Showing papers by "Jiuwen Cao published in 2022"

Multicenter and Multichannel Pooling GCN for Early AD Diagnosis Based on Dual-Modality Fused Brain Network

Abstract: For significant memory concern (SMC) and mild cognitive impairment (MCI), their classification performance is limited by confounding features, diverse imaging protocols, and limited sample size. To address the above limitations, we introduce a dual-modality fused brain connectivity network combining resting-state functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI), and propose three mechanisms in the current graph convolutional network (GCN) to improve classifier performance. First, we introduce a DTI-strength penalty term for constructing functional connectivity networks. Stronger structural connectivity and bigger structural strength diversity between groups provide a higher opportunity for retaining connectivity information. Second, a multi-center attention graph with each node representing a subject is proposed to consider the influence of data source, gender, acquisition equipment, and disease status of those training samples in GCN. The attention mechanism captures their different impacts on edge weights. Third, we propose a multi-channel mechanism to improve filter performance, assigning different filters to features based on feature statistics. Applying those nodes with low-quality features to perform convolution would also deteriorate filter performance. Therefore, we further propose a pooling mechanism, which introduces the disease status information of those training samples to evaluate the quality of nodes. Finally, we obtain the final classification results by inputting the multi-center attention graph into the multi-channel pooling GCN. The proposed method is tested on three datasets (i.e., an ADNI 2 dataset, an ADNI 3 dataset, and an in-house dataset). Experimental results indicate that the proposed method is effective and superior to other related algorithms, with a mean classification accuracy of 93.05% in our binary classification tasks. Our code is available at: https://github.com/Xuegang-S.

3D residual-attention-deep-network-based childhood epilepsy syndrome classification

Abstract: Interictal electroencephalograms (EEGs) usually contain important information for epilepsy analysis and diagnosis. However, the focus of existing research has mainly been on epilepsy seizure onset detection, and only a few studies have been conducted on childhood epilepsy syndrome classification, which is usually more complicated than seizure detection. In this study, a novel 3D residual-attention-module-based deep network (AR3D) is developed to explore the spatial and time–frequency features of multichannel EEGs. The interictal EEGs of 37 patients with five typical childhood epilepsy syndromes, namely, benign childhood epilepsy with centrotemporal spikes, childhood absence epilepsy, febrile seizures plus, infantile spasms, unknown epilepsy syndrome, and one control group, are studied. The proposed AR3D algorithm, with a 97.03% F1 score, outperforms several state-of-the-art 2D and 3D convolution deep networks.

Multidimensional Feature Optimization Based Eye Blink Detection Under Epileptiform Discharges

Abstract: 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.

BECT Spike Detection Algorithm Based on Optimal Template Matching and Morphological Feature Selection

Abstract: Over 15% of children with epilepsy belong to benign childhood epilepsy with centro-temporal spikes (BECT), facing with educational difficulties. The accurate recognition of spikes in electroencephalogram (EEG) signals collected from BECT patients can help the doctor to effectively make diagnosis and give therapeutic schedule. Traditionally, template matching method can extract spike-like waves from EEG signals and is adopted by many researches. However, the patterns of the spikes appeared in different patients or different time in one patient varies greatly. The brief proposes a spike detection algorithm based on optimal template matching and morphological feature selection, which includes universal template matching, spike clustering, universal template optimization based on particle swarm optimization (PSO) algorithm and false positive spike (FPS) elimination based on spike morphological feature. Based on the testing EEG data set adopted in this brief, the sensitivity (Sen), specificity (Spe) and accuracy (AC) of the proposed algorithm reaches 98.2%, 95.1% and 96.5%, respectively.

Childhood epilepsy syndromes classification based on fused features of electroencephalogram and electrocardiogram

Abstract: The paper presents a novel algorithm to classify children's epileptic syndromes based on the fused features of electroencephalogram (EEG) and electrocardiogram (ECG). The purpose is to assess whether multimodal physiological signals could improve the classification performance of epileptic syndromes over a single physiological signal. The study is carried out on the epileptic syndromes database recorded by the Children's Hospital, Zhejiang University School of Medicine (CHZU), that includes the synchronised EEGs and ECGs of 16 children suffered from the infantile spasms (known as the WEST syndrome, named) and the childhood absence epilepsy (CAE), respectively. Experiments are conducted and compared using the EEGs and ECGs in the ictal and interictal periods. The data imbalanced issue between the ictal and interictal periods is also considered by applying a synthetic minority sample generating approach. The experimental results show that using the fused feature of EEG + ECG can achieve an average of 98.15% overall classification accuracy, which is better than using the single physiological signal.

Seizure Prediction Analysis of Infantile Spasms

Abstract: Infantile spasms (IS) is a typical childhood epileptic disorder with generalized seizures. The sudden, frequent and complex characteristics of infantile spasms are the main causes of sudden death, severe comorbidities and other adverse consequences. Effective prediction is highly critical to infantile spasms subjects, but few related studies have been done in the past. To address this, this study proposes a seizure prediction framework for infantile spasms by combining the statistical analysis and deep learning model. The analysis is conducted on dividing the continuous scalp electroencephalograms (sEEG) into 5 phases: Interictal, Preictal, Seizure Prediction Horizon (SPH), Seizure, and Postictal. The brain network of Phase-Locking Value (PLV) of 5 typical brain rhythms is constructed, and the mechanism of epileptic changes is analyzed by statistical methods. It is found that 1) the connections between the prefrontal, occipital, and central regions show a large variability at each stage of seizure transition, and 2) 4 sub-bands of brain rhythms ( $\theta $ , $\alpha $ , $\beta $ , $\gamma $ ) are predominant. Group and individual variabilities are validated by using the Resnet18 deep model on data from 25 patients with infantile spasms, where the consistent results to statistical analyses can be observed. The optimized model achieves an average of $79.78~\%$ , $94.46\%$ , $75.46\%$ accuracy, specificity, and recall rate, respectively. The method accomplishes the analysis of the synergy between infantile spasms mechanism, model, data and algorithm, providing a guideline to build an intelligent and systematic model for comprehensive IS seizure prediction.

Longitudinal study of early mild cognitive impairment via similarity-constrained group learning and self-attention based SBi-LSTM

Abstract: Alzheimer’s disease (AD) is an incurable neurodegenerative disease. Mild cognitive impairment (MCI) is often considered a critical time window for predicting early conversion to Alzheimer’s disease (AD), with approximately 80% of amnestic MCI patients developing AD within 6 years. MCI can be further categorized into two stages (i.e., early MCI (EMCI) and late MCI (LMCI)). To identify EMCI effectively and understand how it changes brain function, the brain functional connectivity network (BFCN) has been widely used. However, the conventional methods mainly focused on detection from a single time-point data, which could not discover the changes during the disease progression without using multi-time points data. Therefore, in this work, we carry out a longitudinal study based on multi-time points data to detect EMCI and validate them on two public datasets. Specifically, we first construct a similarity-constrained group network (SGN) from the resting state functional magnetic resonance imaging (rs-fMRI) data at different time-points, and then use a stacked bidirectional long short term memory (SBi-LSTM) network to extract features for longitudinal analysis. Also, we use a self-attention mechanism to leverage high-level features to further improve the detection accuracy. Evaluated on the public Alzheimer’s Disease Neuroimaging Initiative Phase II and III (ADNI-2 and ADNI-3) databases, the proposed method outperforms several state-of-the-art methods.

Cluster Embedding Joint-Probability-Discrepancy Transfer for Cross-Subject Seizure Detection

Abstract: 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.

DDistill-SR: Reparameterized Dynamic Distillation Network for Lightweight Image Super-Resolution

Abstract: Recent research on deep convolutional neural networks (CNNs) has provided a significant performance boost on efficient super-resolution (SR) tasks by trading off the performance and applicability. However, most existing methods focus on subtracting feature processing consumption to reduce the parameters and calculations without refining the immediate features, which leads to inadequate information in the restoration. In this paper, we propose a lightweight network termed DDistill-SR, which significantly improves the SR quality by capturing and reusing more helpful information in a static-dynamic feature distillation manner. Specifically, we propose a plug-in reparameterized dynamic unit (RDU) to promote the performance and inference cost trade-off. During the training phase, the RDU learns to linearly combine multiple reparameterizable blocks by analyzing varied input statistics to enhance layer-level representation. In the inference phase, the RDU is equally converted to simple dynamic convolutions that explicitly capture robust dynamic and static feature maps. Then, the information distillation block is constructed by several RDUs to enforce hierarchical refinement and selective fusion of spatial context information. Furthermore, we propose a dynamic distillation fusion (DDF) module to enable dynamic signals aggregation and communication between hierarchical modules to further improve performance. Empirical results show that our DDistill-SR outperforms the baselines and achieves state-of-the-art results on most super-resolution domains with much fewer parameters and less computational overhead. We have released the code of DDistill-SR at https://github.com/icandle/DDistill-SR .

Clustering-Guided Pairwise Metric Triplet Loss for Person Reidentification

Abstract: Most of the loss functions proposed for person reidentification (Re-ID) are expected to be easy to deploy, efficiently improve network performance, and will not introduce redundant parameters. This study proposes a no-parameter and generic clustering-guided pairwise metric triplet (CPM-Triplet) loss based on the hard sample mining triplet loss for the metric learning loss. CPM-Triplet loss deploys two metrics: 1) the Euclidean metric and 2) the cosine metric, to complementarily improve the metric learning of the model. Paralleled to the Euclidean metric, the cosine metric quantifies the sample similarity in a different way to the Euclidean metric, which takes a different perspective to explore the distribution of samples. But the pairwise metric mainly improves the precision between dissimilar samples of the same label and could not solve the problem of excessive outliers. Therefore, the clustering-guided correction term was deployed to apply to all samples with the same label to mine the similarity in the samples, while weakening the influence of outliers in CPM-Triplet loss. Experiments conducted on four benchmark data sets show that the combination of the CPM-Triplet loss and the widely used Bag-of-Tricks baseline generally outperforms the baseline and numerous state-of-the-art methods studied in this article. The source code would be available at https://github.com/weiyu-zeng/CPM-Triplet-loss.

Differentiation of MOGAD in ADEM-like presentation children based on FLAIR MRI features.

Abstract: The differences in magnetic resonance imaging (MRI) between children with classic acute disseminated encephalomyelitis (ADEM) and myelinal oligodendrocyte glycoprotein antibody associated disease (MOGAD) with ADEM-like presentation are controversial. The purpose of this study was to investigate whether the radiological characteristics of the MRI-FLAIR sequence can predict MOGAD in children with ADEM-like presentation and to further explore its imaging differences.We extracted 1041 radiomics features from MRI-FLAIR lesions. Then we used the redundancy analysis (Spearman correlation coefficient), significance test (student test or Mann-Whitney U test), least absolute contraction and selection operator (LASSO) to select potential predictors from the feature groups. The selected potential predictors and MOG antibody test results were used to fit the machine learning model for classification. Combined with feature selection and machine learning classifiers, the optimal model for each subgroup was derived. The resulting models have been evaluated using the receiver operator characteristic curve (ROC) at the lesion level and the model performance was evaluated at the case level using decision curve analysis.We retrospectively reviewed and re-diagnosed 70 ADEM-like presentation cases in our center from April 2015 to January 2020. Including 49 cases with classic ADEM and 21 cases with MOGAD. 30(43%) were female, with a median age of 5.3 years. On the four subgroups by age and gender, the area under the curve (AUC) of the optimal models were 89%, 90%, 98%, and 99%, and the MOGAD detection rates (Specificity) were 83%, 83%, 92%, and 75%, respectively.The machine learning model trained on radiomics features of MR-FLAIR images can effectively predict patients' MOGAD. This study provides a fast, objective, and quantifiable method for MOGAD diagnosis.

Screen Content Video Quality Assessment Model Using Hybrid Spatiotemporal Features

Abstract: 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.

SLPR: A Deep Learning Based Chinese Ship License Plate Recognition Framework

Abstract: Automatic ship license plate recognition (SLPR) for ship identification is of great significance to waterway shipping management. But few attention has been paid to SLPR in the past. In this paper, a novel cascaded Chinese SLPR framework consisting of the quadrangle-based ship license plate detection (QSLPD) algorithm and the rectification-based text recognition network (RTRNet) is developed. Concretely, in QSLPD algorithm, detection is performed based on the pyramid feature fusion architecture ameliorated by the proposed variable receptive field feature enhancement strategy and three task-specific output heads. In addition, a new loss function combining the dice coefficient and cross entropy is explored in the proposed SLPR which can generate significant improvement over the baseline. In RTRNet, regions of interest (RoIs) extraction and irregular text line rectification based on the vertices information predicted by QSLPD are performed before text recognition. Data augmentation are also applied to cope with the problem of limited text recognizer training data and the extremely imbalance distribution of corpus. Extensive experiments are carried out to demonstrate the reliability of the proposed cascaded SLPR framework, that can achieve the highest F-measure of 87.78% and 76.59% with IoU and TIoU metric on the collected dataset, surpasses many existing advanced methods.

Multichannel Matrix Randomized Autoencoder

Abstract: The existing randomized autoencoders are generally designed for vectorization data resulting in destroying the original structure information inevitably when dealing with multi-dimension data such as image and video. To address this issue, a one-side matrix randomized AE (OMRAE) is developed that takes the two-dimensional (2D) data as inputs directly by the linear mapping on one-side of inputs with matrix multiplication. For multichannel 2D (M2D) data, a multichannel OMRAE (OMMRAE) is proposed by training the output weights to rebuild each channel of inputs respectively. In this way, the structural information of each channel and the interaction between channels are explored. Then, a double-side structure using 2 OMMRAEs to simultaneously extracts the row and column structure information of M2D is developed. At last, a novel hierarchical matrix randomized neural networks is constructed for one-class classification where each layer passes information by bilinear mapping derived from DMMRAE. Experiments are conducted on 2 benchmark datasets for the effectiveness demonstration. Comparisons to several state-of-the-art AEs reveal that the proposed OMMRAE/DMMRAE can obtain better performance with a compact network size. The source code would be available at https://github.com/ML-HDU/MMRAE .

Vibration-based hypervelocity impact identification and localization

Abstract: Hypervelocity impact (HVI) vibration source identification and localization have found wide applications in many fields, such as manned spacecraft protection and machine tool collision damage detection and localization. In this paper, we study the synchrosqueezed transform (SST) algorithm and the texture color distribution (TCD) based HVI source identification and localization using impact images. The extracted SST and TCD image features are fused for HVI image representation. To achieve more accurate detection and localization, the optimal selective stitching features OSSST+TCD are obtained by correlating and evaluating the similarity between the sample label and each dimension of the features. Popular conventional classification and regression models are merged by voting and stacking to achieve the final detection and localization. To demonstrate the effectiveness of the proposed algorithm, the HVI data recorded from three kinds of high-speed bullet striking on an aluminum alloy plate is used for experimentation. The experimental results show that the proposed HVI identification and localization algorithm is more accurate than other algorithms. Finally, based on sensor distribution, an accurate four-circle centroid localization algorithm is developed for HVI source coordinate localization.

BECT Spike Detection Based on Novel Multichannel Data Weighted Fusion Algorithm

Abstract: 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.

Deterministic Learning-Based WEST Syndrome Analysis and Seizure Detection on ECG

Abstract: 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.

A Novel Multi-task Learning based Automatic Speech Impairment Assessment Algorithm

Abstract: Speech impairment assessment is crucial to the treatment evaluation of speech therapy. The current evaluation methods mainly depend on speech-language pathologists (SLPs). Automatic speech impairment assessment (ASIA), especially the regression of severity scores, has not received enough attention. So we present a novel ASIA algorithm which based on the multi-task leaning with joint severity level classification and score regression. Owing to the auxiliary classification task, the precision of the severity score prediction can be improved effectively. In addition, the residual network (ResNet) and the long short-term memory (LSTM) are cascaded as the backbone. The performance of the model is demonstrated on the Mandarin AphasiaBank dataset and the experiments show that the algorithm achieves promising performance.

Coherence Matrix Based Early Infantile Epileptic Encephalopathy Analysis with ResNet

Abstract: EIEE syndrome, known as early infantile epileptic encephalopathy, is considered to be the earliest onset form of age-dependent epileptic encephalopathy. The main manifestations are tonic-spasmodic seizures in early infancy, accompanied by burst suppressive electroencephalogram (EEG) patterns and severe psychomotor disturbances, with structural brain lesions in some cases. Specific to EIEE syndrome, this paper presents a comprehensive analysis of EEG features at three different periods: pre-seizure, seizure and post-seizure. Coherent features are extracted to characterize EEG signals in EIEE syndrome, and Kruskal-Wallis H Test and Gradient-weighted Class Activation Mapping (Grad-CAM) are used to investigate and visualize the significance of features in different frequency band for distinguishing the three stages. The study found that activity synchrony between temporal and central regions decreased significantly in the $$\gamma $$ band during seizures. And the coherence feature in the $$\gamma $$ band combined with the ResNet18-based seizure detection model achieved an accuracy of 91.86%. It is believed that changes in the $$\gamma $$ band can be considered as a biomarker of seizure cycle changes in EIEE syndrome.

Incremental Quaternion Random Neural Networks

Abstract: Quaternion, as a hypercomplex number with three imaginary elements, is effective in characterizing three- and four-dimensional vector signals. Quaternion neural networks with randomly generated quaternions as the hidden node parameters become attractive for the good learning capability and generalization performance. In this paper, a novel incremental quaternion random neural network trained by extreme learning machine (IQ-ELM) is proposed. To fully exploit the second-order Q-properness statistic of quaternion random variables, the augmented quaternion vector is further applied in IQ-ELM (IAQ-ELM) for hypercomplex data learning. The network is constructed by gradually adding the hidden neuron one-by-one, where the output weight is optimized by minimizing the residual error based on the fundamental of the generalized HR calculus (GHR) of quaternion variable function. Experiments on multidimensional chaotic system regression, aircraft trajectory tracking, face and image recognition are conducted to show the effectiveness of IQ-ELM and IAQ-ELM. Comparisons to two popular quaternion RNNs, Schmidt NN (SNN) and random vector functional-link net (RVFL), are also provided to show the feasibility and superiority of using quaternions in RNN for incremental learning.