Signal denoising based on non-local similarity and wavelet transform

A novel signal denoising method combining translation invariant (TI) wavelet transform with non-local signal similarities is developed. Signal blocks with similar structures are assembled together to build up groups with strong correlations, and then the translation invariant wavelet transform is applied on these groups to produce, in an enhanced sparsity manner, the transformed coefficients, these coefficients are hard-thresholded and inverse transformed back into their denoised versions; finally these denoised blocks are aggregated together to get the final estimate of the true signal. Experimental results confirm that the proposed method can achieve certain improvements in denoising performance compared with the traditional translation invariant wavelet methods.