Emotion estimating system
An emotion estimating system includes a learning model and an estimation unit. The learning model accepts external information and biological information as input, and outputs an emotion of a user. The estimation unit changes a weighting applied to external information about the user detected by a first detector and a weighting applied to biological information about the user detected by a second detector in accordance with a situation around the user, and estimates the emotion output as a result of inputting external information and biological information changed by the respective weightings into the learning model as the emotion of the user.
1. An emotion estimating system, comprising:
a learning model configured to accept external information and biological information as input, and configured to output a determination of an emotion of a user based on the external information and the biological information; and
an estimation unit configured to:
receive the external information about the user detected by a first detector and the biological information about the user detected by a second detector,
determine a weighting factor for the external information and a weighting factor for the biological information based on a situation around the user, and
estimate an emotional state of the user by inputting the weighted external information and the weighted biological information into the learning model.
2. The emotion estimating system according to claim 1 , wherein
the situation around the user is determined on a basis of environment information detected by a third detector that detects environmental information about an environment around the user.
3. The emotion estimating system according to claim 2 , wherein
when the estimation unit determines that the situation around the user is a public situation or a situation that tends to induce nervousness or stress, the estimation unit makes the weighting factor for the biological information larger than the weighting factor for the external information.
4. The emotion estimating system according to claim 2 , wherein
when the estimation unit determines that the situation around the user is a private situation or a situation that tends to induce relaxation or relief, the estimation unit makes the weighting factor for the biological information smaller than the weighting factor for the external information.
5. The emotion estimation system according claim 1 , wherein
when the estimation unit determines that the user has a personality type in which emotion is not readily exhibited in expression and behavior, the estimation unit makes the weighting factor for the biological information larger than the weighting factor for the external information.
6. The emotion estimating system according to claim 1 , wherein
when the estimation unit determines that the user has a personality type in which emotion is readily exhibited in expression and behavior, the estimation unit makes the weighting factor for the biological information smaller than the weighting factor for the external information.
7. An emotion estimating system, comprising:
a plurality of learning models, each of which is configured to accept external information and biological information as input, and configured to output a determination of an emotion of a user based on the external information and the biological information;
a selector that selects a learning model to use, in accordance with a situation around the user; and
an estimation unit configured to:
receive the external information about the user detected by a first detector and the biological information about the user detected by a second detector, and
estimate an emotional state of the user by inputting the external information and the biological information into the selected learning model.
8. The emotion estimating system according to claim 7 , further comprising:
a third detector that detects environmental information about an environment around the user, wherein
the selector determines the situation around the user on a basis of the environmental information detected by the third detector.
9. An emotion estimating system, comprising:
a learning model that is configured to accept external information, biological information, and environmental information about an environment around a user as input, and configured to output an emotion of the user based on the external information, the biological information, and the environmental information; and
an estimation unit configured to:
receive the external information about the user detected by a first detector, the biological information about the user detected by a second detector, and the environmental information about the environment around the user detected by a third detector, and
estimate an emotional state of the user by inputting the external information, the biological information, and the environmental information into the learning model.
10. The emotion estimating system according to claim 1 , wherein
the first detector detects at least one of an expression and speech of the user as the external information, and
the second detector detects at least one of skin potential and heart rate of the user as the biological information.
11. The emotion estimating system according to claim 1 , wherein
the first detector and the second detector are provided in an interactive robot.
12. The emotion estimating system according to claim 11 , wherein
environmental information about an environment around the user is detected by an environmental sensor installed near where the interactive robot is present.
13. The emotion estimating system according to claim 12 , wherein
the emotion estimating system includes a control server,
the interactive robot transmits current position information to the control server, and
the control server transmits, to the interactive robot, environmental information about the environment around the user acquired from an environmental sensor installed at a position corresponding to the current position information acquired from the interactive robot.
14. The emotion estimating system according to claim 13 , wherein
the control server requests the environmental sensor installed at a position corresponding to the current position information acquired from the interactive robot to acquire the environmental information about the environment around the user.