Explore the vast range of titles available.

Background Conventional methods of motor imagery brain computer interfaces (MI-BCIs) suffer from the limited number of samples and simplified features, so as to produce poor performances with spatial-frequency features and shallow classifiers. Methods Alternatively, this paper applies a deep recurrent neural network (RNN) with a sliding window cropping strategy (SWCS) to signal classification of MI-BCIs. The spatial-frequency features are first extracted by the filter bank common spatial pattern (FB-CSP) algorithm, and such features are cropped by the SWCS into time slices. By extracting spatial-frequency-sequential relationships, the cropped time slices are then fed into RNN for classification. In order to overcome the memory distractions, the commonly used gated recurrent unit (GRU) and long-short term memory (LSTM) unit are applied to the RNN architecture, and experimental results are used to determine which unit is more suitable for processing EEG signals. Results Experimental results on common BCI benchmark datasets show that the spatial-frequency-sequential relationships outperform all other competing spatial-frequency methods. In particular, the proposed GRU-RNN architecture achieves the lowest misclassification rates on all BCI benchmark datasets. Conclusion By introducing spatial-frequency-sequential relationships with cropping time slice samples, the proposed method gives a novel way to construct and model high accuracy and robustness MI-BCIs based on limited trials of EEG signals.

Background Epilepsy is a neurological disorder that is usually detected by electroencephalogram (EEG) signals. Since manual examination of epilepsy seizures is a laborious and time-consuming process, lots of automatic epilepsy detection algorithms have been proposed. However, most of the available classification algorithms for epilepsy EEG signals adopted a single feature extraction, in turn to result in low classification accuracy. Although a small account of studies have carried out feature fusion, the computational efficiency is reduced due to too many features, because there are also some poor features that interfere with the classification results. Methods In order to solve the above problems, an automatic recognition method of epilepsy EEG signals based on feature fusion and selection is proposed in this paper. Firstly, the Approximate Entropy (ApEn), Fuzzy Entropy (FuzzyEn), Sample Entropy (SampEn), and Standard Deviation (STD) mixed features of the subband obtained by the Discrete Wavelet Transform (DWT) decomposition of EEG signals are extracted. Secondly, the random forest algorithm is used for feature selection. Finally, the Convolutional Neural Network (CNN) is used to classify epilepsy EEG signals. Results The empirical evaluation of the presented algorithm is performed on the benchmark Bonn EEG datasets and New Delhi datasets. In the interictal and ictal classification tasks of Bonn datasets, the proposed model achieves an accuracy of 99.9%, a sensitivity of 100%, a precision of 99.81%, and a specificity of 99.8%. For the interictal-ictal case of New Delhi datasets, the proposed model achieves a classification accuracy of 100%, a sensitivity of 100%, a specificity of 100%, and a precision of 100%. Conclusion The proposed model can effectively realize the high-precision automatic detection and classification of epilepsy EEG signals. This model can provide high-precision automatic detection capability for clinical epilepsy EEG detection. We hope to provide positive implications for the prediction of seizure EEG.

Electroencephalogram (EEG) have been widely used for the diagnosis of neurological diseases like epilepsy and sleep disorders. Support vector machines (SVMs) are widely used classifiers for the classification of EEG signals due to their better generalization performance. However, SVM suffers due to high computational complexity. To reduce the computations, twin support vector machines (TWSVM) solved smaller size quadratic optimization problems. To enhance the performance of the SVM and TWSVM models, prior information known as universum data has been incorporated in the universum SVM (USVM) and universum twin (UTSVM) models. Both SVM and UTSVM employ hinge loss which results in sensitivity to noise and instability. To overcome these issues and incorporate the prior information of the EEG signals, we propose a novel universum twin support vector machine with pinball loss function (Pin-UTSVM) for the classification of EEG signals. The proposed Pin-UTSVM is more stable for resampling and is noise insensitive. Furthermore, the computational complexity of proposed Pin-UTSVM model is similar to the standard UTSVM model. In the proposed approach, we used the interictal EEG signal as the universum data. Numerical experiments at varying level of noise show that the proposed Pin-UTSVM is more robust to noise compared to standard models. To show the efficiency of the proposed Pin-UTSVM model, we used multiple feature extraction techniques for the classification of the EEG signal. Experimental results reveal that the proposed Pin-UTSVM model is performing better compared to the existing models. Moreover, statistical tests show that the proposed Pin-UTSVM model is significantly better in comparison with the existing baseline models.

This paper refers to the method of using the deep neural long-short-term memory (LSTM) network for the problem of electrocardiogram (ECG) signal classification. ECG signals contain a lot of subtle information analyzed by doctors to determine the type of heart dysfunction. Due to the large number of signal features that are difficult to identify, raw ECG data is usually not suitable for use in machine learning. The article presents how to transform individual ECG time series into spectral images for which two characteristics are determined, which are instantaneous frequency and spectral entropy. Feature extraction consists of converting the ECG signal into a series of spectral images using short-term Fourier transformation. Then the images were converted using Fourier transform again to two signals, which includes instantaneous frequency and spectral entropy. The data set transformed in this way was used to train the LSTM network. During the experiments, the LSTM networks were trained for both raw and spectrally transformed data. Then, the LSTM networks trained in this way were compared with each other. The obtained results prove that the transformation of input signals into images can be an effective method of improving the quality of classifiers based on deep learning.

Neuropsychiatric disorders are one of the leading causes of disability. Mental health problems can occur due to various biological and environmental factors. The absence of definitive confirmatory diagnostic tests for psychiatric disorders complicates the diagnosis. It's critical to distinguish between bipolar disorder, depression, and schizophrenia since their symptoms and treatments differ. Because of brain–heart autonomic connections, electrocardiography (ECG) signals can be changed in behavioral disorders. In this research, we have automatically classified bipolar, depression, and schizophrenia from ECG signals. In this work, a new hand-crafted feature engineering model has been proposed to detect psychiatric disorders automatically. The main objective of this model is to accurately detect psychiatric disorders using ECG beats with linear time complexity. Therefore, we collected a new ECG signal dataset containing 3,570 ECG beats with four categories. The used categories are bipolar, depression, schizophrenia, and control. Furthermore, a new ternary pattern-based signal classification model has been proposed to classify these four categories. Our proposal contains four essential phases, and these phases are (i) multileveled feature extraction using multilevel discrete wavelet transform and ternary pattern, (ii) the best features selection applying iterative Chi2 selector, (iii) classification with artificial neural network (ANN) to calculate lead wise results and (iv) calculation the voted/general classification accuracy using iterative majority voting (IMV) algorithm. tenfold cross-validation is one of the most used validation techniques in the literature, and this validation model gives robust classification results. Using ANN with tenfold cross-validation, lead-by-lead and voted results have been calculated. The lead-by-lead accuracy range of the proposed model using the ANN classifier is from 73.67 to 89.19%. By deploying the IMV method, the general classification performance of our ternary pattern-based ECG classification model is increased from 89.19 to 96.25%. The findings and the calculated classification accuracies (single lead and voted) clearly demonstrated the success of the proposed ternary pattern-based advanced signal processing model. By using this model, a new wearable device can be proposed.

This paper presents a fast and compressible multiple signal classification algorithm (C-MUSIC) based on the manifold separation technique (MST) for a uniform circular array (UCA), named as MC-MUSIC. Making use of the MST, the steering vector (SV) of the UCA can be converted into a new SV, which has a Vandermonde structure similar to a uniform linear array. Then a noise-like subspace cluster (NLSC) can be generated by applying a new linear relation between the true direction-of-arrival (DOA) and virtual DOA to the new given SV. Finally, the angular field-of-view can be compressed to a small sector by the intersection of the NLSC so that the calculation times of the spatial spectrum can be reduced effectively. Simulations demonstrate the presented method can maintain the high precision of the DOA estimation with low computational complexity.

Fatigue detection is a critical application area for machine learning, and variable input data have been utilized to detect fatigue. One of the most commonly used inputs for fatigue detection is electroencephalography (EEG) signals. The main objective of this study is to accurately detect fatigue using a hand-crafted framework. To achieve this, a new signal classification framework has been proposed, and its performance has been tested on an EEG fatigue detection dataset. Wavelet packet decomposition with 16 mother wavelet functions has been utilized to extract features from the frequency domain and create a multilevel feature extraction method to calculate frequency subbands. To generate classification results, two validation techniques, tenfold cross-validation and leave-one-subject-out (LOSO) validation, have been applied to attain robust classification results. The proposed framework achieved high classification performance with 99.90% and 82.08% classification accuracies using tenfold CV and LOSO CV, respectively. Furthermore, the classification performance of each used method in our framework has been analyzed to understand the driving fatigue classification effect of the machine learning functions used. The proposed framework attained superior classification results, demonstrating its efficacy in accurately detecting fatigue.

This study proposes a complete data-processing framework for Geiger-mode avalanche photodiode (GM-APD) light detection and ranging (LiDAR) echo signals. It investigates the feasibility of classifying target and background noise using machine learning. Four feature processing schemes were first compared, among which the PNT strategy (Principal Component Analysis without tail features) was identified as the most effective and adopted for subsequent analysis. Based on this framework, nine models derived from six baseline algorithms—Decision Trees (DTs), Support Vector Machines (SVMs), Backpropagation Neural Networks (NN-BPs), Linear Discriminant Analysis (LDA), Logistic Regression (LR), and k-Nearest Neighbors (KNN)—were systematically assessed under Monte Carlo simulations with varying echo signal-to-noise ratio (ESNR) and statistical frame number (SFN) conditions. Model performance was evaluated using eight metrics: accuracy, precision, recall, FPR, FNR, F1-score, Kappa coefficient, and relative change percentage (RCP). Monte Carlo simulations were employed to generate datasets, and Principal Component Analysis (PCA) was applied for feature extraction in the machine learning training process. The results show that LDA achieves the shortest training time (0.38 s at SFN = 20,000), DT maintains stable accuracy (0.7171–0.8247) across different SFNs, and NN-BP models perform optimally under low-SNR conditions. Specifically, NN-BP-3 achieves the highest test accuracy of 0.9213 at SFN = 20,000, while NN-BP-2 records the highest training accuracy of 0.9137. Regarding stability, NN-BP-3 exhibits the smallest RCP value (0.0111), whereas SVM-3 yields the largest (0.1937) at the same frame count. In conclusion, NN-BP-based models demonstrate clear advantages in classifying sky-background noise. Building on this, we design a ResNet based on NN-BP, which achieves further accuracy gains over the best baseline at 400, 2000, and 20,000 frames—12.5% (400), 9.16% (2000), and 2.79% (20,000)—clearly demonstrating the advantage of NN-BP for GM-APD LiDAR signal classification. This research thus establishes a novel framework for GM-APD LiDAR signal classification, provides the first systematic comparison of multiple machine learning models, and highlights the trade-off between accuracy and computational efficiency. The findings confirm the feasibility of applying machine learning to GM-APD data and offer practical guidance for balancing detection performance with real-time requirements in field applications.

A radial basis probabilistic process neuron (RBPPN) and radial basis probabilistic process neural network (RBPPNN) model are proposed to fuse a priori knowledge for application to time-varying signal pattern classification. RBPPN inputs were multi-channel time-varying signals and a generalized inner product was used to perform spatio-temporal aggregation of input signals in the kernel. Typical signal samples from various pattern subsets in the sample set were used as kernel center functions, which use morphological distribution characteristics and combination relationships to implicitly express prior knowledge for the signal category. The exponential probability function was used as the activation function to achieve kernel transformation and RBPPN probability output. The RBPPNN is composed of process signal input layers, an RBPPN hidden layer, a pattern layer, and a Softmax classifier developed through stacking. Generalized inner product operations were used to conduct probability similarity measurements of distribution characteristics between process signals. The pattern layer selectivity summed inputs from the RBPPN hidden layer to the pattern layer according to the category of the kernel center function. Its outputs were then used as inputs in the Softmax classifier. The proposed RBPPNN information processing mechanism was extended to the time domain, and through learning time-varying signal training samples, achieved extraction, expression, and information association of time-varying signal characteristics, as well as direct classification. It can improve the deficiencies of existing neural networks, such as a complete large-scale training dataset is needed, and the information processing flow is complex. In this paper, the properties of the RBPPNN are analyzed and a specific learning algorithm is presented which synthesizes dynamic time warping, dynamic C-means clustering, and the mean square error algorithm. A series of 12-lead electrocardiogram (ECG) signals were used for classification testing of heart disease diagnosis results. The ECG classification accuracy across ten disease types was 75.52% and sinus arrhythmia was identified with an accuracy of 86.75%, verifying the effectiveness of the model and algorithm.

Sparse frequency agile orthogonal frequency division multiplexing (SFA-OFDM) signal brings excellent performance to electronic counter-countermeasures (ECCM) and reduces the complexity of the radar system. However, frequency agility makes coherent processing a much more challenging task for the radar, which leads to the discontinuity of the echo phase in a coherent processing interval (CPI), so the fast Fourier transform (FFT)-based method is no longer a valid way to complete the coherent integration. To overcome this problem, we proposed a novel scheme to estimate both super-resolution range and velocity. The subcarriers of each pulse are firstly synthesized in time domain. Then, the range and velocity estimations for the SFA-OFDM radar are regarded as the parameter estimations of a linear array. Finally, both the super-resolution range and velocity are obtained by exploiting the multiple signal classification (MUSIC) algorithm. Simulation results are provided to demonstrate the effectiveness of the proposed method.