On the efficient application of compressive sensing of physiological signals in medical diagnostics

Wireless telemonitoring of physiological signals is an evolving direction in personalized medicine and home-based e-Health. There are several constraints in designing such systems. The three important constraints are energy consumption, data compression and device cost. Compressive Sensing (CS) is an emerging data compression technique that overcomes those constraints. Nevertheless, the non-sparsity of physiological signals presents a major issue to the existing compressive sensing algorithms. This research proposes to use a developed compressive sensing algorithm which has the ability to recover such non-sparse physiological signals. This algorithm is Block Sparse Bayesian Learning (BSBL). The proposed algorithm and the conventional CS algorithm were used to compress Fetal ECG (FECG) signals. Results showed that using BSBL to recover non-sparse FECG is more efficient comparing with the conventional CS algorithm, SL0.