Adaptive Wavelet Denoising Based on Cross-Validation
Jiasheng Li · Journal of Hunan University · 2008
A new wavelet-based adaptive denoising algorithm was presented.By using a modified twofold cross-validation,a noise-corrupted signal was divided into two parts: one for estimating,and the other as a reference signal,and they made it possible to search for optimal thresholds by using gradient-based adaptive algorithms.The numerical results indicate that the proposed method outperforms the standard wavelet thresholding denoising methods,like Donoho's VisuShrink and SureShrink,in MSE sense.The proposed algorithm does not need any prior information of the noise-distorted signal,and its convergence speed is faster.So it is fit for real-time signal processing.