Abstract | ||
---|---|---|
In this work, we outline an automated system for continuous assessment of patient recovery levels generically in movement-related disorders using wrist-worn accelerometers, with visualisation of aspects of recovery trends. We demonstrate our model on assessment of recoveries from hemiparetic stroke.
A clinically-assessed functional recovery score from hemiparetic stroke is periodically collected alongside daily-living wrist-based accelerometer data. After accelerometer preprocessing steps, inferences on individual sliding windows are performed, followed by an overall assessment over a period of time (e.g., past three days). Using isotonic regression, excluding less relevant datapoints, and model averaging over hyperparameters further improve predictions.
Through time series decomposition methods of the sliding window inferences, we visualise informative aspects of recovery patterns: overall recovery trend, diurnal exertion patterns and times of outliers in capabilities exhibited.
|
Year | DOI | Venue |
---|---|---|
2019 | 10.1145/3341163.3347731 | Proceedings of the 23rd International Symposium on Wearable Computers |
Keywords | Field | DocType |
accelerometer, chedoke arm and hand inventory, health, human activity recognition, machine learning, wearables | Rehabilitation,Accelerometer,Wearable computer,Computer science,Stroke,Physical medicine and rehabilitation,Embedded system | Conference |
ISBN | Citations | PageRank |
978-4503-6870-4 | 1 | 0.36 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shane Halloran | 1 | 41 | 2.00 |
Lin Tang | 2 | 1 | 0.36 |
Yu Guan | 3 | 195 | 22.59 |
Jianqing Shi | 4 | 1 | 3.07 |
Janet A. Eyre | 5 | 8 | 2.32 |