Sparse Ramanujan Sequences (SRS) transform based signal denoising

Sparse Ramanujan sequences (SRS) are very recently introduced sparse sequences derived from Ramanujan sums (RS). RS are integer-valued sequences proposed by famous Indian mathematician S. Ramanujan. They have some attractive properties such as orthogonality and periodicity. SRS also possess similar properties and in addition, these sequence elements contain only values from the set {1, 0, -1}. The transforms based on these sequences are also sparse which approximate signal to a fewer number of coefficients. This sparse approximation has many advantages in several signal processing applications like signal denoising and signal compression. SRS transform-based signal denoising is proposed in this paper. The denoising performance of SRS transform is evaluated using performance metrics like mean square error (MSE) and output signal to noise ratio (SNR) and compared with existing transform-based denoising techniques.