CPSS: Fusing consistency regularization and pseudo-labeling techniques for semi-supervised deep cardiovascular disease detection using all unlabeled electrocardiograms.

BACKGROUND AND OBJECTIVE: Deep learning usually achieves good performance in the supervised way, which requires a large amount of labeled data. However, manual labeling of electrocardiograms (ECGs) is laborious that requires much medical knowledge. Semi-supervised learning (SSL) provides an effective way of leveraging unlabeled data to improve model performance, providing insight for solving this problem. The objective of this study is to improve the performance of cardiovascular disease (CVD) detection by fully utilizing unlabeled ECG. METHODS: A novel SSL algorithm fusing consistency regularization and pseudo-labeling techniques (CPSS) is proposed. CPSS consists of supervised learning and unsupervised learning. For supervised learning, the labeled ECGs are mapped into prediction vectors by the classifier. The cross-entropy loss function is used to optimize the classifier. For unsupervised learning, the unlabeled ECGs are weakly and strongly augmented, and a consistency loss is used to minimize the difference between the classifier's predictions for the two augmentations. Pseudo-labeling techniques include positive pseudo-labeling (PL) and ranking-based negative pseudo-labeling (RNL). PL introduces pseudo-labels for data with high prediction confidence. RNL assigns negative pseudo-labels to the lower-ranked categories in the prediction vectors to leverage data with low prediction confidence. In this study, VGGNet and ResNet are used as classifiers, which are jointly optimized by labeled and unlabeled ECGs. RESULTS: CPSS has been validated on several databases. With the same number of labeled ECGs (10%), it improves the accuracies over pure supervised learning by 13.59%, 4.60%, and 5.38% in the CPSC2018, PTB-XL, and Chapman databases, respectively. CPSS achieves comparable results to the fully supervised method with only 10% of labeled ECGs, which reduces the labeling workload by 90%. In addition, to verify the practicality of CPSS, a cardiovascular disease monitoring system is designed by heterogeneously deploying the trained classifiers on an SoC (system-on-a-chip), which can detect CVD in real time. CONCLUSION: The results of this study indicate that the proposed CPSS can significantly improve the performance of CVD detection using unlabeled ECG, which reduces the burden of ECG labeling in deep learning. In addition, the designed monitoring system makes the proposed CPSS promising for real-world applications.

A High-Performance Pixel-Level Fully Pipelined Hardware Accelerator for Neural Networks.

The design of convolutional neural network (CNN) hardware accelerators based on a single computing engine (CE) architecture or multi-CE architecture has received widespread attention in recent years. Although this kind of hardware accelerator has advantages in hardware platform deployment flexibility and development cycle, it is still limited in resource utilization and data throughput. When processing large feature maps, the speed can usually only reach 10 frames/s, which does not meet the requirements of application scenarios, such as autonomous driving and radar detection. To solve the above problems, this article proposes a full pipeline hardware accelerator design based on pixel. By pixel-by-pixel strategy, the concept of the layer is downplayed, and the generation method of each pixel of the output feature map (Ofmap) can be optimized. To pipeline the entire computing system, we expand each layer of the neural network into hardware, eliminating the buffers between layers and maximizing the effect of complete connectivity across the entire network. This approach has yielded excellent performance. Besides that, as the pixel data stream is a fundamental paradigm in image processing, our fully pipelined hardware accelerator is universal for various CNNs (MobileNetV1, MobileNetV2 and FashionNet) in computer vision. As an example, the accelerator for MobileNetV1 achieves a speed of 4205.50 frames/s and a throughput of 4787.15 GOP/s at 211 MHz, with an output latency of 0.60 ms per image. This extremely shorts processing time and opens the door for AI's application in high-speed scenarios.

Temporary Consolidation of Marine Artifact Based on Polyvinyl Alcohol/Tannic Acid Reversible Hydrogel.

Underwater artefacts are vulnerable to damage and loss of archaeological information during the extraction process. To solve this problem, it is necessary to apply temporary consolidation materials to fix the position of marine artifacts. A cross-linked network hydrogel composed of polyvinyl alcohol (PVA), tannic acid (TA), borax, and calcium chloride has been created. Four hydrogels with varying concentrations of tannic acid were selected to evaluate the effect. The hydrogel exhibited exceptional strength, high adhesion, easy removal, and minimal residue. The PVA/TA hydrogel and epoxy resin were combined to extract waterlogged wooden artifacts and marine archaeological ceramics from a 0.4 m deep tank. This experiment demonstrates the feasibility of using hydrogel for the extraction of marine artifacts.

ST-ReGE: A Novel Spatial-Temporal Residual Graph Convolutional Network for CVD.

Recently, deep learning (DL) has enabled rapid advancements in electrocardiogram (ECG)-based automatic cardiovascular disease (CVD) diagnosis. Multi-lead ECG signals have lead systems based on the potential differences between electrodes placed on the limbs and the chest. When applying DL models, ECG signals are usually treated as synchronized signals arranged in Euclidean space, which is the abstraction and generalization of real space. However, conventional DL models typically merely focus on temporal features when analyzing Euclidean data. These approaches ignore the spatial relationships of different leads, which are physiologically significant and useful for CVD diagnosis because different leads represent activities of specific heart regions. These relationships derived from spatial distributions of electrodes can be conveniently created in non-Euclidean data, making multi-lead ECGs better conform to their nature. Considering graph convolutional network (GCN) adept at analyzing non-Euclidean data, a novel spatial-temporal residual GCN for CVD diagnosis is proposed in this work. ECG signals are firstly divided into single-channel patches and transferred into nodes, which will be connected by spatial-temporal connections. The proposed model employs residual GCN blocks and feed-forward networks to alleviate over-smoothing and over-fitting. Moreover, residual connections and patch dividing enable the capture of global and detailed spatial-temporal features. Experimental results reveal that the proposed model achieves at least a 5.85% and 6.80% increase in F1 over other state-of-the-art algorithms with similar parameters and computations in both PTB-XL and Chapman databases. It indicates that the proposed model provides a promising avenue for intelligent diagnosis with limited computing resources.

Removal of Calcareous Concretions from Marine Archaeological Ceramics by Means of a Stimuli-Responsive Hydrogel.

The presence of calcareous concretions on the surface of marine archaeological ceramics is a frequently observed phenomenon. It is necessary to remove these materials when the deposits obscure the feature of ceramics. Unfortunately, calcareous concretions provide distinctive documentation of the burning history of ceramics. The interaction of acid solution or detachment of the deposit layers in physical ways leads to the loss of archeological information. To prevent the loss of archeological information and to achieve precise and gentle concretion removal, responsive hydrogel cleaning systems have been developed. The hydrogels synthesized are composed of networks of poly(vinyl acetate)/sodium alginate that exhibit desirable water retention properties, are responsive to Ca2+ ions, and do not leave any residues after undergoing cleaning treatment. Four distinct compositions were selected. The study of water retention properties involved quantifying the weight changes. The composition was obtained from Fourier transform infrared spectra. The microstructure was obtained from scanning electron microscopy. The mechanical properties were obtained from rheological measurements. To demonstrate both the efficiency and working mechanism of the selected hydrogels, a representative study of mocked samples is presented first. After selecting the most appropriate hydrogel composite, a cleaning process was implemented on the marine archaeological ceramics. This article demonstrates the advantages of stimuli-responsive hydrogels in controlling the release of acid solution release, thereby surpassing the limitations of traditional cleaning methods.

Optimized Solutions of Electrocardiogram Lead and Segment Selection for Cardiovascular Disease Diagnostics.

Most of the existing multi-lead electrocardiogram (ECG) detection methods are based on all 12 leads, which undoubtedly results in a large amount of calculation and is not suitable for the application in portable ECG detection systems. Moreover, the influence of different lead and heartbeat segment lengths on the detection is not clear. In this paper, a novel Genetic Algorithm-based ECG Leads and Segment Length Optimization (GA-LSLO) framework is proposed, aiming to automatically select the appropriate leads and input ECG length to achieve optimized cardiovascular disease detection. GA-LSLO extracts the features of each lead under different heartbeat segment lengths through the convolutional neural network and uses the genetic algorithm to automatically select the optimal combination of ECG leads and segment length. In addition, the lead attention module (LAM) is proposed to weight the features of the selected leads, which improves the accuracy of cardiac disease detection. The algorithm is validated on the ECG data from the Huangpu Branch of Shanghai Ninth People's Hospital (defined as the SH database) and the open-source Physikalisch-Technische Bundesanstalt diagnostic ECG database (PTB database). The accuracy for detection of arrhythmia and myocardial infarction under the inter-patient paradigm is 99.65% (95% confidence interval: 99.20-99.76%) and 97.62% (95% confidence interval: 96.80-98.16%), respectively. In addition, ECG detection devices are designed using Raspberry Pi, which verifies the convenience of hardware implementation of the algorithm. In conclusion, the proposed method achieves good cardiovascular disease detection performance. It selects the ECG leads and heartbeat segment length with the lowest algorithm complexity while ensuring classification accuracy, which is suitable for portable ECG detection devices.

A fully-mapped and energy-efficient FPGA accelerator for dual-function AI-based analysis of ECG.

In this paper, a fully-mapped field programmable gate array (FPGA) accelerator is proposed for artificial intelligence (AI)-based analysis of electrocardiogram (ECG). It consists of a fully-mapped 1-D convolutional neural network (CNN) and a fully-mapped heart rate estimator, which constitute a complementary dual-function analysis. The fully-mapped design projects each layer of the 1-D CNN to a hardware module on an Intel Cyclone V FPGA, and a virtual flatten layer is proposed to effectively bridge the feature extraction layers and fully-connected layer. Also, the fully-mapped design maximizes computational parallelism to accelerate CNN inference. For the fully-mapped heart rate estimator, it performs pipelined transformations, self-adaptive threshold calculation, and heartbeat count on the FPGA, without multiplexed usage of hardware resources. Furthermore, heart rate calculation is elaborately analyzed and optimized to remove division and acceleration, resulting in an efficient method suitable for hardware implementation. According to our experiments on 1-D CNN, the accelerator can achieve 43.08× and 8.38× speedup compared with the software implementations on ARM-Cortex A53 quad-core processor and Intel Core i7-8700 CPU, respectively. For the heart rate estimator, the hardware implementations are 25.48× and 1.55× faster than the software implementations on the two aforementioned platforms. Surprisingly, the accelerator achieves an energy efficiency of 63.48 GOPS/W, which obviously surpasses existing studies. Considering its power consumption is only 67.74 mW, it may be more suitable for resource-limited applications, such as wearable and portable devices for ECG monitoring.

The diagnostic potential of plasma SCUBE-1 concentration for pulmonary embolism: A pilot study.

INTRODUCTION: This study aimed to investigate the potential application of plasma signal peptide-complement C1r/C1s, Uegf and Bmp1-epidermal growth factor domain-containing protein 1 (SCUBE-1) as a biomarker in the diagnosis of pulmonary embolism (PE). METHODS: This cross-sectional study enrolled 177 patients who underwent PE diagnostic test and 87 healthy controls. The results of CT pulmonary angiogram (CTPA) were used as reference standards for PE diagnosis. The levels of SCUBE-1 and D-dimer in participants' plasma were detected with enzyme-linked immunosorbent assay and compared among patients with confirmed PE, suspicious PE and healthy controls. The diagnostic values were analysed using receiver operating characteristic (ROC) curve analysis. In addition, differences in plasma SCUBE-1 levels were compared among patients with different risk stratifications. RESULTS: The plasma SCUBE-1 concentration levels in patients with CTPA confirmed PE (14.28 ± 7.74 ng/ml) was significantly higher than those in the suspicious patients (11.11 ± 4.48 ng/ml) and in healthy control (4.40 ± 3.23 ng/ml) (P < 0.01). ROC curve analysis showed that at the cut-off of 7.789 ng/ml, SCUBE-1 has significant diagnostic value in differentiating PE patients from healthy control (AUC = 0.919, sensitivity = 81.25%, specificity = 92.13%), and the performance is more accurate than D-dimer (cut-off 273.4 ng/ml, AUC = 0.648, sensitivity = 65.75%, specificity = 67.42%). The combination of D-dimer with SCUBE-1 did not further improve the diagnostic value. However, SCUBE-1 did not show significant diagnostic value in identifying PE among suspicious patients There was no significant difference in SCUBE-1 level among different risk groups (P > 0.05). CONCLUSION: We believe that SCUBE-1 could be a potential coagulation-related marker for the diagnosis of PE.

A joint cross-dimensional contrastive learning framework for 12-lead ECGs and its heterogeneous deployment on SoC.

The utilization of unlabeled electrocardiogram (ECG) data is always a critical topic in artificial intelligence healthcare, as the manual annotation for ECG data is a time-consuming task that requires much medical expertise. The recent development of self-supervised learning, especially contrastive learning, has provided helpful inspirations to solve this problem. In this paper, a joint cross-dimensional contrastive learning algorithm for unlabeled 12-lead ECGs is proposed. Unlike existing studies about ECG contrastive learning, our algorithm can simultaneously exploit unlabeled 1-dimensional ECG signals and 2-dimensional ECG images. A cross-dimensional contrastive learning method enhances the interaction between 1-dimensional and 2-dimensional ECG data, resulting in a more effective self-supervised feature learning. Combining this cross-dimensional contrastive learning, a 1-dimensional contrastive learning with ECG-specific transformations is employed to constitute a joint model. To pre-train this joint model, a new hybrid contrastive loss balances the 2 algorithms and uniformly describes the pre-training target. In the downstream classification task, the features learned by our algorithm shows impressive advantages. Compared with other representative methods, it achieves a at least 5.99% increase in accuracy. For real-world applications, an efficient heterogenous deployment on a "system-on-a-chip" (SoC) is designed. According to our experiments, the model can process 12-lead ECGs in real-time on the SoC. Furthermore, this heterogenous deployment can achieve a 14 × faster inference than the pure software deployment on the same SoC. In summary, our algorithm is a good choice for unlabeled 12-lead ECG utilization, the proposed heterogenous deployment makes it more practical in real-world applications.

New ECG Compression Method for Portable ECG Monitoring System Merged with Binary Convolutional Auto-Encoder and Residual Error Compensation.

In the past few years, deep learning-based electrocardiogram (ECG) compression methods have achieved high-ratio compression by reducing hidden nodes. However, this reduction can result in severe information loss, which will lead to poor quality of the reconstructed signal. To overcome this problem, a novel quality-guaranteed ECG compression method based on a binary convolutional auto-encoder (BCAE) equipped with residual error compensation (REC) was proposed. In traditional compression methods, ECG signals are compressed into floating-point numbers. BCAE directly compresses the ECG signal into binary codes rather than floating-point numbers, whereas binary codes take up fewer bits than floating-point numbers. Compared with the traditional floating-point number compression method, the hidden nodes of the BCAE network can be artificially increased without reducing the compression ratio, and as many hidden nodes as possible can ensure the quality of the reconstructed signal. Furthermore, a novel optimization method named REC was developed. It was used to compensate for the residual between the ECG signal output by BCAE and the original signal. Complemented with the residual error, the restoration of the compression signal was improved, so the reconstructed signal was closer to the original signal. Control experiments were conducted to verify the effectiveness of this novel method. Validated by the MIT-BIH database, the compression ratio was 117.33 and the root mean square difference (PRD) was 7.76%. Furthermore, a portable compression device was designed based on the proposed algorithm using Raspberry Pi. It indicated that this method has attractive prospects in telemedicine and portable ECG monitoring systems.

Pathogenic variants identified using whole-exome sequencing in Chinese patients with primary ciliary dyskinesia.

The genetic factors contributing to primary ciliary dyskinesia (PCD), a rare autosomal recessive disorder, remain elusive for ~20%-35% of patients with complex and abnormal clinical phenotypes. Our study aimed to identify causative variants of PCD-associated pathogenic candidate genes using whole-exome sequencing (WES). All patients were diagnosed with PCD based on clinical phenotype or transmission electron microscopy images of cilia. WES and bioinformatic analysis were then conducted on patients with PCD. Identified candidate variants were validated by Sanger sequencing. Pathogenicity of candidate variants was then evaluated using in silico software and the American College of Medical Genetics and Genomics (ACMG) database. In total, 13 rare variants were identified in patients with PCD, among which were three homozygous causative variants (including one splicing variant) in the PCD-associated genes CCDC40 and DNAI1. Moreover, two stop-gain heterozygous variants of DNAAF3 and DNAH1 were classified as pathogenic variants based on the ACMG criteria. This study identified novel potential pathogenic genetic factors associated with PCD. Noteworthy, the patients with PCD carried multiple rare causative gene variants, thereby suggesting that known causative genes along with other functional genes should be considered for such heterogeneous genetic disorders.

A Multi-Classification Hybrid Quantum Neural Network Using an All-Qubit Multi-Observable Measurement Strategy.

Quantum machine learning is a promising application of quantum computing for data classification. However, most of the previous research focused on binary classification, and there are few studies on multi-classification. The major challenge comes from the limitations of near-term quantum devices on the number of qubits and the size of quantum circuits. In this paper, we propose a hybrid quantum neural network to implement multi-classification of a real-world dataset. We use an average pooling downsampling strategy to reduce the dimensionality of samples, and we design a ladder-like parameterized quantum circuit to disentangle the input states. Besides this, we adopt an all-qubit multi-observable measurement strategy to capture sufficient hidden information from the quantum system. The experimental results show that our algorithm outperforms the classical neural network and performs especially well on different multi-class datasets, which provides some enlightenment for the application of quantum computing to real-world data on near-term quantum processors.

Tourism Experience and Construction of Personalized Smart Tourism Program Under Tourist Psychology.

The present work aims to boost tourism development in China, grasp the psychology of tourists at any time, and provide personalized tourist services. The research object is the tourism industry in Macau. In particular, tourists' experiences are comprehensively analyzed in terms of dining, living, traveling, sightseeing, shopping, and entertaining as per their psychological changes using approaches including big data analysis, literature analysis, and field investigation. In this case, a model of tourism experience formation path is summarized, and a smart travel solution is proposed based on psychological experience. In the end, specific and feasible suggestions are put forward for the Macau tourism industry. Results demonstrate that the psychology-based smart travel solution exerts a significant impact on tourists' tourism experience. Specifically, the weight of secular tourism experience is 0.523, the weight of aesthetic tourism experience is 0.356, and the weight of stimulating tourism experience is 0.121. Tourists prefer travel destinations with excellent urban security and scenic authenticity. They give the two indexes comprehensive scores of 75.14 points and 73.12 points, respectively. The proposed smart travel solution can grasp the psychology of tourists and enhance their tourism experiences. It has strong practical and guiding significances, which can promote constructing smart travel services in Macau and enhancing tourism experiences.

EvoMBN: Evolving Multi-Branch Networks on Myocardial Infarction Diagnosis Using 12-Lead Electrocardiograms.

Multi-branch Networks (MBNs) have been successfully applied to myocardial infarction (MI) diagnosis using 12-lead electrocardiograms. However, most existing MBNs share a fixed architecture. The absence of architecture optimization has become a significant obstacle to a more accurate diagnosis for these MBNs. In this paper, an evolving neural network named EvoMBN is proposed for MI diagnosis. It utilizes a genetic algorithm (GA) to automatically learn the optimal MBN architectures. A novel fixed-length encoding method is proposed to represent each architecture. In addition, the crossover, mutation, selection, and fitness evaluation of the GA are defined to ensure the architecture can be optimized through evolutional iterations. A novel Lead Squeeze and Excitation (LSE) block is designed to summarize features from all the branch networks. It consists of a fully-connected layer and an LSE mechanism that assigns weights to different leads. Five-fold inter-patient cross validation experiments on MI detection and localization are performed using the PTB diagnostic database. Moreover, the model architecture learned from the PTB database is transferred to the PTB-XL database without any changes. Compared with existing studies, our EvoMBN shows superior generalization and the efficiency of its flexible architecture is suitable for auxiliary MI diagnosis in real-world.

DMMAN: A two-stage audio-visual fusion framework for sound separation and event localization.

Videos are used widely as the media platforms for human beings to touch the physical change of the world. However, we always receive the mixed sound from the multiple sound objects, and cannot distinguish and localize the sounds as the separate entities in videos. In order to solve this problem, a model named the Deep Multi-Modal Attention Network (DMMAN), is established to model the unconstrained video datasets for further finishing the sound source separation and event localization tasks in this paper. Based on the multi-modal separator and multi-modal matching classifier module, our model focuses on the sound separation and modal synchronization problems using two stage fusion of the sound and visual features. To link the multi-modal separator and multi-modal matching classifier modules, the regression and classification losses are employed to build the loss function of the DMMAN. The estimated spectrum masks and attention synchronization scores calculated by the DMMAN can be easily generalized to the sound source and event localization tasks. The quantitative experimental results show the DMMAN not only separates the high quality of the sound sources evaluated by Signal-to-Distortion Ratio and Signal-to-Interference Ratio metrics, but also is suitable for the mixed sound scenes that are never heard jointly. Meanwhile, DMMAN achieves better classification accuracy than other contrast baselines for the event localization tasks.

A Real Time QRS Detection Algorithm Based on ET and PD Controlled Threshold Strategy.

As one of the important components of electrocardiogram (ECG) signals, QRS signal represents the basic characteristics of ECG signals. The detection of QRS waves is also an essential step for ECG signal analysis. In order to further meet the clinical needs for the accuracy and real-time detection of QRS waves, a simple, fast, reliable, and hardware-friendly algorithm for real-time QRS detection is proposed. The exponential transform (ET) and proportional-derivative (PD) control-based adaptive threshold are designed to detect QRS-complex. The proposed ET can effectively narrow the magnitude difference of QRS peaks, and the PD control-based method can adaptively adjust the current threshold for QRS detection according to thresholds of previous two windows and predefined minimal threshold. The ECG signals from MIT-BIH databases are used to evaluate the performance of the proposed algorithm. The overall sensitivity, positive predictivity, and accuracy for QRS detection are 99.90%, 99.92%, and 99.82%, respectively. It is also implemented on Altera Cyclone V 5CSEMA5F31C6 Field Programmable Gate Array (FPGA). The time consumed for a 30-min ECG record is approximately 1.3 s. It indicates that the proposed algorithm can be used for wearable heart rate monitoring and automatic ECG analysis.

The reconstruction of the symmetry between sublattices: a strategy to improve the transport properties of edge-defective graphene nanoribbon transistors.

A numerical study that combines device simulation and first-principle calculations is performed, aiming to alleviate the performance degradation of graphene nanoribbon field-effect devices with edge defects. We believe that investigating the symmetry between the sublattices of graphene is a novel approach to understand this key problem. The results show that the edge defects that break the symmetry between the sublattices of graphene cause more severe degradation of the device performance because they induce highly localized electronic states, which dramatically affect the transport of carriers. We propose a strategy to alleviate the localization of electronic states by rebuilding the symmetry between the sublattices. This strategy can be realized by introducing foreign radicals to modify the defective edge. A stability analysis is performed to find the most stable modified structures. The final effect of our strategy on the corresponding devices demonstrates that it can effectively address specific edge defects and remarkably improve the ON-state current and subthreshold swing.

Multi-information fusion neural networks for arrhythmia automatic detection.

BACKGROUND AND OBJECTIVES: . The electrocardiograms (ECGs) are widely used to diagnose a variety of arrhythmias. Generally, the abnormalities of ECG signals mainly consist of ill-shaped ECG beat morphologies and irregular intervals. The ill-shaped ECG beat morphologies represent morphological information, while the irregular intervals denote the temporal information of ECG signals. But it is difficult to utilize morphological information and temporal information simultaneously when dealing with single ECG heartbeats, because RR interval is not contained in a single short heartbeat. Therefore, to handle this problems, a novel Multi-information Fusion Convolutional Bidirectional Recurrent Neural Network (MF-CBRNN) is proposed for arrhythmia automatic detection. METHODS: . The MF-CBRNN is designed with two parallel hybrid branches that can simultaneously focus on the beat-based information in the ECG beats and the segment-based information in the adjacent segments of the beats. A single ECG beat provides the morphological information. At the same time, the adjacent segment of the ECG beat enriches the temporal information, so the two branches are designed to exploit the multiple information contained in ECGs. Furthermore, a combination of convolutional neural networks (CNNs) and a bidirectional long short memory (BLSTM) in each branch is utilized to capture the information from the two inputs. And all the features extracted from the two branches are fused for information aggregation. RESULTS: . To evaluate the performance of the proposed model, the ECG signals from MIT-BIH databases are used for intra-patient and inter-patient paradigms. The proposed model yields an accuracy of 99.56% and an F1-score of 96.40% under the intra-patient paradigm. And it obtains an overall accuracy of 96.77% and F1-score of 77.83% under the inter-patient paradigm. CONCLUSIONS: . Compared with other studies on arrhythmia detection, our method achieves a state-of-the-art performance. It indicates that the proposed model is a promising arrhythmia detection algorithm for computer-aided diagnostic systems.

Classification of VLF/LF Lightning Signals Using Sensors and Deep Learning Methods.

Lightning waveform plays an important role in lightning observation, location, and lightning disaster investigation. Based on a large amount of lightning waveform data provided by existing real-time very low frequency/low frequency (VLF/LF) lightning waveform acquisition equipment, an automatic and accurate lightning waveform classification method becomes extremely important. With the widespread application of deep learning in image and speech recognition, it becomes possible to use deep learning to classify lightning waveforms. In this study, 50,000 lightning waveform samples were collected. The data was divided into the following categories: positive cloud ground flash, negative cloud ground flash, cloud ground flash with ionosphere reflection signal, positive narrow bipolar event, negative narrow bipolar event, positive pre-breakdown process, negative pre-breakdown process, continuous multi-pulse cloud flash, bipolar pulse, skywave. A multi-layer one-dimensional convolutional neural network (1D-CNN) was designed to automatically extract VLF/LF lightning waveform features and distinguish lightning waveforms. The model achieved an overall accuracy of 99.11% in the lightning dataset and overall accuracy of 97.55% in a thunderstorm process. Considering its excellent performance, this model could be used in lightning sensors to assist in lightning monitoring and positioning.

MFB-CBRNN: A Hybrid Network for MI Detection Using 12-Lead ECGs.

This paper proposes a novel hybrid network named multiple-feature-branch convolutional bidirectional recurrent neural network (MFB-CBRNN) for myocardial infarction (MI) detection using 12-lead ECGs. The model efficiently combines convolutional neural network-based and recurrent neural network-based structures. Each feature branch consists of several one-dimensional convolutional and pooling layers, corresponding to a certain lead. All the feature branches are independent from each other, which are utilized to learn the diverse features from different leads. Moreover, a bidirectional long short term memory network is employed to summarize all the feature branches. Its good ability of feature aggregation has been proved by the experiments. Furthermore, the paper develops a novel optimization method, lead random mask (LRM), to alleviate overfitting and implement an implicit ensemble like dropout. The model with LRM can achieve a more accurate MI detection. Class-based and subject-based fivefold cross validations are both carried out using Physikalisch-Technische Bundesanstalt diagnostic database. Totally, there are 148 MI and 52 healthy control subjects involved in the experiments. The MFB-CBRNN achieves an overall accuracy of 99.90% in class-based experiments, and an overall accuracy of 93.08% in subject-based experiments. Compared with other related studies, our algorithm achieves a comparable or even better result on MI detection. Therefore, the MFB-CBRNN has a good generalization capacity and is suitable for MI detection using 12-lead ECGs. It has a potential to assist the real-world MI diagnostics and reduce the burden of cardiologists.