Finite Rate of Innovation Scheme for Identifying L0 Norm in the Reconstruction of Sparse Signals

Sparse signals are the finite rate of innovation (FRI) signals that can recover from the lower dimensional signals by utilizing appropriate recovery schemes. Estimating the positions and weights of each nonzero coefficient allows the recovery of the sparse signal. Therefore, it is crucial to determine the number of nonzero coefficients for complete recovery. Because of this, we proposed a new FRI scheme that perfectly identifies the number of nonzero coefficients (L0 norm) before recovering the sparse signal in a noiseless environment. This makes the proposed scheme universal and able to handle signals with an unknown L0 norm. The scheme is also effective in noisy environments and performs better reconstruction than the traditional scheme.