 |
| EP3739573A1 (en) | 2020-11-18 |
| CN111542876B (zh) | 2025-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP7276158B2 (ja) | 情報処理装置、情報処理方法およびプログラム | |
| US20120077172A1 (en) | Presentation system | |
| JP7062126B1 (ja) | 端末、情報処理方法、プログラム、および記録媒体 | |
| JP2018205638A (ja) | 集中度評価機構 | |
| CN114554135A (zh) | 线上会议方法及电子设备 | |
| JP7841578B2 (ja) | 配信システム | |
| WO2022137547A1 (ja) | コミュニケーション支援システム | |
| JP6901190B1 (ja) | 遠隔対話システム、遠隔対話方法及び遠隔対話プログラム | |
| CN112312062A (zh) | 多人会议记录回放的3d展示方法、存储介质及终端设备 | |
| JPWO2009119288A1 (ja) | コミュニケーションシステム及びコミュニケーションプログラム | |
| US20220189200A1 (en) | Information processing system and information processing method | |
| JP6267819B1 (ja) | 授業システム、授業サーバ、授業支援方法、及び授業支援プログラム | |
| JP7152453B2 (ja) | 情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム | |
| US20230196632A1 (en) | Information processing device and information processing method | |
| JP6456163B2 (ja) | 情報処理装置、音声出力方法及びコンピュータプログラム | |
| JP6849228B2 (ja) | 教室システム | |
| Bergstrom et al. | Conversation votes: enabling anonymous cues | |
| JP2024147690A (ja) | 音声会議装置、音声会議システム及び音声会議方法 | |
| JP2009060220A (ja) | コミュニケーションシステム及びコミュニケーションプログラム | |
| CN104539873B (zh) | 远程会议系统和进行远程会议的方法 | |
| JP2018165978A (ja) | 授業システム、授業サーバ、授業支援方法、及び授業支援プログラム | |
| Bergstrom et al. | Vote and be heard: Adding back-channel signals to social mirrors | |
| CN117793607A (zh) | 一种播放控制方法及装置 | |
| JP2019057061A (ja) | 情報出力装置及びプログラム | |
| JP2012003698A (ja) | 会議支援装置、会議支援方法、会議支援プログラムおよび記録媒体 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19738367 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2019564747 Country of ref document: JP Kind code of ref document: A |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2019738367 Country of ref document: EP Effective date: 20200812 |
|
| WWG | Wipo information: grant in national office |
Ref document number: 201980007295.6 Country of ref document: CN |