Title | ||
---|---|---|
Exploring clinical correlations in centroid-based gait metrics from depth data collected in the home. |
Abstract | ||
---|---|---|
A longitudinal study in the home setting using inexpensive depth cameras was done over 34 months to investigate the ability to predict clinical events. Previous work developed a set of metrics based upon the movement of the centroid computed from segmented depth data [14]. A predictive analysis method is developed allowing the identification of significant changes in the subjectu0027s gait. These changes are compared to the subjectu0027s clinical events and correlated with standard Fall Risk Assessments (FRA). The method developed here allows the proper clustering of all purposeful walks in the residence to isolate the subject from visitors, and identification of significant changes using a set of metrics unique to each subject. Correct detection of events and non-events ranged between 75% and 94% across a set of 7 residents. These predicted events were also found to correlate strongly with established monthly FRAs. |
Year | Venue | Field |
---|---|---|
2017 | PervasiveHealth | Longitudinal study,Gait,Pattern recognition,Computer science,Fall risk,Artificial intelligence,Cluster analysis,Centroid,Distributed computing |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
3 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Robert Wallace | 1 | 0 | 0.68 |
Carmen Abbott | 2 | 66 | 7.51 |
Marjorie Skubic | 3 | 1045 | 105.36 |