Comparing on sparse heart sound recovery algorithms
Indrarini Dyah Irawati, Ervin Masita Dewi · 2016
This paper applied the concept of Compressive Sensing on practical problem for sparse heart sound recovery. The sparse representation matrix used Haar wavelet transform, while the measurement matrix used random orthogonal matrix. We compare the performance of different recovery algorithm such as Subspace Pursuit (SP), Iterative Hard Thresholding (IHT), Compressive Sampling Matching Pursuit(CoSaMP), Orthogonal Matching Pursuit (OMP), and L1 norm optimization. The experiment results show that all of method can recovery a periodic heart beat sound. The best recovery performance is L1 norm, but the computational time is worse.