Assessing Parkinson's disease symptoms
AssigneeInternational Business Machines Corporation
InventorsTian Hao, Jeffrey L. Rogers, Umar Asif, Erhan Bilal, Deval Samirbhai Mehta, Stefan Harrer, Jianbin Tang, Stefan von Cavallar, Paolo Fraccaro
Life of this patent
Estimating the term…
Abstract
The exemplary embodiments disclose a system and method, a computer program product, and a computer system for assessing one or more Parkinson's disease symptoms. The exemplary embodiments may include collecting data of a user's motion, extracting one or more features from the collected data, and assessing one or more Parkinson's disease symptoms of the user based on applying one or more models to the data.
Claims20
A computer implemented method for assessing a plurality of Parkinson's disease symptoms, the method comprising:
- training a plurality of models, wherein each of the plurality of models corresponds to a different symptom of the Parkinson's disease symptoms;
- collecting video data corresponding to motion of a user;
- performing motion reconstruction of the user from the video data by:
- identifying, by a pipeline of pre-trained deep learning neural networks, at least one of skeletal, facial, and hand key points;
- generating a time series that maps movement of the key points in degrees per second within at least one of a 2D and 3D space;
- and
- smoothing the time series of the key points using linear trend fitting to remove an effect of gravity on the user;
- extracting a plurality of features from the motion reconstruction;
- and
- assessing the Parkinson's disease symptoms of the user based on applying the plurality of trained models to the plurality of features.
A computer program product for assessing a plurality of Parkinson's disease symptoms, the computer program product comprising:
- one or more non-transitory computer-readable storage media and program instructions stored on the one or more non-transitory computer-readable storage media capable of performing a method, the method comprising:
- training a plurality of models, wherein each of the plurality of models corresponds to a different symptom of the Parkinson's disease symptoms;
- collecting video data corresponding to motion of a user;
- performing motion reconstruction of the user from the video data by:
- identifying, by a pipeline of pre-trained deep learning neural networks, at least one of skeletal, facial, and hand key points;
- generating a time series that maps movement of the key points in degrees per second within at least one of a 2D and 3D space;
- and
- smoothing the time series of the key points using linear trend fitting to remove an effect of gravity on the user;
- extracting a plurality of features from the motion reconstruction;
- and
- assessing the Parkinson's disease symptoms of the user based on applying the plurality of trained models to the plurality of features.
A computer system for assessing a plurality of Parkinson's disease symptoms, the computer system comprising:
- one or more computer processors, one or more computer-readable storage media, and program instructions stored on the one or more of the computer-readable storage media for execution by at least one of the one or more processors capable of performing a method, the method comprising:
- training a plurality of models, wherein each of the plurality of models corresponds to a different symptom of the Parkinson's disease symptoms;
- collecting video data corresponding to motion of a user;
- performing motion reconstruction of the user from the video data by:
- identifying, by a pipeline of pre-trained deep learning neural networks, at least one of skeletal, facial, and hand key points;
- generating a time series that maps movement of the key points in degrees per second within at least one of a 2D and 3D space;
- and
- smoothing the time series of the key points using linear trend fitting to remove an effect of gravity on the user;
- extracting a plurality of features from the motion reconstruction;
- and
- assessing the Parkinson's disease symptoms of the user based on applying the plurality of trained models to the plurality of features.
Dependencies are inferred from the claim text.