Blind Source Separation Research Articles

Multi-leakage localization in water supply pipes based on convolutional blind source separation

In acoustic-based methods for water pipeline leak localization, there are few studies on complex multi-leak localization since these sources mix with and affect each other during propagation. In this paper, a novel multi-leak localization method based on the convolutional blind source separation (CBSS) algorithm is proposed, in which the time delays are estimated by the quadrature filter of the demixing matrix or the cross-correlation function between the separated signals and the measured signals, and then the leakage locations are found by linear localization formula. In theory, this paper analyzes the basic principle of the algorithm for locating multiple leaks. In simulation results, the algorithm’s feasibility is verified. In addition, it is obtained that the location distribution of the leakage sources will affect the location of the peak of the delay, and that strong noise will reduce the applicability of the algorithm. In prototype experiments, the algorithm is verified to locate the leaks effectively. This paper is the first to apply CBSS to multi-leakage localization, providing a feasible solution to this complex problem.

Nonlinear spatio-temporal filter to reduce crosstalk in bipolar electromyogram

Objective. The wide detection volume of surface electromyogram (EMG) makes it prone to crosstalk, i.e. the signal from other muscles than the target one. Removing this perturbation from bipolar recordings is an important open problem for many applications. Approach. An innovative nonlinear spatio-temporal filter is developed to estimate the EMG generated by the target muscle by processing noisy signals from two bipolar channels, placed over the target and the crosstalk muscle, respectively. The filter is trained on some calibration data and then can be applied on new signals. Tests are provided in simulations (considering different thicknesses of the subcutaneous tissue, inter-electrode distances, locations of the EMG channels, force levels) and experiments (from pronator teres and flexor carpi radialis of 8 healthy subjects). Main results. The proposed filter allows to reduce the effect of crosstalk in all investigated conditions, with a statistically significant reduction of its root mean squared of about 20%, both in simulated and experimental data. Its performances are also superior to those of a blind source separation method applied to the same data. Significance. The proposed filter is simple to be applied and feasible in applications in which single bipolar channels are placed over the muscles of interest. It can be useful in many fields, such as in gait analysis, tests of myoelectric fatigue, rehabilitation with EMG biofeedback, clinical studies, prosthesis control.

An optimized approach to speech transcription using blind source separation and speech-to-text machine learning models

The use of speech-to-text transcription has a multitude of applications in various industries, including accessibility support, language processing, and automatic subtitling. In recent years, there has been greater interest in incorporating automatic speech source separation features to improve the accuracy and efficiency of transcription mechanisms. This paper aims to design a transcription mechanism that utilizes DUET algorithm to separate speech sources in a stereo setup. The separated sources are then transcribed into text using a machine learning model. The study evaluates the effectiveness of this approach using a dataset of speech recordings. The results of the study indicate high accuracy in speech separation and transcription, highlighting the potential of this approach for practical applications. However, the study also revealed potential issues with the mechanism, indicating the need for further exploration and refinement. These findings indicate the potential of the proposed approach for practical applications, and propose insight for further development and researches in this area.

Advancing real-time close captioning: blind source separation and transcription for hearing impairments

This project investigates the potential of integrating Blind Source Separation (DUET algorithm) and Automatic Speech Recognition (Wav2Vec2 model) for real-time, accurate transcription in multi-speaker scenarios. Specifically targeted towards improving accessibility for individuals with hearing impairments, the project addresses the challenging task of separating and transcribing speech from simultaneous speakers in various contexts. The DUET algorithm effectively separates individual voices from complex audio scenarios, which are then accurately transcribed into text by the machine learning model, Wav2Vec2. However, despite their remarkable capabilities, both techniques present limitations, particularly when handling complicated audio scenarios and in terms of computational efficiency. Looking ahead, the research suggests incorporating a feedback mechanism between the two systems as a potential solution for these issues. This innovative mechanism could contribute to a more accurate and efficient separation and transcription process by enabling the systems to dynamically adjust to each other's outputs. Nevertheless, this promising direction also brings with it new challenges, particularly in terms of system complexity, defining actionable feedback parameters, and maintaining system efficiency in real-time applications.

Two new methods for the estimation and interpretation of the range of feasible profiles in multivariate curve resolution and their implications to analytical chemistry

AbstractTwo new models have been recently introduced for studying the remaining rotational ambiguity in the bilinear decomposition of matrix data. One of the models is N‐BANDS, which yields two extreme profiles per sample component, corresponding to maximum or minimum signal contribution function or relative component area under its concentration profile. It is highly useful for computing the relative root mean square error due to rotational ambiguity in estimated analyte concentrations (RMSERA), which numerically quantifies the impact of the phenomenon in terms of prediction uncertainty. Since N‐BANDS successfully consider the presence of instrumental noise in the data, it is extremely useful for the analysis of real data sets. The other model is SW‐N‐BANDS, which is similar to N‐BANDS, but is applied in a sensor wise manner, that is, computing the maximum and minimum intensity value at each sensor. It provides the boundaries of the full set of feasible profiles, and helps to better understand the behavior of a given component under the application of several constraints. Both models are described in light of both simulations and experimental data, illustrating their main characteristics of importance to analytical chemistry studies.

Dynamic Gain Decomposition Reveals Functional Effects of Dendrites, Ion Channels, and Input Statistics in Population Coding.

Modern, high-density neuronal recordings reveal at ever higher precision how information is represented by neural populations. Still, we lack the tools to understand these processes bottom-up, emerging from the biophysical properties of neurons, synapses, and network structure. The concept of the dynamic gain function, a spectrally resolved approximation of a population's coding capability, has the potential to link cell-level properties to network-level performance. However, the concept is not only useful but also very complex because the dynamic gain's shape is co-determined by axonal and somato-dendritic parameters and the population's operating regime. Previously, this complexity precluded an understanding of any individual parameter's impact. Here, we decomposed the dynamic gain function into three components corresponding to separate signal transformations. This allowed attribution of network-level encoding features to specific cell-level parameters. Applying the method to data from real neurons and biophysically plausible models, we found: (1) The encoding bandwidth of real neurons, approximately 400 Hz, is constrained by the voltage dependence of axonal currents during early action potential initiation. (2) State-of-the-art models only achieve encoding bandwidths around 100 Hz and are limited mainly by subthreshold processes instead. (3) Large dendrites and low-threshold potassium currents modulate the bandwidth by shaping the subthreshold stimulus-to-voltage transformation. Our decomposition provides physiological interpretations when the dynamic gain curve changes, for instance during spectrinopathies and neurodegeneration. By pinpointing shortcomings of current models, it also guides inference of neuron models best suited for large-scale network simulations.

Multi-stream signals separation based on space-time-isomeric (SPATIO) array using metasurface antennas

Multi-stream signals separation based on space-time-isomeric (SPATIO) array using metasurface antennas

A general dual-pathway network for EEG denoising.

Scalp electroencephalogram (EEG) analysis and interpretation are crucial for tracking and analyzing brain activity. The collected scalp EEG signals, however, are weak and frequently tainted with various sorts of artifacts. The models based on deep learning provide comparable performance with that of traditional techniques. However, current deep learning networks applied to scalp EEG noise reduction are large in scale and suffer from overfitting. Here, we propose a dual-pathway autoencoder modeling framework named DPAE for scalp EEG signal denoising and demonstrate the superiority of the model on multi-layer perceptron (MLP), convolutional neural network (CNN) and recurrent neural network (RNN), respectively. We validate the denoising performance on benchmark scalp EEG artifact datasets. The experimental results show that our model architecture not only significantly reduces the computational effort but also outperforms existing deep learning denoising algorithms in root relative mean square error (RRMSE)metrics, both in the time and frequency domains. The DPAE architecture does not require a priori knowledge of the noise distribution nor is it limited by the network layer structure, which is a general network model oriented toward blind source separation.

Raman and photoluminescence signal separation in Raman hyperspectral imagery including noise reduction

AbstractRaman hyperspectral imaging (RHSI) is a valuable tool for gaining crucial information about the chemical composition of materials. However, obtaining clear Raman signals is not always a trivial task. Raw Raman signals can be susceptible to photoluminescence interference and noise. Hence, the preprocessing of RHSI is a required step for an effective and reliable chemical analysis. The main challenge is splitting the measured RHSI into separate Raman photoluminescence signals. Since no golden‐standard exists, it is non‐trivial to validate the correctness of the separated signals. While current state‐of‐the‐art preprocessing methods are effective, they require expert knowledge and involve unintuitive hyperparameters. Current approaches also lack generalizability, requiring extensive hyperparameter tuning on a case‐by‐case basis, while even then results are not always as expected. To this end, this work proposes a novel iterative RHSI preprocessing pipeline for splitting raw Raman signals and noise removal based on linear spline and radial basis function regression (IlsaRBF). The proposed method involves hyperparameters based on the physical properties of Raman spectroscopy, making them intuitive to use. This leads to more robust and stable hyperparameters, reducing the necessity for extensive hyperparameter tuning. A thorough evaluation shows that the proposed method outperforms the current state‐of‐the‐art. Additionally, a cosmic ray identification and removal algorithm (CRIR) and dynamic PCA for noise reduction are introduced. A standalone tool containing our proposed methods is provided, making RHSI preprocessing available to a broader audience, aiding further research and advancements in the field of Raman spectroscopy.

Seismic resolution enhancement with variational modal-based fast-matching pursuit decomposition

Enhancing vertical resolution and signal-to-noise ratio (S/N) is a key objective in seismic data processing. Considering the underground medium is inhomogeneous and incompletely elastic, seismic wave energy attenuation occurs during underground propagation, which has a significant impact on seismic data resolution and S/N. Traditional fast-matching pursuit (FMP) algorithms make it difficult to separate valid signals and noise effectively while reconstructing the noisy signals. Therefore, an improved FMP algorithm that combines the variational-mode decomposition (VMD) strategy is developed. The VMD algorithm is used to obtain intrinsic mode functions with varying amplitudes, frequencies, and center times. It can achieve a multiscale decomposition of nonstationary seismic data. Based on the intrinsic mode functions of different scales, the FMP algorithm can reconstruct prior information of the amplitude, frequency, and center time of valid signals and noise signals in the mode functions. Thus, the high-resolution sparse representation of intrinsic mode functions is achieved. The numerical results indicate that our method not only separates the effective signal and noise but also preserves the valid signal as much as possible. In addition, the feasibility of the method is further verified by field exploration data. The results indicate that this strategy can enhance the resolution of seismic data while restoring the attenuated energy using multiscale seismic data.

Uv–Vis molecular spectral resolution of multi-level mixture design-fingerprints of yerba mate leaves by chemometric methods

This study proposes the use of multivariate curve resolution with alternating least squares (MCR-ALS) to comprehensively analyze mixture design-fingerprints in yerba mate leaves, investigating the influence of multi-level factors such as sexual dimorphism (male and female plants) and cultivation systems (monoculture and agroforestry). Plant extracts were obtained according to a mixture design with ethanol, dichloromethane, and hexane and their binary and ternary combinations. Through simultaneous resolution of the diverse profiles obtained using these different solvents, seven components were successfully identified, collectively explaining 98.57% of the data variance. Each component was associated with distinct absorption bands indicative of their chemical composition with compounds such as chlorogenic acids, xanthophylls, catechin, caffeine, kaempferol, lycopene, carotene, and chlorophyll. The MCR-ALS resolved profiles revealed significant variations in component concentrations across experimental conditions, facilitating the determination of the effects of sexual dimorphism and cultivation systems on yerba mate metabolism. The ANOVA-simultaneous component analysis (ASCA) highlighted the chemical differences of yerba mate leaves due to the cultivation system. Binary and ternary mixture extractants resulted to be more effective than pure solvents in determining the significance of effects on the chemical composition of yerba mate. Partial least squares discriminant analysis (PLS-DA) demonstrated the most distinct metabolic profiles of yerba mate leaves under monoculture and agroforestry conditions, with specific components playing crucial roles in their discrimination. This study shows the robustness of the MCR-ALS method in resolving complex chemical data, offering valuable insights into the diverse chemical composition of yerba mate under different experimental conditions. The congruence of results across various analyses reinforced the reliability and validity of the methodology, emphasizing its potential use for exploring the composition and properties of natural products.

Study on weak sound signal separation and pattern recognition under strong background noise in marine engineering

The extraction of weak acoustic signals under strong background noise is of great significance in the applications of target identification and localization. In this paper, the pulse signal with high randomness is set as the weak signal sound source, random noise and sine sound are used as the background noise. Under the condition of a signal-to-noise ratio of −20 dB, combined with blind source separation and neural network methods, the collected observation signals are subjected to weak sound signal separation and recognition research. The optimization method of centralization and scaling processing is used to eliminate the unfavorable influence of the uncertainty of the separated signal amplitude caused by the blind source separation method on the pattern recognition. The recognition result is verified by the combination of “weak impulse acoustic signal” and “random noise signal,” and the output vector (0.99 0.01 0.01) approaches (1 0 0), which is recognized as impulse acoustic signal. By combining blind source separation and neural network methods, the separation and identification of weak pulse signals under the condition of a signal-to-noise ratio of −20 dB can be achieved.

Spatial Blind Source Separation in the Presence of a Drift

Multivariate measurements taken at different spatial locations occur frequently in practice. Proper analysis of such data needs to consider not only dependencies on-sight but also dependencies in and in-between variables as a function of spatial separation. Spatial Blind Source Separation (SBSS) is a recently developed unsupervised statistical tool that deals with such data by assuming that the observable data is formed by a linear latent variable model. In SBSS the latent variable is assumed to be constituted by weakly stationary random fields which are uncorrelated. Such a model is appealing as further analysis can be carried out on the marginal distributions of the latent variables, interpretations are straightforward as the model is assumed to be linear, and not all components of the latent field might be of interest which acts as a form of dimension reduction. The weakly stationarity assumption of SBSS implies that the mean of the data is constant for all sample locations, which might be too restricting in practical applications. Therefore, an adaptation of SBSS that uses scatter matrices based on differences was recently suggested in the literature. In our contribution we formalize these ideas, suggest a novel adapted SBSS method and show its usefulness on synthetic data and illustrate its use in a real data application.

Multivariate Multiscale Higuchi Fractal Dimension and Its Application to Mechanical Signals

Fractal dimension, as a common nonlinear dynamics metric, is extensively applied in biomedicine, fault diagnosis, underwater acoustics, etc. However, traditional fractal dimension can only analyze the complexity of the time series given a single channel at a particular scale. To characterize the complexity of multichannel time series, multichannel information processing was introduced, and multivariate Higuchi fractal dimension (MvHFD) was proposed. To further analyze the complexity at multiple scales, multivariate multiscale Higuchi fractal dimension (MvmHFD) was proposed by introducing multiscale processing algorithms as a technology that not only improved the use of fractal dimension in the analysis of multichannel information, but also characterized the complexity of the time series at multiple scales in the studied time series data. The effectiveness and feasibility of MvHFD and MvmHFD were verified by simulated signal experiments and real signal experiments, in which the simulation experiments tested the stability, computational efficiency, and signal separation performance of MvHFD and MvmHFD, and the real signal experiments tested the effect of MvmHFD on the recognition of multi-channel mechanical signals. The experimental results show that compared to other indicators, A achieves a recognition rate of 100% for signals in three features, which is at least 17.2% higher than for other metrics.

Image quality improvement in multi-mode laser ultrasound imaging using reconstructed synthetic aperture focusing technique

The synthetic aperture focusing technique (SAFT) is used as a laser ultrasonic imaging algorithm for non-destructive internal defect testing, and multi-mode time domain SAFT (MMT-SAFT) imaging is developed to enhance the imaging capability of defect detection. However MMT-SAFT algorithm has the problem of artifacts caused by the laser ultrasound (LU) modes waves. Thus to solve this problem, an automatic imaging technique based on MMT-SAFT called Reconstructed synthetic aperture focusing technique (Re-SAFT) is proposed. This technique consists of two steps: a preliminary MMT-SAFT is firstly performed to locate the defect position tentatively and futher separate LU modes wave signals using the defect-reflected signal; then a secondary reconstructed MMT-SAFT based on the priori defect location information is achieved for artifacts removal. In addition, directivity patterns of LU modes wave have been used in Re-SAFT to achieve more accurate defect imaging in the second step. Results indicate that by using Re-SAFT, the typical artifact caused by laser ultrasonic waves can be removed and the imaging quality can be significantly improved from 19.67 to 33.85 dB. Further, the consumed time of Re-SAFT is economized by optimizing the imaging pixel size without compromising the image quality. Therefore, Re-SAFT has potential applications in industrial laser ultrasonic testing.

Comprehensive Separation Algorithm for Single-Channel Signals Based on Symplectic Geometry Mode Decomposition.

This paper aims to explore the difficulty of obtaining source signals from complex mixed signals and the issue that the FastICA algorithm cannot directly decompose the received single-channel mixed signals and distort the signal separation in low signal-to-noise environments. Thus, in this work, a comprehensive single-channel mixed signal separation algorithm was proposed based on the combination of Symplectic Geometry Mode Decomposition (SGMD) and the FastICA algorithm. First, SGMD-FastICA uses SGMD to decompose single-channel mixed signals, and then it uses the Pearson correlation coefficient to select the Symplectic Geometry Components that exhibit higher correlation coefficients with the mixed signals. Then, these components are expanded with the single-channel mixed signals into virtual multi-channel signals and input into the FastICA algorithm. The simulation results show that the SGMD algorithm could eliminate noise interference while keeping the raw time series unchanged, which is achievable through symplectic geometry similarity transformation during the decomposition of mixed signals. Comparative experiment results also show that compared with the EMD-FastICA and VMD-FastICA, the SGMD-FastICA algorithm has the best separation effect for single-channel mixed signals. The SGMD-FastICA algorithm represents an improved solution that addresses the limitations of the FastICA algorithm, enabling the direct separation of single-channel mixed signals, while also addressing the challenge of proper signal separation in noisy environments.

Enhanced efficiency of MS/MS all-ion fragmentation for non-targeted analysis of trace contaminants in surface water using multivariate curve resolution and data fusion

Data-independent acquisition–all-ion fragmentation (DIA-AIF) mode of mass spectrometry can facilitate wide-scope non-target analysis of contaminants in surface water due to comprehensive spectral identification. However, because of the complexity of the resulting MS2 AIF spectra, identifying unknown pollutants remains a significant challenge, with a significant bottleneck in translating non-targeted chemical signatures into environmental impacts. The present study proposes to process fused MS1 and MS2 data sets obtained from LC-HRMS/MS measurements in non-targeted AIF workflows on surface water samples using multivariate curve resolution-alternating least squares (MCR-ALS). This enables straightforward assignment between precursor ions obtained from resolved MS1 spectra and their corresponding MS2 spectra. The method was evaluated for two sets of tap water and surface water contaminated with 14 target chemicals as a proof of concept. The data set of surface water samples consisting of 3506 MS1 and 2170 MS2 AIF mass spectral features was reduced to 81 components via a fused MS1-MS2 MCR model that describes at least 98.8% of the data. Each component summarizes the distinct chromatographic elution of components together with their corresponding MS1 and MS2 spectra. MS2 spectral similarity of more than 82% was obtained for most target chemicals. This highlights the potential of this method for unraveling the composition of MS/MS complex data in a water environment. Ultimately, the developed approach was applied to the retrospective non-target analysis of an independent set of surface water samples.Graphical abstract

Research on a Resource Modeling and Power Prediction Method Based on Virtual Aggregation

Distributed resources at a grid’s end cannot upload operational power data to local centers due to data transmission and privacy issues. This leaves the centers with incomplete information, thus impacting decision making. This paper presents a virtual aggregation-based model for such scenarios. We define four virtual aggregate types based on resource response characteristics. Using characteristic coefficients, we identify these aggregates’ categories and proportions from bus power. To address blind source separation in single-channel power signals, we apply the Ensemble Empirical Mode Decomposition-Fast Independent Component Analysis (EEMD-FastICA) method. This helps extract and analyze bus power, thereby deriving power curves for different aggregates. Moreover, we use a graph convolutional network to explore how factors like date, time, weather, and pricing intertwine with aggregate power. We develop a predictive model with an advanced SpatioTemporal Graph Convolutional Network (STGCN), thus facilitating proactive power forecasting for virtual aggregates. Case studies show our method’s efficacy in extracting power curves under limited information, with the STGCN ensuring accurate, forward-looking predictions.

Early pseudoprogression and progression lesions in glioblastoma patients are both metabolically heterogeneous.

The standard treatment in glioblastoma includes maximal safe resection followed by concomitant radiotherapy plus chemotherapy and adjuvant temozolomide. The first follow-up study to evaluate treatment response is performed 1 month after concomitant treatment, when contrast-enhancing regions may appear that can correspond to true progression or pseudoprogression. We retrospectively evaluated 31 consecutive patients at the first follow-up after concomitant treatment to check whether the metabolic pattern assessed with multivoxel MRS was predictive of treatment response 2 months later. We extracted the underlying metabolic patterns of the contrast-enhancing regions with a blind-source separation method and mapped them over the reference images. Pattern heterogeneity was calculated using entropy, and association between patterns and outcomes was measured with Cramér's V. We identified three distinct metabolic patterns-proliferative, necrotic, and responsive, which were associated with status 2 months later. Individually, 70% of the patients showed metabolically heterogeneous patterns in the contrast-enhancing regions. Metabolic heterogeneity was not related to the regions' size and only stable patients were less heterogeneous than the rest. Contrast-enhancing regions are also metabolically heterogeneous 1 month after concomitant treatment. This could explain the reported difficulty in finding robust pseudoprogression biomarkers.

Deep signal separation for adaptive estimation of instantaneous phase from vibration signals

Accurate extraction of instantaneous phase information is an essential step in condition monitoring and intelligent diagnosis of rotating machinery under variable speed conditions, which will directly affect the reliability of the tacho-less order tracking (TLOT) results. At present, instantaneous phase information extraction techniques are mainly divided into time–frequency analysis and signal decomposition. These methods are constrained because they are not available to extract harmonic components adaptively. Therefore, it is important to develop an intelligent model to accurately separate harmonic component with strict physical meaning for TLOT. In order to address the aforementioned issues, a deep binary mask (BM) signal separation model is presented for rotating machinery TLOT. The deep BM signal separation method can adaptively separate the fundamental harmonic component of vibration signals without prior knowledge. The outcomes indicate that the deep BM signal separation model is more productive and flexible in terms of accuracy and self-adaptation compared to some advanced TLOT algorithms.