An Adaptive Computation of Contour Representations for Mode Decomposition

Abstract

Abstract-This letter addresses the problem of the detection and estimation of the modes of a multicomponent signal using the reassignment framework. More precisely, we propose a new algorithm to estimate the ridges representing the time-frequency (TF) signatures of the modes based on the local orientation of the reassignment vector (RV), and use them to define the socalled "basins of attraction" enabling modes' retrieval. Compared with previous approaches, this new technique not only enables reconstruction of AM/FM modes but also Dirac impulses, which is of great interest in many practical situations. Numerical experiments conducted with both synthetic and real data illustrate the effectiveness of the technique.

Magnetization reversal in bent nanofibers of different cross sections

“…Similar to quantum computers which are expected to solve a defined class of mathematical problems on very short time-scales, [1][2][3] a novel adaptive computation technology can be estimated to calculate different problems similar to the human brain and thus in a much more flexible and efficient way than common technology. [4][5][6] Besides developing hard-and software architecture further, however, only few attempts have been made to create completely new approaches for a new generation of computers.…”

Unsupervised classification of the spectrogram zeros with an application to signal detection and denoising

“…We compared the performance of the denoising strategy proposed in Sec. 7.2 with the DT method [19], an approach [7,37] based on synchrosqueezing and ridge detection (SST+RD), a simple hard thresholding (HT) [30] of the STFT based on the spectrogram, and a contoursbased reconstruction [29,34] (denoted "Contours").…”

Phonocardiogram Signal Denoising Based on Nonnegative Matrix Factorization and Adaptive Contour Representation Computation

“…The ACRC introduced in [16], [17] is computed on a lowpass filtered (LPF) version ofŝ 1 with a cutoff frequency 80Hz to enable not only the elimination of high-frequency noises but also a significant reduction of the computational cost of ACRC. From now on, we denoteŝ 1,LP F such a signal.…”

“…CPs are then linked to form so-called contours. To avoid both the issue of choosing θ and numerical instabilities, an adaptive technique, proposed in [17], defined CPs as the zeros of the projection of RV in a direction corresponding to the average of the orientation of RV over a squared neighborhood centered at the point of study. This resulted in a robust estimation of TF signatures for a wide class of MCSs.…”

“…Unfortunately, the procedure relied on a threshold, fixed a priori, thus limiting its adaptivity. In another direction, a novel TF method, called adaptive contour representation computation (ACRC), was proposed to adaptively retrieve and denoise the modes of multicomponent signals (MCSs) [15]- [17]. While this technique cannot efficiently deal with highly energetic noises which are common in PCG signals, our goal is here to show that to apply it to a low-pass version of the signal obtained after application of NMF-denoising results in improved denoising performance and adaptivity.…”