EOG artifact removal using a wavelet neural network

As is well known, it is the application of runoff flood forecasting that is extraordinary significant for us. A detailed detection of the flood forecasting process has been carried out using powerful artificial neural network in this paper. Learning algorithm of wavelet neural network was produced by extruding it in BP idea.The determination of network hidden layer nodes utilizes the method of tring fault. Activation function belongs to morlet wavelet fund ion,and the module of net structure belongs to 371. It is shown that the reliable prediction accury could be provided by using this model for predicting and analysing for the flood data of solar Da in 1996.

ABSTRACTIn order to solve the disadvantages of the traditional wavelet neural network (WNN) algorithm applied in transformer fault diagnosis, such as uneven sample distribution of training samples and high diagnostic error rate and long training time, an improved fault diagnosis method is proposed based on fuzzy clustering and the flower pollination algorithm. Firstly, fuzzy clustering is applied to deal with transformer fault sample data so as to remove the bad data; secondly, the flower pollination algorithm is applied to obtain the optimal parameters of the WNN. The example analysis results show that WNN based on the flower pollination algorithm (FPA-WNN) has better convergence, lower diagnosis error rate and shorter training time compared with WNN based on the particle swarm algorithm (PWA-WNN) and it is more suitable for transformer fault diagnosis.

In this paper we present algorithms which are adaptive and based on neural networks and wavelet series to build wavenets function approximators. Results are shown in numerical simulation of two wavenets approximators architectures: the first is based on a wavenet for approach the signals under study where the parameters of the neural network are adjusted online, the other uses a scheme approximators with an IIR filter in the output of wavenet, which helps to reduce convergence time to a minimum time desired.

Validating and improving the correction of ocular artifacts in electro-encephalography

With the rapid development of the global manufacturing industry, an efficient and accurate quality control system has become key to enhancing competitiveness. Ultrasonic Nondestructive Testing (NDT), as an efficient means of quality inspection, plays a crucial role in improving manufacturing quality through the precision of its data analysis. This study aims to explore the application of ultrasonic NDT data in manufacturing quality control by integrating machine learning technologies, with a specific focus on the Wavelet Neural Network optimized by Genetic Algorithms (GA-WNN). This study achieved significant prediction and evaluation results by applying a GA-WNN to quality control in manufacturing. Compared to traditional Wavelet Neural Network (WNN) models, the GA-WNN more effectively identifies and predicts potential quality issues, especially in noisy data and complex production environments, demonstrating higher accuracy and stability. When predicting possible defect types in the manufacturing process, the GA-WNN showed a notable improvement in accuracy over other models. Additionally, in quality stability evaluation, GA-WNN was able to capture production fluctuations more accurately, providing more valuable results for decision-making. The methodologies and discoveries of this study offer new perspectives and tools for quality control in manufacturing and the analysis of ultrasonic NDT data, presenting broad application prospects.

This paper proposed the discrete transform and neural network algorithms to obtain the monthly peak load demand in mid term load forecasting. The mother wavelet daubechies2 (db2) is employed to decomposed, high pass filter and low pass filter signals from the original signal before using feed forward back propagation neural network to determine the forecasting results. The historical data records in 1997-2007 of Electricity Generating Authority of Thailand (EGAT) is used as reference. In this study, historical information of peak load demand(MW), mean temperature(Tmean), consumer price index (CPI), and industrial index (economic:IDI) are used as feature inputs of the network. The experimental results show that the Mean Absolute Percentage Error (MAPE) is approximately 4.32%. This forecasting results can be used for fuel planning and unit commitment of the power system in the future.

Objective: The activity of the brain can be recorded by means of an electroencephalogram (EEG). An EEG is a multichannel signal related to brain activity. However, EEG presents a wide variety of undesired artefacts. Removal of these artefacts is often done using blind source separation methods (BSS) and mainly those based on Independent Component Analysis (ICA). ICA-based methods are well-accepted in the literature for filtering artefacts and have proved to be satisfactory in most scenarios of interest. Our goal is to develop a generic and unsupervised ICA-based algorithm for EEG artefacts removal. Approach: The proposed algorithm makes use of a new unsupervised artefact detection, ICA and a statistical criterion to automatically select the artefact related independent components (ICs) requiring no human intervention. The algorithm is evaluated using both simulated and real EEG data with artefacts (SEEG and AEEG). A comparison between the proposed unsupervised selection of ICs related to the artefact and other supervised selection is also presented. Main results: A new unsupervised ICA-based algorithm to filter artefacts, where ICs related to each artefact are automatically selected. It can be used in online applications, it preserves most of the original information among the artefacts and removes different types of artefacts. Significance: ICA-based methods for filtering artefacts prevail in the literature. The work in this article is important insofar as it addresses the problem of automatic selection of ICs in ICA-based methods. The selection is unsupervised, avoiding the manual ICs selection or a learning process involved in other methods. Our method is a generic algorithm that allows removing EEG artefacts of various types and, unlike some ICA-based algorithms, it retains most of the original information among the artefacts. Within the algorithm, the artefact detection method implemented does not require human intervention either.

The Sixth International Symposium on Neural Networks (ISNN 2009)

On the basis of wavelet theory, a novel Adaptive wavelet thresholding method (AWT) is proposed for the ECG signal enhancement. The best base wavelet for ECG signal filtering can be automatically obtained through the cross correlation coefficient and the energy to entropy ratio. The variable universal threshold (VarUniversal) is applied to different decomposition level so as to suppress diverse noise. To achieve a smooth cut-off transition, an identical correlation shrinkage function (IcoShrinkage) is also adopted in the AWT according to its correlation coefficients with the hard thresholding and the soft thresholding. The performance of AWT is compared with four threshold approaches and six shrinkage functions, respectively, on the basis of 150 practical ECG signals of 30 subjects. The filtering results reveal that the AWT can adaptively choose an optimal base wavelet for a specific ECG signal. With the VarUniversal threshold and IcoShrinkage, the AWT obtains the better filtering results than the other compared methods.

Electrocardiographic (ECG) signals are affected by several kinds of artifacts that may hide vital signs of interest. In this study we apply independent component analysis (ICA) to isolate motion artifacts. Standard or instantaneous ICA, which is currently the most addressed ICA model within the context of artifact removal, is compared to two other ICA techniques. The first technique is a frequency domain approach to convolutive mixture separation. The second is based on temporally constrained ICA, which enables the estimation of only one component close to a particular reference signal. Performance indexes evaluate ECG complex enhancement and relevant heart rate errors. Our results show that both convolutive and constrained ICA implementations perform better than standard ICA, thus opening up a new field of application for these two methods. Moreover, statistical analysis reveals that constrained ICA and convolutive ICA do not significantly differ concerning heart rate estimation, even though the latter overcomes the former in ECG morphology recovery.

This paper addresses the overlearning problem in the independent component analysis (ICA) used for the removal of muscular artifacts from electroencephalographic (EEG) records. We note that for short EEG records with high number of channels the ICA fails to separate artifact-free EEG and muscular artifacts, which has been previously attributed to the phenomenon called overlearning. We address this problem by projecting an EEG record into several subspaces with a lower dimension, and perform the ICA on each subspace separately. Due to a reduced dimension of the subspaces, the overlearning is suppressed, and muscular artifacts are better separated. Once the muscular artifacts are removed, the signals in the individual subspaces are combined to provide an artifact free EEG record. We show that for short signals and high number of EEG channels our approach outperforms the currently available ICA based algorithms for muscular artifact removal. The proposed technique can efficiently suppress ICA overlearning for short signal segments of high density EEG signals.

In this paper, a set of multi-channel sEMG was given to examine the coordination between quadriceps and hamstring. To extract valuable muscle information and also describe muscles’ activities from noisy multi-channel sEMG, advanced signal processing techniques are required to cancel additive noise, remove artifacts and reduce crosstalkes among electrodes. In the last two decades, Independent Component Analysis (ICA) techniques have been successfully used to remove crosstalkes among electrodes. Besides that, ICA are multi-channel adaptive signal processing techniques and do not require any prior information of the signal. In our study, ICA is used to eliminate the crosstalks among the electrodes. To improve the performance of our system, spectral descriptors were also used to describe muscle activities. Many experiments and various simulations have been conducted. Our experimental results showed that ICA can remove the cross-talks and further improve the signal quality while three different spectral descriptors which are energy, MPF and VMF can work together to describe a muscle activity effectively.KeywordsMulti-channel sEMGICASpectral Curve DescriptorQuadricepsHamstring

Automated artifact removal as preprocessing refines neonatal seizure detection

Quantitative ultrasound (QUS) imaging complements the standard B-mode images with a quantitative represen-tation of the target's acoustic properties. Attenuation coefficient is an important parameter characterizing these properties, with applications in medical diagnosis and tissue characterization. Traditional QUS methods use analytical models to estimate this coefficient from the acquired signal. Propagation effects, such as diffraction, which are difficult to model analytically are usually ignored, affecting their estimation accuracy. To tackle this issue, reference phantom measurements are commonly used. These are, however, time-consuming and may not always be feasible, limiting the existing approaches' practical applicability. To overcome these challenges, we leverage recent advances in the deep learning field and propose a neural network approach which takes the magnitude spectra of the backscattered ultrasound signal at different axial depths as the input and provides the target's attenuation coefficient as the output. For the presented proof-of-concept study, the network was trained on a simulated dataset, and learnt a proper model from the training data, thereby avoiding the need for an analytical model. The trained network was tested on both simulated and tissue-mimicking phantom datasets, demonstrating the capability of neural networks to provide accurate attenuation estimates from diffraction affected recordings without a reference phantom measurement.

Nonlinear dynamic systems are characterized with uncertainties in terms of structure and parameters. These uncertainties cannot be described by deterministic models. The modelling and identification of nonlinear dynamic systems through the measured experimental data is a problem in engineering and technical processes. Therefore, field of system identification have become an important area of research. Fuzzy technology is an effective tool for dealing with complex nonlinear processes that are characterized with uncertain factors. In this paper, a novel approach based on Local Linear method learning in dynamical filter weights neurons for the identification of non-linear dynamic systems is presented. The fuzzy wavelet neural network combines wavelet theory with fuzzy logic and neural networks. Learning fuzzy rules and parameter update in fuzzy wavelet neural network is based on gradient decent method. The proposed approach is said to be Fuzzy Local Linear Wavelet Neural Network based model. It has been explained through examples. The structure is tested for the identification with both wavelet neural network and Fuzzy Local Linear Wavelet Neural Network that shows the comparative performance.KeywordsSystem IdentificationNon-linear SystemWavelet Neural NetworkFuzzy Wavelet Neural NetworkLocal Linear Wavelet Neural NetworkFuzzy Local Linear Wavelet Neural Network