Soft Thresholding Using Moore–Penrose Inverse

The acquisition of a discrete-time signal is an important part compressive sensing problem. A high-accuracyalgorithm that could bring better recovery performance often called for. In this work, two thresholding algorithms involve soft decision are proposed using the Moore-Penrose inverse. Numerical examples conducted and illustrate in optimal case, both methods consume computational time same level as conventional homotopy algorithm (SHA). Under no knowledge regularization parameter, will perform than SHA with less amount required for computation. Taking non-sparse electroencephalogram from real measurement into account, all provide nearly error several compression ratios, while SHA.