Papers by Keyword: MEME Microstructure

Abstract: A new signal-denoising approach based on DT-CWT (Dual-Tree Complex Wavelet Transform) is presented in this paper to extract feature information from microstructure profile. It takes advantage of shift invariance of DT-CWT, non-Gaussian probability distribution for the wavelet coefficients and the statistical dependencies between a coefficient and its parent. This approach substantially improved the performance of classical wavelet denoising algorithms, both in terms of SNR and in terms of visual artifacts. A simulated MEMS microstructure signal is analyzed.