Training a gesture recognition model
AssigneeTDK Corporation
InventorsRémi Louis Clément PONÇOT, Juan S Mejia SANTAMARIA, Abbas ATAYA, Etienne De FORAS, Bruno FLAMENT
Life of this patent
Estimating the term…
Abstract
In a method for training a gesture recognition model, gesture data is collected from an inertial measurement unit (IMU) positioned on one side of a user, wherein the IMU is capable of collecting data when positioned on either side of the user. A transformation is applied to the gesture data, wherein the transformation generates transformed gesture data that is independent of either side of the user. A gesture recognition model is trained using the transformed gesture data.
Claims20
A method for training a gesture recognition model, the method comprising:
- collecting gesture data from an inertial measurement unit (IMU) positioned on one side of a user, wherein the IMU is capable of collecting data when positioned on either side of the user;
- applying a transformation to the gesture data, wherein the transformation generates transformed gesture data that is independent of either side of the user;
- and
- training a gesture recognition model using the transformed gesture data.
A wearable system utilizing gesture recognition, the system comprising:
- an inertial measurement unit (IMU) positioned on one side of a user for collecting gesture data based on movement of the IMU;
- a gesture recognition model trained on one side of the user;
- and
- a processing unit for performing gesture recognition using the gesture data, the processing unit configured to:
- determine a side of the user on which the IMU is positioned;
- provided the IMU is positioned on an opposite side of the user on which the gesture recognition is trained, applying a transformation to the gesture data, wherein the transformation transforms the gesture data such that transformed gesture data is transformed to a side of the user on which the gesture recognition model is trained;
- and
- performing gesture recognition using the transformed gesture data.
A wearable system utilizing gesture recognition, the system comprising:
- an inertial measurement unit (IMU) positioned on one side of a user for collecting gesture data based on movement of the IMU;
- a gesture recognition model that is trained independent of which side of the user the IMU is positioned;
- and
- a processing unit for performing gesture recognition using the gesture data, the processing unit configured to:
- applying a transformation to the gesture data, wherein the transformation is an absolute value transformation that transforms the gesture data such that transformed gesture data is independent of which side of the user the IMU is positioned;
- and
- performing gesture recognition using the transformed gesture data.
A non-transitory computer readable storage medium having computer readable program code stored thereon of a method for training a gesture recognition model, the method comprising:
- collecting gesture data from inertial measurement units (IMUs) positioned on one side of a plurality of users, wherein the IMUs are capable of collecting data when positioned on either side of the plurality of users;
- determining which side of the user on which the IMU is positioned, wherein a transformation is configured to be applied to a first side of the user;
- and
- provided the IMU is positioned on the first side of the user on which the transformation is configured to be applied, applying a transformation to the gesture data, wherein the transformation transforms the gesture data such that transformed gesture data is representative of data collected from an IMU positioned on a second side of the user;
- and
- training a gesture recognition model using the gesture data received from an IMU positioned on the second side of the user and the transformed gesture data for the gesture data received from an IMU positioned on the first side of the user.
Dependencies are inferred from the claim text.