Time Series Extrinsic Regression for Physical Rehabilitation Assessment

Abstract

Rehabilitation is the process of assisting people with disabilities in regaining their function and independence. As artificial neural networks are trained on large datasets using deep learning, rehabilitation can be improved by providing individualized and efficient treatment options. As human rehabilitation involves multivariate time series data, we review well-known algorithms for the classification of time series data. We also discuss the challenges and opportunities presented by the use of deep learning in rehabilitation, including the need for large and diverse datasets and the potential for bias in algorithms. Overall, our analysis indicates that deep learning has the potential to improve rehabilitation outcomes and the lives of disabled individuals. A comparison of many methodologies was conducted in order to establish a framework capable of supporting and reliably evaluating patients’ workouts throughout recovery programs. In order to assess the algorithms, two datasets pertaining to human rehabilitation are used: KIMORE, and UI-PRMD for regression tasks.