THz signal denoising via redundant representation

Denoising is widely used in THz signal processing. In recent years, there have been many papers about applying wavelet-based denoising techniques to achieve clean THz signal. Traditionally, denoising by shrinkage on orthogonal, decimated wavelet transforms satisfies most THz applications with dramatic improvements in SNR for TDS pulses. In this paper, we apply the shrinkage technique to non-orthogonal, redundant THz signal representations to TDS pulses and it is shown that this scheme is capable of even an even greater SNR gains.