Independent Component Analysis Algorithm Research Articles

Second-Order Approximation of Minimum Discrimination Information in Independent Component Analysis

Independent Component Analysis (ICA) is intended to recover the mutually independent sources from their linear mixtures, and <inline-formula><tex-math notation="LaTeX">$FastICA$</tex-math></inline-formula> is one of the most successful ICA algorithms. Although it seems reasonable to improve the performance of <inline-formula><tex-math notation="LaTeX">$FastICA$</tex-math></inline-formula> by introducing more nonlinear functions to the negentropy estimation, the original fixed-point method (approximate Newton method) in <inline-formula><tex-math notation="LaTeX">$FastICA$</tex-math></inline-formula> degenerates under this circumstance. To alleviate this problem, we propose a novel method based on the second-order approximation of minimum discrimination information (MDI). The joint maximization in our method is consisted of minimizing single weighted least squares and seeking unmixing matrix by the fixed-point method. Experimental results validate its efficiency compared with other popular ICA algorithms.

Stochastic Higher-Order Independent Component Analysis for Hyperspectral Dimensionality Reduction

Hyperspectral imaging is a remote sensing technique that measures the spectrum of each pixel in the image of a scene. Various methods have been developed to reduce the spectral dimension of hyperspectral images in order to facilitate their analysis. Independent Component Analysis (ICA) is a class of algorithms which extract statistically independent features. FastICA, is one of the most used ICA algorithms because it is simple and fast. However, FastICA often finds irrelevant stationary points (e.g., minima instead of maxima) and is not scalable as it uses at each iteration the whole set of pixels. In this paper, we present a new stochastic algorithm, called SHOICA, which smoothly approximates the non-convex loss functions of ICA using higher-order Taylor minorizers. Because SHOICA guarantees ascent of its objective function, it identifies (local) maxima. Moreover, because SHOICA is stochastic, it facilitates minibatching and thus is scalable and appropriate for large datasets. Hence, we show that our method is faster and the extracted features are better than those of FastICA. The quality of features extracted, as well as time and epochs required by FastICA and SHOICA are compared practically, on dimensionality reduction and classification tasks of real hyperspectral images.

Extraction of Reduced Fault Subspace Based on KDICA and Its Application in Fault Diagnosis

Independent component analysis (ICA) is a commonly used non-Gaussian process fault diagnosis method. A fault detection algorithm of kernel dynamic ICA (KDICA) has been proposed for the non-Gaussian process with dynamic and nonlinear characteristics. However, a lack of studies tackling the fault reconstruction and fault diagnosis algorithm exists. Hence, a fault reconstruction model based on KDICA is proposed in this article. In this model, a reduced fault subspace extraction method is proposed. It consists in dividing the fault subspace into the kernel dynamic independent component reduced fault subspace and the residual reduced fault subspace (RRFS). Based on the RRFS, a fault diagnosis approach is designed for online process monitoring. Using the proposed method, the computational complexity can be efficiently reduced and the specific fault type can be accurately identified. The Tennessee–Eastman process is used to verify the feasibility and efficiency of the proposed method and its fault diagnosis application.

FiCA: A Fixed-Point Custom Architecture FastICA for Real-Time and Latency-Sensitive Applications.

Independent Component Analysis (ICA) is a common method exploited in different biomedical signal processing applications, especially in noise removal of electroencephalography (EEG) signals. Among different existing ICA algorithms, FastICA is a popular method with less complexity, which makes it more suitable for practical implementation. However, and due to its inherent computationally intensive nature, development of a custom FastICA hardware is the best way to utilize it in high-performance real-time applications. On the other hand, development of a custom hardware in a fixed-point manner is also a complex and challenging task due to the algorithm's iterative nature. Moreover, the algorithm intrinsically suffers from some convergence problems which prevents to be practically exploited in latency-sensitive applications. In this paper, a fixed-point fully customized, scalable, and high-performance FastICA processor architecture has been presented. The proposed architecture is developed in an algorithm-aware manner to mitigate the inherent FastICA algorithmic failures. The synthesis results in a 90 nm technology show that the design proposes a computational time of 0.32 ms to perform an 8-channel ICA with a frequency of 555 MHz. The performance-related measurements prove that its normalized throughput is 10 times more, compared to the closest rival.

Discussion on "distributional independent component analysis for diverse neuroimaging modalities" by Ben Wu, Subhadip Pal, Jian Kang, and Ying Guo.

I applaud the authors on their innovative generalized independent component analysis (ICA) framework for neuroimaging data. Although ICA has enjoyed great popularity for the analysis of functional magnetic resonance imaging (fMRI) data, its applicability to other modalities has been limited because standard ICA algorithms may not be directly applicable to a diversity of data representations. This is particularly true for single-subject structural neuroimaging, where only a single measurement is collected at each location in the brain. The ingenious idea of Wu etal. (2021) is to transform the data to a vector of probabilities via a mixture distribution with K components, which (following a simple transformation to ) can be directly analyzed with standard ICA algorithms, such as infomax (Bell and Sejnowski, 1995) or fastICA (Hyvarinen, 1999). The underlying distribution forming the basis of the mixture is customized to the particular modality being analyzed. This framework, termed distributional ICA (DICA), is applicable in theory to nearly any neuroimaging modality. This has substantial implications for ICA as a general tool for neuroimaging analysis, with particular promise for structural modalities and multimodal studies. This invited commentary focuses on the applicability and potential of DICA for different neuroimaging modalities, questions around details of implementation and performance, and limitations of the validation study presented in thepaper.

Classification of Alcoholic and Non-Alcoholic EEG Signals Based on Sliding-SSA and Independent Component Analysis

Alcoholism is a widely affected disorder that leads to critical brain deficiencies such as emotional and behavioural impairments. One of the prominent sources to detect alcoholism is by analysing Electroencephalogram (EEG) signals. Previously, most of the works have focused on detecting alcoholism using various machine and deep learning algorithms. This paper has used a novel algorithm named Sliding Singular Spectrum Analysis (S-SSA) to decompose and de-noise the EEG signals. We have considered independent component analysis (ICA) to select the prominent alcoholic and non-alcoholic components from the preprocessed EEG data. Later, these components were used to train and test various machine learning models like SVM, KNN, ANN, GBoost, AdaBoost and XGBoost to classify alcoholic and non-alcoholic EEG signals. The sliding SSA-ICA algorithm helps in reducing the computational time and complexity of the machine learning models. To validate the performance of the ICA algorithm, we have compared the computational time and accuracy of ICA with its counterpart, like principal component analysis (PCA) and linear discriminant analysis (LDA). The proposed algorithm is tested on a publicly available UCI alcoholic EEG dataset. To verify the performance of machine learning models, we have calculated various metrics like accuracy, precision, recall and F1 score. Our work reported the highest accuracy of 98.97% with the XGBoost classifier. The validation of the proposed method is done by comparing the classification metrics with the latest state-of-the-art works.

Convolutive Complex-Valued Independent Component Analysis for Nonlinear Radar Signal Processing and Maritime Weak Target Detection

This paper first establishes a new complex independent component analysis (cICA) algorithm based on the spatiotemporal extension of complex-valued entropy bound minimization (CEBM) to separate received complex-valued radar signals. Next, we propose a new cICA-based detector with an open structure to find Swerling model targets, lognormal targets, and sea-surface small floating targets in coherent high-resolution maritime surveillance radars. The detector encountered three major problems when adopting cICA for detection and solved them using three effective suggestions. After performing cICA on the time series received by the radar, we obtained two different sources. Using the first and second theoretical and empirical moment estimates of the K-distribution, the target was selected between these two output source signals. Detector performance was verified quantitatively and qualitatively using the real-life IPIX radar database. Comprehensive experiments on this database with synthetic injected targets showed promising results. The computational time and sample size dependency of the proposed cICA algorithm were also discussed. Finally, a comparison of the detector with several novel detectors for detecting sea-surface floating small targets of the IPIX radar database demonstrated the proposed detector’s superiority.

An improved feature extraction algorithms of EEG signals based on motor imagery brain-computer interface

The electroencephalogram (EEG) signals based on the Brian-computer Interface (BCI) equipment is weak, non-linear, non-stationary and time-varying, so an effective feature extraction method is the key to improving the recognition accuracy. Electrooculogram and electrocardiogram artifacts are common noises in the process of EEG signals acquisition, it seriously affects the extraction of useful information. This paper proposes a processing method on EEG signals by combing independent component analysis (ICA), wavelet transform (WT) and common spatial pattern (CSP). First, the independent component analysis algorithm is used to break the EEG signals into independent components; and then these independent components are decomposed by WT to obtain the wavelet coefficient of each independent source. The soft and hard compromise threshold function is used to process the wavelet packet coefficients. Then the CSP algorithm is used to extract the features of the denoised EEG data. Finally, four common classification algorithms are used for classification to verify the effectiveness of the improved algorithm. The experimental results show that the EEG signals processed by the proposed method has obvious advantages in identify and remove electrooculogram (EOG) and electrocardiogram (ECG) artifacts, meanwhile, it can preserve the neural activity that is missed in the noise component. Cross-comparison experiments also proved that the proposed method has higher classification accuracy than other algorithms.

Fractional Order Butterworth Filter for Fetal Electrocardiographic Signal Feature Extraction

The non-invasive Fetal Electrocardiogram (FECG) signal has become a significant method for monitoring the fetus's physiological conditions, extracted from the Abdominal Electrocardiogram (AECG) during pregnancy. The current techniques are limited during delivery for detecting and analyzing fECG. The non - intrusive fECG recorded from the mother's abdomen is contaminated by a variety of noise sources, can be a more challenging task for removing the maternal ECG. These contaminated noises have become a major challenge during the extraction of fetal ECG is managed by uni-modal technique. In this research, a new method based on the combination of Wavelet Transform (WT) and Fast Independent Component Analysis (FICA) algorithm approach to extract fECG from AECG recordings of the pregnant woman is proposed. Initially, preprocessing of a signal is done by applying a Fractional Order Butterworth Filter (FBWF). To select the Direct ECG signal which is characterized as a reference signal and the abdominal signal which is characterized as an input signal to the WT, the cross-correlation technique is used to find the signal with greater similarity among the available four abdominal signals. The model performance of the proposed method shows the most frequent similarity of fetal heartbeat rate present in the database can be evaluated through MAE and MAPE is 0.6 and 0.041209 respectively. Thus the proposed methodology of de-noising and separation of fECG signals will act as the predominant one and assist in understanding the nature of the delivery on further analysis.

Peak-Load-Regulation Nuclear Power Unit Fault Diagnosis Using Thermal Sensors Combined with Improved ICA-RF Algorithm.

Owing to the fact that large-scale peak-load-regulation nuclear power turbine units’ thermal signal is greatly influenced by background noise and has non-stationary and nonlinear characteristics, this paper proposes a new fault diagnosis method for thermal sensors based on an improved independent component analysis (Improved-ICA) algorithm and random forest (RF) algorithm. This method is based on independent component analysis (ICA), which is not capable of extracting components independently. Therefore, we propose the use of the maximum approximate information negative entropy optimization model in order to improve the ICA algorithm’s independent principal component extraction ability and obtain better non-Gaussian physical source signal separation results. The improved ICA algorithm is used for the blind source separation of the thermal parameters of peak-load-regulation nuclear power units. A series of stationary physical source functions and a series of non-stationary noise signals are obtained. Then, according to the specific signal format and data volume of the nuclear power parameter signal, the network parameters of the random forest algorithm are determined, giving rise to the fault diagnosis model. Finally, the real-time operation data of an 1121 MW nuclear power unit are used to complete the training and fault diagnosis of the random forest network and analyze the diagnosis results. The results indicate that the model can effectively mine the abnormal sample points of thermal parameters and classify the fault type of the thermal sensor during peak load operation of the nuclear power unit. The accuracy rate is found to be at the threshold of 99%.

RETRACTED ARTICLE: Application of music in relief of driving fatigue based on EEG signals

In order to solve the problem of traffic accidents caused by fatigue driving, the research of EEG signals is particularly important, which can timely and accurately determine the fatigue state and take corresponding measures. Effective fatigue improvement measures are an important research topic in the current scientific field. The purpose of this article is to use EEG signals to analyze fatigue driving and prevent the dangers and injuries caused by fatigue driving. We designed the electroencephalogram (EEG) signal acquisition model to collect the EEG signal of the experimenter, and then removed the noise through the algorithm of Variational Mode Decomposition (VMD) and independent component analysis (ICA). On the basis of in-depth analysis and full understanding, we learned about the EEG signal of the driver at different driving times and different landscape roads, and provided some references for the study of music in relieving driving fatigue. The results of the study show that in the presence of music, the driver can keep the EEG signal active for more than 2 h, while in the absence of music, the driver’s EEG signal is active for about 1.5 h. Under different road conditions, the driver’s EEG signal activity is not consistent. The β wave and (α + θ)/β ratio of the driver in mountainous roads and grassland road landscape environments are highly correlated with driving time, and β wave is negatively correlated with driving time, and (α + θ)/β is positively correlated with driving time. In addition, the accumulation of changes in the two indicators is also strongly correlated with driving time.

Novel Process Monitoring Approach Enhanced by a Complex Independent Component Analysis Algorithm with Applications for Wastewater Treatment

Novel Process Monitoring Approach Enhanced by a Complex Independent Component Analysis Algorithm with Applications for Wastewater Treatment

Developing an attack detection framework for wireless sensor network‐based healthcare applications using hybrid convolutional neural network

AbstractAttack detection is the major issue in healthcare‐based wireless sensor networks (H‐WSNs). Due to their low processing speed, very low storage space, poor attack detection rate, longer deployment time, poor communication range, and reduced energy, H‐WSNs are subjected to difficult implementation and have their own limitations. To tackle these issues, we have presented a hybrid deep learning model using convolutional neural network and long short term memory (HDMCL) for attack detection in H‐WSN. This research is divided into three steps: preprocessing, dimensionality reduction, and classification (attack detection). At first, the raw input data (patient's health data) is preprocessed using the one‐hot encoding method. Next, the modified Huber independent component analysis based squirrel search algorithm (MHICA‐SSA) effectively reduces the data dimensionality in which the MHICA‐SSA is the amalgamation of both modified Huber independent component analysis and squirrel search algorithm. The novel algorithm designed overcomes the complexities associated with existing techniques such as the curse of dimensionality problem, improves the parameter interpretation, minimizes the time and storage space, minimizes space complexity, and enhances the attack detection accuracy of the convolutional neural network‐based long short‐term memory (CNN‐LSTM). After that, the deep learning CNN‐LSTM model is utilized for normal, black hole, and gray hole attack detection. The NS2.34 network simulator implements the proposed work thereby the proposed work efficiency is validated using different performance measures such as packet delivery ratio, false alarm rate, network lifetime, energy consumption, throughput, attack detection time, and rate. The proposed work performance is evaluated using different existing methods such as fish swarm optimization based particle swarm optimization, intelligent opportunistic routing algorithm, Jensen‐Shannon divergence‐based independent component analysis, and radio frequency identification based WSN. The throughput of the proposed work is increased up to 3.2 times and the network lifetime is increased up to 4 times when compared to the existing techniques.

Decoupling of Multivariable Control Systems Using Mod-ICA / Dissociação de sistemas de controle multivariável usando Mod-ICA

Sensors usage in process control systems is of vital importance for industrial plants proper operation and monitoring. In turn, process signals may have interference from other sources, and in some cases, it is not possible to observe the individual signs of the sources directly. Because of this, techniques of processing and separation of signals has been used to extract the information from the sources contained in mixed signals. The main signal separation techniques are associated with the technique ICA (Independent Component Analysis), which has undergone significant evolution since its creation in the 80’s. This growth had also received contributions from PCA technique (Principal Component Analysis), and the development of computational processing power. However, these methods have two fundamental problems, which are the deviation of amplitude, and phase change, with issues such as limiting its use in control systems. Thus, this work aims to present a solution to the amplitude problem of ICA's techniques for use in obtaining uncoupling of multivariable systems. The proposed correction, based on the stage of whitening of ICA algorithms, generating the technique MOD-ICA, has been used as an alternative to the breaking of the correlation between variables of multivariate systems, with the goal of achieving the decoupling of MIMO (Multiple Input Multiple Output) systems. In the study case, proposed in this work, it was observed a better estimation of the parameters of decoupling models for the control system. The variables, after using the ICA modified technique, are independent and do not present the influence of disturbances arising from other variables in the process, resulting in a more robust control system for process variations.

Random Forest (RF) based identification of rice powder mixture using terahertz spectroscopy

Adulteration is a severe problem in agriculture field. It may cause some dishonest traders in the marketing side. For their own wish, dishonest traders make adulteration products for higher profit. In consumer side, agro product adulterations provide some unhealthy situation. It may cause severe health problem and also some incurable disease. Its right time, need to find an adulteration various level of agro products. For detecting rice adulteration, to collect spectral data for mixed rice powder (low quality rice powder and high quality rice powder), and a Fast Independent Component Analysis (FICA) algorithm is used to extract valuable information. Then, apply Random Forest Classifier algorithm for classifying purpose, whether the mixture is low quality or high quality. It achieves high accuracy with a single model and also easy to implement.

Mechanical Compound Fault Analysis Method Based on Shift Invariant Dictionary Learning and Improved FastICA Algorithm

For mechanical compound fault, it is of great significance to employ the vibration signal of a single-channel compound fault to analyze and realize the separation of multiple fault sources, which is essentially the problem of single-channel blind source separation. Shift invariant K-means singular value decomposition (shift invariant K-SVD) dictionary learning is suitable to extract the periodic and repeated fault features of a rotating machinery fault, hence in this article a single-channel compound fault analysis method is put forward which combines shift invariant K-SVD with improved fast independent component analysis (improved FastICA) algorithm. Firstly, based on single-channel compound fault signal, the shift invariant K-SVD algorithm can be used for learning multiple latent components that can be constructed as a virtual multi-channel signal. Then the improved FastICA algorithm is utilized to realize the separation of multiple fault source signals. With regard to the FastICA algorithm, the third-order convergence Newton iteration method is adopted to improve convergence speed. Moreover, in order to address the problem that FastICA is very sensitive to initialization, a steepest descent method can be applied. The experimental analysis of the compound fault of rolling bearing verifies that the presented method is effective to separate multiple fault source signals and the improved FastICA algorithm can increase convergence rate and overcome the problem of sensitivity to initialization.

Robust ICA based mode demultiplexing

In this paper, we present a mode demultiplexing algorithm based on robust independent component analysis (ICA) in electrical domain. Multi-mode fiber (MMF) based spatial division multiplexing system is modeled as a behavioral model. Different modes can carry the data to increase the system capacity with the proposed algorithm. In the system, two orthogonal frequency division multiplexing (OFDM) signals are modulated on two modes by two vertical cavity surface emitting lasers (VCSEL) and then coupled into the MMF with an optical coupler. Two channel signals are separated by the proposed robust ICA mode demultiplexing algorithm after photodetectors at the receiver side. The OFDM signals can be demodulated successfully after the separation. Compared with a single mode transmission in MMF, the transmission performance of two mode channels is not degraded significantly with the proposed mode demultiplexing algorithm. Our results show that the proposed algorithm can carry out the modes demultiplexing successfully and the cross-talk between modes can be reduced or be negligible after processing. It also demonstrates a new application of ICA algorithm in the electrical time-domain mode demultiplexing.

Fast Independent Component Analysis Algorithm-Based Diagnosis of L5 Nerve Root Compression and Changes of Brain Functional Areas Using 3D Functional Magnetic Resonance Imaging.

In this paper, the application of 3-dimensional (3D) functional magnetic resonance imaging (FMRI) in the diagnosis of the 5th lumbar (L5) nerve root compression and brain functional areas in patients with lumbar disc herniation (LDH) was analyzed. The traditional fast independent component analysis (Fast ICA) algorithm was optimized based on the modified whitening matrix to establish a new type of Modified-Fast ICA (M-Fast ICA) algorithm that was compared with the introduced traditional Fast ICA and ICA. M-Fast ICA was applied to the 3D FMRI diffusion tensor imaging (DTI) evaluation of 65 patients with L5 nerve root pain due to LDH (group A) and 50 healthy volunteers (group B). The values of fractional anisotropy (FA) and apparent diffusion coefficient (ADC) in the lumbar nerve roots (L3, L4, L5, and the 1st sacral vertebra (S1)) were recorded among subjects from the two groups. Besides, the score of edema degree in the lumbar nerve roots (L5 and S1) and activity of brain functional areas were also recorded among all subjects of the two groups. The results showed that the mean square error of M-Fast ICA was smaller than that of traditional Fast ICA and ICA, while its signal-to-noise ratio (SNR) was greater than that of Fast ICA and ICA (P < 0.05). The FA of L5 and S1 nerve roots in patients of group A was sharply lower than the values of group B, while the ADC of patients in group A was greater than that of the control group (P < 0.05). Besides, the score of edema in L5 and S1 nerve roots of patients in group A increased in contrast to group B (P < 0.05). The brain areas were activated after surgery including bilateral temporal lobe, left thalamus, splenium of corpus callosum, and right internal capsule. In conclusion, the 3D image denoising performance of M-Fast ICA optimized and constructed in this study was superior to that of the traditional Fast ICA and ICA. The FA of patients with L5 nerve root pain due to LDH decreased steeply, while the ADC increased dramatically. L5 nerve root pain caused by LDH resulted in changes in brain functional areas of the patients to inhibit the resting state default network activity, and the corresponding brain functional areas could be activated through treatment.

NIRS-ICA: A MATLAB Toolbox for Independent Component Analysis Applied in fNIRS Studies.

Independent component analysis (ICA) is a multivariate approach that has been widely used in analyzing brain imaging data. In the field of functional near-infrared spectroscopy (fNIRS), its promising effectiveness has been shown in both removing noise and extracting neuronal activity-related sources. The application of ICA remains challenging due to its complexity in usage, and an easy-to-use toolbox dedicated to ICA processing is still lacking in the fNIRS community. In this study, we propose NIRS-ICA, an open-source MATLAB toolbox to ease the difficulty of ICA application for fNIRS studies. NIRS-ICA incorporates commonly used ICA algorithms for source separation, user-friendly GUI, and quantitative evaluation metrics assisting source selection, which facilitate both removing noise and extracting neuronal activity-related sources. The options used in the processing can also be reported easily, which promotes using ICA in a more reproducible way. The proposed toolbox is validated and demonstrated based on both simulative and real fNIRS datasets. We expect the release of the toolbox will extent the application for ICA in the fNIRS community.

Improvement and Assessment of a Blind Image Deblurring Algorithm Based on Independent Component Analysis

The aim of the present paper is to improve an existing blind image deblurring algorithm, based on an independent component learning paradigm, by manifold calculus. The original technique is based on an independent component analysis algorithm applied to a set of pseudo-images obtained by Gabor-filtering a blurred image and is based on an adapt-and-project paradigm. A comparison between the original technique and the improved method shows that independent component learning on the unit hypersphere by a Riemannian-gradient algorithm outperforms the adapt-and-project strategy. A comprehensive set of numerical tests evidenced the strengths and weaknesses of the discussed deblurring technique.