Underdetermined Convolutive Source Separation Using GEM-MU With Variational Approximated Optimum Model Order NMF2D

Al-Tmeme, Ahmed; Woo, Wai Lok; Dlay, S.S.; Gao, Bin

doi:10.1109/taslp.2016.2620600

Cited by 23 publications

(19 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These adverse factors include various noises such as Gaussian noise, muscle artifacts, power-line interference, and baseline wander. To suppress noise and to obtain a clean ECG signal, many different algorithms are proposed [ 11 , 12 , 13 ]. The commonly used ECG signal de-noising methods include the morphological filtering method, the adaptive filtering method, the wavelet-based method, and the empirical mode decomposition (EMD) method.…”

Section: Introductionmentioning

confidence: 99%

ECG Signal De-noising and Baseline Wander Correction Based on CEEMDAN and Wavelet Threshold

Yang

Luo

Tao

et al. 2017

Sensors

100

View full text Add to dashboard Cite

A novel electrocardiogram (ECG) signal de-noising and baseline wander correction method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and wavelet threshold is proposed. Although CEEMDAN is based on empirical mode decomposition (EMD), it represents a significant improvement of the original EMD by overcoming the mode-mixing problem. However, there has been no previous study on using CEEMDAN to de-noise ECG signals, to the authors’ best knowledge. In the proposed method, the original noisy ECG signal is decomposed into a series of intrinsic mode functions (IMFs) sorted from high to low frequency by CEEMDAN. Each IMF is then analyzed by the autocorrelation method to find out the first few high frequency IMFs containing random noise, and these IMFs should be de-noised by the wavelet threshold. The zero-crossing rate (ZCR) of all IMFs, including final residue, are computed, and the IMFs with ZCR less than a certain value are removed. Finally, the remaining IMFs are reconstructed to obtain the clean ECG signal. The proposed algorithm is validated through experiments using the MIT–BIH ECG databases, and the results show that the random noise in the ECG signal can be effectively suppressed, and at the same time the baseline wander can be corrected efficiently.

show abstract

Section: Introductionmentioning

confidence: 99%

ECG Signal De-noising and Baseline Wander Correction Based on CEEMDAN and Wavelet Threshold

Yang

Luo

Tao

et al. 2017

Sensors

100

View full text Add to dashboard Cite

show abstract

“…Let K ≥ I is known in advance, and {K i } I i=1 be a nontrivial partition of K = 1, ..., K. Following [16,17], a coefficient s i, f n is modeled as the sum of latent components c k, f n , such that…”

Section: The Nmf Source Modelmentioning

confidence: 99%

“…The NMF model can be used to efficiently exploit the low-rank nature of the speech spectrogram and its dependency across the frequencies. In some NMF-based methods [14][15][16][17], non-negative matrix factor two-dimensional deconvolution is an effective machine learning method in audio source separation field. In particular, in the convolutive frequency-domain model, the well-known permutation alignment problem cannot be solved without using additional a priori knowledge about the sources or the mixing filters.…”

Section: Introductionmentioning

confidence: 99%

Underdetermined Blind Source Separation Combining Tensor Decomposition and Nonnegative Matrix Factorization

Xie

Yang

et al. 2018

Symmetry

View full text Add to dashboard Cite

Underdetermined blind source separation (UBSS) is a hot topic in signal processing, which aims at recovering the source signals from a number of observed mixtures without knowing the mixing system. Recently, expectation-maximization algorithm shows a great potential in the UBSS. However, the final separation results depend strongly on the parameter initialization, leading to poor separation performance. In this paper, we propose an effective algorithm that combines tensor decomposition and nonnegative matrix factorization (NMF). In the proposed algorithm, we first employ tensor decomposition to estimate the mixing matrix, and NMF source model is used to estimate the source spectrogram factors. Then a series of iterations are derived to update the model parameters. At the same time, the spatial images of source signals are estimated with Wiener filters constructed from the learned parameters. Therefore, time-domain sources can be obtained through inverse short-time Fourier transform. Finally, plenty of experimental results demonstrate the effectiveness and advantages of our proposed algorithm over the compared algorithms.

show abstract

“…The PNMF algorithm adds constraints to the NMF algorithm, reduces an unknown variable, and increases the sparsity of the matrix [11]. Ahmed et al proposed an approximately optimal nonnegative matrix factor two-dimensional deconvolution (NMF2D) algorithm for separating underdetermined convolution integrals [12]. After that, they proposed a mono audio separation method based on multicomponent nonnegative matrix factor two-dimensional deconvolution (NMF2D) [13].…”

Section: Introductionmentioning

confidence: 99%

Multimode Process Monitoring Method Based on Multiblock Projection Nonnegative Matrix Factorization

Wang

Zhao

et al. 2020

Advances in Mathematical Physics

View full text Add to dashboard Cite

A multimode process monitoring method based on multiblock projection nonnegative matrix factorization (MPNMF) is proposed for traditional process monitoring methods which often adopt global model of data and ignore local information of data. Firstly, the training data set of each mode is partitioned by the complete link algorithm and the multivariate data space is divided into several subblocks. Then, the projection nonnegative matrix factorization (PNMF) algorithm is used to model each subspace of each mode separately. A joint probabilistic statistic index is defined to identify the running modes of the process data. Finally, the Bayesian information criterion (BIC) is used to synthesize the statistics of each subblock and construct a new statistic for process monitoring. The proposed process monitoring method is applied to the TE process to verify its effectiveness.

show abstract

Underdetermined Convolutive Source Separation Using GEM-MU With Variational Approximated Optimum Model Order NMF2D

Cited by 23 publications

References 40 publications

ECG Signal De-noising and Baseline Wander Correction Based on CEEMDAN and Wavelet Threshold

ECG Signal De-noising and Baseline Wander Correction Based on CEEMDAN and Wavelet Threshold

Underdetermined Blind Source Separation Combining Tensor Decomposition and Nonnegative Matrix Factorization

Multimode Process Monitoring Method Based on Multiblock Projection Nonnegative Matrix Factorization

Contact Info

Product

Resources

About