A novel patient-specific dictionary matrix optimization method for ECG compressed sensing improves signal reconstruction quality and reduces execution time.
This paper proposes a new method for dictionary matrix optimization with the aim of improving the reconstruction quality of ECG signals delivered by a Compressed Sensing (CS) algorithm. The method exploits the features common to all the records of the ECG signal of the same patient to obtain an optimized dictionary with reduced size. In this way, the signal reconstruction from the compressed samples is performed in a do-main defined by a base with a reduced cardinality, thus allowing to increase the signal’s reconstruction quality and to reduce the execution time of the reconstruction algorithm. The mathematical model for the patient specific ECG signals dictionary optimization is described, and a preliminary experimental assessment is presented. The obtained results clearly demonstrates that the proposed method exhibits a reconstruction quality in terms of Percentage of Root-mean-squared Difference (PRD) lower than a method adopting the non-optimized dictionary matrix.
Picariello et al. (Mon,) studied this question.