Wavelet transform based features of skin blood flow response signal for pressure ulcer evaluation

Abstract

Pressure Ulcers (PUs) are localized tissue damage that usually occur over the soft tissue of body prominence when a subject is exposed to prolonged external mechanical loading. Several studies have proposed that skin blood flow response (SBFR) can be used in PU evaluation to determine tissue ischemic accumulation due to insufficient time of tissue recovery. In previous study, twenty one Sprague Dawley rats weigh 388-481g that were divided into three different group have been used to investigate the trends of SBFR signal using time domain features like peak reactive hyperaemia (RH), time to peak RH and area under the RH curve as well as frequency domain features like peak power spectral density (PSD) and total PSD. However, the results indicate that both frequency domain features are not effective at determining individual insufficient recovery time. In this study, Continuous Wavelet Transform (CWT) based features such as average amplitude and relative amplitude based on Morlet wavelet function scale 200 are investigated. The results show that the samples representing inconsistent trend of average amplitude for metabolic frequency range are dominant in all short (3 samples) , moderate (4 samples) and prolonged groups (4 samples), while no clear pattern can be established for relative amplitude. Hence both features may not suitable at distinguishing between sufficient and insufficient recovery time due to the low percentage in number of samples.