engleski

Wavelet transforms with application in signal denoising

This paper is focused on application of the wavelet transforms in signal denoising. Wavelet transforms have ability to localize the signal properties in both the frequency and the time. Signal decomposition results in packing the main signal properties in a small number of large wavelet coefficients and constraining noise to small coefficients. To remove noise from the signal, wavelet shrinkage technique is used – thresholding is applied to wavelet coefficients at each scale. This way, only larger coefficients are retained, while the coefficients representing noise are eliminated before the reconstruction. An example of signal decomposition and denoising is provided, in which these effects can be easily observed.

wavelets; denoising; shrinkage; thresholding

nije evidentirano