Analyzing center of pressure progression during bed exits

Ren, Zhaofen; Grant, Theresa; Goubran, Rafik; El-Tanany, M.; Knoefel, Frank; Sveistrup, Heidi; Beznoska, Martin; Jutai, Jeffrey W.

doi:10.1109/embc.2014.6943955

Cited by 6 publications

(2 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…to future research.Ren et al took an image-based approach for observing postural control behaviour in bed-exit data prior to and post-fall using the PSM and the COP[14]. They observed the magnitude and frequency of movement in the y-direction by x-values of the COP.…”

mentioning

confidence: 99%

“…The PSM's capabilities have been explored in healthcare research as an unobtrusive monitoring tool for a range of clinical areas and conditions. To name a few applications of the PSM, the PSM has also been used as a real-time data collection tool to support geriatric in-bed mobility detection for fall prevention, ulcer prevention, sleep state classification, respiration rate estimation, and the detection of detrimental sitting positions in children to avoid musculoskeletal disorders[1],[14]-[17]. Applications of the PSM in motion detection and healthcare are further discussed in Section 2.2.2 and Section 2.2.3.As mentioned above, the pressure sensor mat is a tool for measuring contact pressure but does not necessarily require direct contact with skin or other objects that will apply pressure on it.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Non-Invasive Motion Detection and Classification in NICU Patients using Ballistographic Signals from a Pressure Sensitive Mat

Aziz¹

View full text Add to dashboard Cite

Patient movements can cause motion artifacts on physiological signals and can result in false alarms in a continuous patient care environment. This thesis explores the use of centre of pressure (COP) signals from a pressure sensitive mat, placed below neonates in the neonatal intensive care unit to (a) detect patient movement in real-time, and (b) classify the source of movement as upper or lower body. The COP is the sum of all vectors of pressure acting on the PSM; it is the point where the total force due to pressure is equal on both halves of the PSM. To achieve (a) the sum distance travelled by the COP is tracked over time using a sliding window with data from seven patients. Windows exhibiting large deviations in the COP are indicative of patient motion. Window boundary suppression led to improved movement detection with precision scores of 0.84 and recall of 0.71 at a window length of 10 seconds in a real-time sliding window approach. To address problem (b) -which has not previously been attempted in non-invasive monitoring research -six features were derived from the COP, and feature selection was done with out-of-bag error feature importance ranking with a random forest to remove anomalous features, then the feature set was further reduced using sequential forward selection for use in SVMs trained with leave-one-groupout cross-validation. Balancing training and test data with SMOTE or majority under sampling improved classifier performance by approximately two-fold. It was found that using a sample imputation approach of adding ~13 minutes of hand-annotated new subject data to the training set makes the classifier most useful to a new patient, producing accuracy scores of ~87.29%, precision of 0.90, and recall of 0.84. These models and insights may be used in a real-time motion detection algorithm and create a foundation for future work in limb movement classification and detection. v Table of Contents ABSTRACT .

show abstract

mentioning

confidence: 99%

mentioning

confidence: 99%