Search

Transfer learning (TL) has been applied in seizure detection to deal with differences between different subjects or tasks. In this paper, we consider cross-subject seizure detection that does not rely on patient history records, that is, acquiring knowledge from other subjects through TL to improve seizure detection performance. We propose a novel domain adaptation method, named the Cluster Embedding Joint-Probability-Discrepancy Transfer (CEJT), for data distribution structure learning. Specifically, 1) The joint probability distribution discrepancy is minimized to reduce the distribution shift in the source and target domains, and strengthen the discriminative knowledge of classes. 2) A clustering is performed on the target domain, and the class centroids of sources is used as the clustering prototype of the target domain to enhance data structure. It is worth noting that the manifold regularization is used to improve the quality of clustering prototypes. In addition, a correlation-alignment-based source selection metric (SSC) is designed for most favorable subject selection, reducing the computational cost as well as avoiding some negative transfer. Experiments on 15 patients with focal epilepsy from the Children’s Hospital, Zhejiang University School of Medicine (CHZU) database shown that CEJT outperforms several state-of-the-art approaches, and can promote the application of seizure detection. [ABSTRACT FROM AUTHOR]

Benign epilepsy with spinous waves in the central temporal region (BECT) is the most common epilepsy syndromes in children. Spike discharges in the Rolandic area are important biomarkers for diagnosis evaluation. Conventional single-channel electroencephalogram (EEG) based spike detection methods are generally susceptible to artifact interference. To address this issue, a novel spike detection method based on multichannel EEG weighted fusion strategy is developed in this brief. The proposed algorithm mainly includes multichannel spike candidate sample screening, data weighted fusion, time-series feature extraction and long-short-time memory neural networks (LSTM) detection. Studies on 15 BECT children show that the proposed algorithm can obtain an average of 95.74% F1 scores, 93.94% sensitivity, 97.73% precision for all subjects. [ABSTRACT FROM AUTHOR]

WEST syndrome is an unknown etiology infant epilepsy, which is characterized by the flexion spastic seizure, intellectual motion development lag, electrode abnormalities, arrhythmia. In this brief, we present a novel electrocardiogram (ECG) based WEST syndrome epilepsy seizure detection method. Based on deterministic learning (DT) theory, the dynamic model of ECG is firstly constructed. The cardiodynamicsgrams (CDGs) of ECGs in seizure and interictal periods are then derived. Nonlinear features on CDGs are extracted for WEST syndrome characterization. For performance evaluation, experiments on ECGs of 12 WEST syndrome patients from the Children’s Hospital of Zhejiang University School of Medicine (CHZU) is carried out. The proposed method can obtain an average of $94.49{\%}$ F1-score, $93.76{\%}$ precision and $95.58{\%}$ accuracy, that outperforms the heart rate variability (HRV) based methods. [ABSTRACT FROM AUTHOR]

In this paper, a full-reference video quality assessment (VQA) model is designed for the perceptual quality assessment of the screen content videos (SCVs), called the hybrid spatiotemporal feature-based model (HSFM). The SCVs are of hybrid structure including screen and natural scenes, which are perceived by the human visual system (HVS) with different visual effects. With this consideration, the three dimensional Laplacian of Gaussian (3D-LOG) filter and three dimensional Natural Scene Statistics (3D-NSS) are exploited to extract the screen and natural spatiotemporal features, based on the reference and distorted SCV sequences separately. The similarities of these extracted features are then computed independently, followed by generating the distorted screen and natural quality scores for screen and natural scenes. After that, an adaptive screen and natural quality fusion scheme through the local video activity is developed to combine them for arriving at the final VQA score of the distorted SCV under evaluation. The experimental results on the Screen Content Video Database (SCVD) and Compressed Screen Content Video Quality (CSCVQ) databases have shown that the proposed HSFM is more in line with the perceptual quality assessment of the SCVs perceived by the HVS, compared with a variety of classic and latest IQA/VQA models. [ABSTRACT FROM AUTHOR]

Objectives: Eye blink artifact detection in scalp electroencephalogram (EEG) of epilepsy patients is challenging due to its similar waveforms to epileptiform discharges. Developing an accurate detection method is urgent and critical. Methods: In this paper, we proposed a novel multi-dimensional feature optimization based eye blink artifact detection algorithm for EEGs containing rich epileptiform discharges. An unsupervised clustering algorithm based on smoothed nonlinear energy operator (SNEO) and variational mode extraction (VME) is proposed to detect epileptiform discharges in the frontal leads. Then, multi-dimensional time/frequency EEG features extracted from forehead electrodes (FP1 and FP2 channels) combining with the improved VME (IVME) threshold are derived for EEG representation. A variance filtering method is further applied for discriminative feature selection and a machine learning model is finally learned to perform detection. Results: Experiments on EEGs of 16 subjects from the Children’s Hospital of Zhejiang University School of Medicine (CHZU) show that our method achieves the highest average sensitivity, specificity and accuracy of 95.04, 89.52, and 93.01, respectively. That outperforms 5 recent and state-of-the-art (SOTA) eye blink detection algorithms. Significance: The proposed method is robust in eye blink artifact detection for EEGs containing high-frequency epileptiform discharges. It is also effective in dealing with individual differences in EEGs, which is usually ignored in conventional methods. [ABSTRACT FROM AUTHOR]

Spike detection is of great significance in the detection of epileptic seizures. Many spike detection algorithms have been proposed, but available algorithms often miss spikes during high firing rate epochs or with artifacts. Also, there exists methodological limitations when there are great variations in spike morphology between different patients or the same patient in different EEG epoch. Thus a data-driven spike detection method using multi-template matching, feature extraction and threshold method is proposed in this work to improve the detection performance under different circumstances. First, universal template matching and feature extraction are used to get putative spikes. Then clustering algorithm and adaptive template matching algorithm are applied to maximize the number of detected spikes for these specific candidate single units obtained from the clustering method. In addition, a data-driven approach based on receiver operating characteristic (ROC) is used for low false identification rate detection. The performance is evaluated on real EEG samples and the evaluating results shows that this algorithm has achieved 97.12% average sensitivity and 0.55 average false negative rate per minute. The least sensitivity is 94.32% and the most false negative rate per minute is 1.23. [ABSTRACT FROM AUTHOR]

The benign epilepsy with spinous waves in the central temporal region (BECT) is the one of the most common epileptic syndromes in children, that seriously threaten the nervous system development of children. The most obvious feature of BECT is the existence of a large number of electroencephalogram (EEG) spikes in the Rolandic area during the interictal period, that is an important basis to assist neurologists in BECT diagnosis. With this regard, the paper proposes a novel BECT spike detection algorithm based on time domain EEG sequence features and the long short-term memory (LSTM) neural network. Three time domain sequence features, that can obviously characterize the spikes of BECT, are extracted for EEG representation. The synthetic minority oversampling technique (SMOTE) is applied to address the spike imbalance issue in EEGs, and the bi-directional LSTM (BiLSTM) is trained for spike detection. The algorithm is evaluated using the EEG data of 15 BECT patients recorded from the Children’s Hospital, Zhejiang University School of Medicine (CHZU). The experiment shows that the proposed algorithm can obtained an average of 88.54% F1 score, 92.04% sensitivity, and 85.75% precision, that generally outperforms several state-of-the-art spike detection methods. [ABSTRACT FROM AUTHOR]

Accurate eye blink artifact detection is essential for electroencephalogram (EEG) analysis and auxiliary analysis of nervous system diseases, especially in the presence of the frontal epileptiform discharges. In this paper, we develop a novel eye blink artifact detection algorithm based on optimally selected multi-dimensional EEG features. Specific efforts have been paid to filtering the frontal epileptiform discharges, where an unsupervised learning exploiting the EEG signal physiological characteristics and smooth nonlinear energy operator (SNEO) based on the K-means clustering has been firstly proposed. Multiple statistical EEG features derived from the frontal electrodes and other electrodes are then extracted to characterize eye blink artifacts. Discriminative feature selection scheme based on the variance filtering and Relief algorithms has been respectively studied, and the average correlation coefficient (ACC) is applied for feature optimization evaluation. The eye blink artifact detection is finally achieved based on the support vector machine (SVM) trained on the optimized EEG features. The effectiveness of the proposed algorithm is demonstrated by experiments carried out on the EEG database of 11 subjects recorded from the Children’s Hospital, Zhejiang University School of Medicine (CHZU). Comparisons to several state-of-the-art (SOTA) eye blink artifact detection methods are also presented. [ABSTRACT FROM AUTHOR]

Seizure onset detection and epileptic preictal prediction based on electroencephalogram (EEG) signals have been a challenge problem in the research community. In this brief, a novel epileptic states classification algorithm based on the multichannel EEGs representation using multiple hand-crafted features, the feature fusion and transfer learning (TL) with multiple pre-trained deep neural networks (DNNs), the discriminative feature extraction and epileptic state classification with a hierarchical neural network (HNN), is developed. The mean amplitude spectrum (MAS), mean power spectral density (MPSD) and wavelet packet features (WPFs) are firstly derived and fused into an image feature for multichannel EEGs representation. Then, 5 classical pre-trained DNNs are directly adopted as feature extractors on the fused image feature. A 7-layer fully-connected (FC) HNN is finally constructed for discriminative feature learning and epileptic state classification. The effectiveness is demonstrated through experiments on the CHB-MIT and the iNeuro epilepsy EEG databases. [ABSTRACT FROM AUTHOR]

Hierarchical neural network based one-class anomaly detection algorithms generally rely on stacked autoencoders (AEs) for feature learning. But existing AEs are not specifically designed to exploit the discriminative characteristics of target data in one-class classification (OCC), and thus may lead to poor generalization performance. In this brief, a novel randomized AE that imposes the constraint of the within-class scatter information is developed in feature learning. The correntropy criterion is applied to replace the mean square error criterion (MSE) to enhance the algorithm performance in outlier and noise rejection. The algorithm is further extended to kernel learning to improve its generalization capability. Experiments on benchmark datasets are carried out to show the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Benign childhood epilepsy with centro-temporal spikes (BECT) is one of the most common epilepsy syndromes in childhood which is typically characterized by localized discharges in the central and temporal regions. Traditionally, the recognition of spikes requires visual assessment of long-term EEG recordings which is time consuming and subjective because it depends on the knowledge and experience of the doctor. Therefore, a novel multi-step spike detection algorithm based on average reference (AV) channel and bipolar (BP) channel BECT EEG is proposed, including candidate spike detection algorithm, false positive spike (FPS) elimination, spike feature extraction and random forest (RF) classification. The proposed method is evaluated using 7 routine EEG recordings. This brief shows that the sensitivity (Sen), specificity (Spe), selectivity (Sel) and accuracy (AC) obtained by the proposed method are 97.4%, 96.5%, 96.6% and 96.9%, respectively. Experimental results show that the proposed method is capable of detecting BECT spikes efficiently. [ABSTRACT FROM AUTHOR]

Radar Emitter Identification (REI) has been broadly used in military and civil fields. In this paper, a novel method is proposed for radar emitter signal identification, where the bispectrum estimation of radar signal is extracted and the recent hierarchical extreme learning machine (BS + H-ELM) is adopted for further feature learning and recognition. Conventional REI methods generally rely on the time-difference-of-arrival, carrier frequency, pulse width, pulse amplitude, direction-of-arrival, etc., for signal representation and recognition. However, the increasingly violent electronic confrontation and the emergence of new types of radar signals generally degrade the recognition performance. With this objective, we explore radar emitter signal representation and classification method with the high order spectrum and deep network based H-ELM. After extracting the bispectrum of radar signals, the sparse autoencoder (AE) in H-ELM is employed for feature learning. Simulations on four representative radar signals, namely, the continuous wave (CW), linear frequency modulation wave(LFM), nonlinear frequency modulation wave(NLFM) and binary phase shift keying wave (BPSK), are conducted for performance validation. In comparison to the existing multilayer ELM algorithm and the popular histogram of gradient (HOG) based feature extraction method are proved that the proposal is feasible and potentially applicable in real applications. [ABSTRACT FROM AUTHOR]

Eye blink is the most common artifact in electroencephalogram (EEG), which usually affects the performance of EEG-based applications, such as neurological aided diagnostic analysis. For low spatial resolution EEG signals, current methods generally lack of spatial filtering, leading to a degraded performance. In this paper, a novel eye blink artifact detection algorithm based on multiple EEG feature fusion and PSO optimization is proposed in a few-channel data environment. The forehead FP1 and FP2 electrodes EEGs are decomposed based on empirical mode decomposition (EMD) through autocorrelation coefficients for signal filtering. The EEG variance features are extracted by the Common Spatial Pattern (CSP) filtering to enhance the feature discrimination. The particle swarm optimization (PSO) combined with support vector machine (SVM) is applied for feature fusion and optimization. We evaluate the performance on real recorded EEG dataset by the Children's Hospital of Zhejiang University School of Medicine (CHZU). There contain EEGs with eye blink artifacts of 20 subjects. The results show that the proposed method can achieve the highest accuracy, recall rate, precision and F1 value. • An improved MEMD-CSP is developed for eye blinking artifact detection. • Autocorrelation coefficient extracted from IMF components is developed for screening. • PSO-SVM is developed to solve the influence of individual differences. • Spatial/time–frequency domain features are optimized to enhance the robustness. [ABSTRACT FROM AUTHOR]

Seizures are induced by sudden abnormal discharge of brain electrical activity. Accurate judgment of seizure type is a prerequisite for correct treatment, rational use of drugs and prognosis. In this paper, a novel feature extraction algorithm based on multi-class specific bands common spatial pattern (MSBCSP), which applies the original CSP algorithm to multi-classification tasks with the help of joint approximation diagonalization (JAD), is presented. Complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is applied to extract the energy of intrinsic mode functions (IMFs). Next, this study employs linear discriminant analysis (LDA) and random forest (RF) for feature selection. A penalized ensemble method was established to address imbalanced data and improve model generalization. Specifically, this study first introduces penalized strategy to handle imbalanced data, which punishes the majority classes by giving them lower weights while giving minority classes higher weights. Furthermore, this study introduces weighted voting strategy to combine logistic regression (LR) algorithm with LightGBM (LGB) algorithm. For evaluating the performance of proposed model, Temple University Hospital EEG dataset v1.5.0 is adopted, accuracy, kappa score, precision, recall and f1-score are computed with the value of 96.14%, 0.9335, 97.24%, 96.36% and 0.9679, respectively. [ABSTRACT FROM AUTHOR]

Heart sound classification plays a vital role in the early detection of cardiovascular disorders, especially for small primary health care clinics. Despite that much progress has been made for heart sound classification in recent years, most of them are based on conventional segmented features and shallow structure based classifiers. These conventional acoustic representation and classification methods may be insufficient in characterizing heart sound, and generally suffer from a degraded performance due to the complicated and changeable cardiac acoustic environment. In this paper, we propose a new heart sound classification method based on improved Mel-frequency cepstrum coefficient (MFCC) features and convolutional recurrent neural networks. The Mel-frequency cepstrums are firstly calculated without dividing the heart sound signal. A new improved feature extraction scheme based on MFCC is proposed to elaborate the dynamic characteristics among consecutive heart sound signals. Finally, the MFCC-based features are fed to a deep convolutional and recurrent neural network (CRNN) for feature learning and later classification task. The proposed deep learning framework can take advantage of the encoded local characteristics extracted from the convolutional neural network (CNN) and the long-term dependencies captured by the recurrent neural network (RNN). Comprehensive studies on the performance of different network parameters and different network connection strategies are presented in this paper. Performance comparisons with state-of-the-art algorithms are given for discussions. Experiments show that, for the two-class classification problem (pathological or non-pathological), a classification accuracy of 98% has been achieved on the 2016 PhysioNet/CinC Challenge database. • We propose a new MFCC-based heart sound classification method. • A new improved MFCC feature extraction scheme is developed. • CNN and RNN are combined for heart sound classification. • We show promising performance on benchmark PCG databases. [ABSTRACT FROM AUTHOR]

• We develop a multi-branch network based on ResNet-34 model, which can extract more deep level features than single network. In addition, an ASPP module is leveraged to refine the extracted deep level features by expanding the receptive field. • A depth-wise attention module is devised to fuse the extracted high-level features, which is used to fuse the features from ASPP module and the last feature extraction layer of the original ResNet-34 model that enforces the network to focus on the salient targets of interest. • We also devise a cross-attention module that learns the distinction and relationship amongst different diseases by channel and spatial attention strategy and integrates the extracted attention map via cross-fusion mode to gain the relevant features of specific-disease. Fundus diseases classification is vital for the health of human beings. However, most of existing methods detect diseases by means of single angle fundus images, which lead to the lack of pathological information. To address this limitation, this paper proposes a novel deep learning method to complete different fundus diseases classification tasks using ultra-wide field scanning laser ophthalmoscopy (SLO) images, which have an ultra-wide field view of 180–200˚. The proposed deep model consists of multi-branch network, atrous spatial pyramid pooling module (ASPP), cross-attention and depth-wise attention module. Specifically, the multi-branch network employs the ResNet-34 model as the backbone to extract feature information, where the ResNet-34 model with two-branch is followed by the ASPP module to extract multi-scale spatial contextual features by setting different dilated rates. The depth-wise attention module can provide the global attention map from the multi-branch network, which enables the network to focus on the salient targets of interest. The cross-attention module adopts the cross-fusion mode to fuse the channel and spatial attention maps from the ResNet-34 model with two-branch, which can enhance the representation ability of the disease-specific features. The extensive experiments on our collected SLO images and two publicly available datasets demonstrate that the proposed method can outperform the state-of-the-art methods and achieve quite promising classification performance of the fundus diseases. [Display omitted] [ABSTRACT FROM AUTHOR]