Natural Thresholding Algorithms for Signal Recovery With Sparsity

The algorithms based on the technique of optimal $k$-thresholding (OT) were recently proposed for signal recovery, and they are very different from traditional family hard thresholding methods. However, computational cost OT-based remains high at current stage their development. This stimulates development so-called natural (NT) algorithm its variants in this paper. NT is developed through first-order approximation regularized model, thus significantly lower than that algorithms. guaranteed performance NT-type recovery noisy measurements shown under restricted isometry property concavity objective function model. Empirical results indicate robust comparable to several mainstream sparse recovery.