Activity recognition for physical therapy: fusing signal processing features and movement patterns

This paper discusses the classification of activities in the context of physical therapy. Usually, specific standardized activities are subjectively assessed, often by means of a patient-reported questionnaire, to estimate a patient's activity capacity, defined as the ability to execute a task. Automatic recognition of these activities is of vital importance for a more objective and quantitative approach to the problem. The proposed accelerometry-based algorithm merges standard signal processing features with information obtained from direct activity pattern matching using dynamic time warping (DTW) in a linear model. This study with 28 spondyloarthritis patients performing 10 activities shows the improvement in activity classification accuracy due to the fusion of the two approaches, up to 93.6%. This is a significant increase compared to similar models based on either of the approaches alone (p < 0.01). Although this paper mainly contributes to the activity recognition step, it also briefly discusses the advantages of the approach with regard to further automated evaluation of the recognized activities.