Identification of single-breaths and respiration patterns by ribcage and abdomen signals

A novel algorithm is proposed in this paper for accurate identification of single-breaths and respiration patterns. The algorithm works on normal breathing segments of Ribcage (RC) and Abdomen (ABD) records and consists of three stages. In the first stage, normal breathing segments are pre-processed by applying a polynomial-curve-fitting method to remove artifacts. Then during the second stage, the frequencies of normal breathing, phase difference and coherence between RC and ABD, were studied in the frequency domain. In the third stage, based on the frequency analysis results, an algorithm was developed in the time domain for identification of single-breaths and respiration patterns. Tests on six babies prove the validity and efficiency of our algorithm. This work is supported by and conducted at the Montreal Children's Hospital.