Exact risk analysis of wavelet spectrum thresholding rules

摘要

The signal analysis and processing based on wavelets has recently been a much used discipline, whose range of applications spreads steadily. One of its most important applications is signal denoising. In this area, the so-called thresholding is an important technique of signal processing. Besides common thresholding rules, several others has been developed. Their statistical properties obtained via so-called exact risk analysis are described in this paper.