BREATHING PATTERN RECOGNITION OF ABDOMINAL WALL MOVEMENT BY USING ENSEMBLE EMPIRICAL MODE DECOMPOSITION
Thoracic breathing (TB), abdominal breathing (AB), and mixing breathing are common respiratory functions. Individuals usually breathe thoracically, whereas the breathing pattern of AB is vague. Despite the statistical representation of the physiological benefits of AB, coping with a time-variant environment still remains challenging. Therefore, based on ensemble empirical mode decomposition (EEMD), this study compares the identification types of using R value, power proportion, and modified significant test (MST). Respiratory maneuver of 26 subjects results that MST varied with a paced breathing frequency is the highest accurate recognition rate of TB (80.8% in 0.2 Hz and 88.5% in 0.1 Hz) and of AB (73.1% in 0.2 and 0.1 Hz). Results of this study demonstrate that EEMD is an adaptive algorithm to decompose respiratory movement. Furthermore, MST is a highly promising feature extraction method for breathing type recognition.