Search

Abstract Tourist movement patterns among attractions are complex and variable, and understanding such patterns can help manage tourist destinations more effectively. However, previous studies on tourist movement utilising complex networks have not explored the network motif approach comprehensively. Therefore, we adopted a network motif approach using social media data to extract and analyse motifs in a city network. This study analyses the attractions corresponding to the nodes in each motif, revealing the connection patterns between these attractions. We also discuss motifs between attractions with different types and titles. Popular attractions play a significant role in a local network while other attractions serve distinct functions within the network. This study’s findings enhance the significance of network motifs in examining tourist movement and deepen the understanding of recurring movement patterns between attractions. Moreover, they assist managers in developing policy tools for intelligent tourism destination marketing and planning that cater to tourists’ needs.

Abstract Current strategies to improve the throughput of continuous directed evolution technologies often involve complex mechanical fluid‐controlling system or robotic platforms, which limits their popularization and application in general laboratories. Inspired by our previous study on bacterial range expansion, in this study, we report a system termed SPACE for rapid and extensively parallelizable evolution of biomolecules by introducing spatial dimensions into the landmark phage‐assisted continuous evolution system. Specifically, M13 phages and chemotactic Escherichia coli cells were closely inoculated onto a semisolid agar. The phages came into contact with the expanding front of the bacterial range, and then comigrated with the bacteria. This system leverages competition over space, wherein evolutionary progress is closely associated with the production of spatial patterns, allowing the emergence of improved or new protein functions. In a prototypical problem, SPACE remarkably simplified the process and evolved the promoter recognition of T7 RNA polymerase (RNAP) to a library of 96 random sequences in parallel. These results establish SPACE as a simple, easy to implement, and massively parallelizable platform for continuous directed evolution in general laboratories.

To study the effects of graphene oxide (GO) dimension on the mechanical and dielectric performance of epoxy resin (EP), two kinds of GO with different dimensions were selected and characterized by Fourier transform infrared spectrometer (FTIR), transmission electron microscopy (TEM), scanning electron microscopy (SEM) and X-ray diffractometry (XRD). The results showed that the two kinds of GO had similar chemical structure, but different average particle sizes of 190.1 nm (SGO) and 1510 nm (MGO), respectively. The mechanical properties, dynamic mechanical properties, and dielectric properties of the EP/GO composites are analyzed using electronic universal testing machine, Rockwell hardness tester, dynamic mechanical analysis (DMA) and broadband dielectric impedance relaxation spectrometer. Results revealed that the GO dimension plays important role in determining the dispersion of the filler in EP matrix, and further determines the mechanical properties and dielectric performance of the composites. EP/SGO with better dispersion of GO in the EP matrix exhibits better mechanical properties, higher glass transition temperature, higher dielectric constant. Related mechanism has been proposed. [ABSTRACT FROM AUTHOR]

In this paper, a magnetic field-assisted WEDM-ECM complex process (MF-assisted WEDM-ECM) is proposed to solve the problems of traditional low-speed wire electrical discharge machining (WEDM-LS) of NiTi SMA, including low machining efficiency, high workpiece surface roughness, and large recast layer thickness. Electrolytic machining is used to remove the recast layer, and magnetic field-assisted method is used to reduce the surface roughness and improve the machining efficiency. Firstly, based on the double-layer model the removal thickness of recast layer is deduced, and the removal thickness of recast layer under different machining parameters is analyzed. In addition, based on the influence of magnetic field on discharge channel, the single spark crater analytical model is modified. Based on this model, combined with the electrochemical corrosion rate and the wire electrode vibration response equation, a continuous discharge corrosion mathematical model is established to predict the surface topography of the workpiece, and the simulation analysis is carried out. Finally, comparison of the experimental results of traditional WEDM-LS, MF-assisted WEDM-LS, ultrasonic vibration (USV)-MF complex-assisted WEDM-LS, and MF-assisted WEDM-ECM in machining NiTi SMA reveals that MF-assisted WEDM-ECM has the best machining performance. Compared with traditional WEDM-LS, the material removal rate (MRR) of MF-assisted WEDM-ECM was increased by 26.2%, the surface roughness value was reduced by 19.2%, and the thickness of recast layer was reduced by 44.25%. The average errors of the experimental and predicted values of workpiece surface roughness, recast layer removal thickness, and kerf width are 8.3%, 6.41%, and 3.58%, respectively, which verified the accuracy of the theoretical model proposed in this paper. [ABSTRACT FROM AUTHOR]

Brain neural networks are regarded as dynamical systems in neural science, in which memories are interpreted as attractors of the systems. Mathematically, ordinary differential equations (ODEs) can be utilized to describe dynamical systems. Any ODE that is employed to describe the dynamics of a neural network can be called a neuralODE. Inspired by rethinking the nonlinear representation ability of existing artificial neural networks together with the functions of columns in the neocortex, this paper proposes a theory of memory-based neuralODE, which is composed of two novel artificial neural network models: nmODE and ϵ -net, and two learning algorithms: nmLA and ϵ -LA. The nmODE (neural memory Ordinary Differential Equation) is designed with a special structure that separates learning neurons from memory neurons, making its dynamics clear. Given any external input, the nmODE possesses the global attractor property and is thus embedded with a memory mechanism. The nmODE establishes a nonlinear mapping from the external input to its associated attractor and does not have the problem of learning features homeomorphic to the input data space, as occurs frequently in most existing neuralODEs. The nmLA (neural memory Learning Algorithm) is developed by proposing an interesting three-dimensional inverse ODE (invODE) and has advantages in memory and parameter efficiency. The proposed ϵ -net is a discrete version of the nmODE, which is particularly feasible for digital computing. The proposed ϵ -LA (ϵ learning algorithm) requires no prior knowledge of the number of network layers. Both nmLA and ϵ -LA have no problem with gradient vanishing. Experimental results show that the proposed theory is comparable to state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Transcatheter aortic valve replacement (TAVR) has recently emerged as an effective treatment for patients with severe symptomatic aortic stenosis. Accurate prediction of postoperative conduction disturbance after TAVR is crucial for successful surgery planning. Most current risk prediction models for conduction disturbance rely on monomodal data methods, such as images. However, predicting conduction disturbance based on multimodal features becomes a challenge. We propose a new memory-augmented graph neural network (MemGCN) to deal with this challenge. The GCN in the proposed network is constructed by taking the patient's image features as nodes' features and creating the edges based on the patient's clinical features, thus effectively fusing the patients' images and clinical feature information. Moreover, the traditional GCN model trains feature extractors and GCNs independently due to limited computing resources, which can negatively impact the feature learning process. We introduce the memory bank and propose a local update learning algorithm to address this issue. Specifically, the proposed method connects the feature extractors and GCN through a memory bank. The CNN module extracts patients' features from their ECG images and updates the memory bank. The GCN constructs the graph using the patients' characteristics in the memory bank as node features. In this way, the node features are cached in the memory bank and gradually updated at each optimization iteration so that the feature extractors and GCN are optimized jointly. The experiments on a real-world TAVR dataset show the proposed method's superior performance. It improves the prediction accuracy by over 4 % compared to the traditional GCN model. Furthermore, to demonstrate the applicability of our proposed method to other multimodal data analysis tasks, we evaluate the performance of the proposed method on an ocular disease intelligent recognition benchmark. The proposed method achieves competitive prediction performance. [ABSTRACT FROM AUTHOR]

Retinopathy of prematurity (ROP) is a retinal proliferative disorder, and it is the primary cause of childhood blindness. Accurate and convenient automatic diagnostic tools are required to assist ophthalmologists in diagnosing ROP. Existing methods only extract information from fundus image captured from posterior angle, while images captured from other angles are ignored, which limits the performance of the algorithm. In this paper, we propose a graph-based feature fusion network (GFF-Net) that can jointly analyze multiple images and make full use of the relevant information between these images to diagnose the plus disease in ROP. The convolutional features of different fundus images are connected into a graph, where the edges of the graph model the correlation between these images. A graph-based feature fusion module is proposed to aggregate features from the constructed feature graph and produce the final prediction. We compared the proposed GFF-Net with state-of-the-art methods on a clinical dataset and a low-quality "attack dataset". The GFF-Net achieved superior performance compared to other methods on both datasets. The results show that the proposed GFF-Net could be more effective than existing methods in clinical practice. [ABSTRACT FROM AUTHOR]

Purpose: Deep neural networks (DNNs) have made great achievements in computer-aided diagnostic systems, but the success highly depends on massive data with high-quality labels. However, for many medical image datasets, a considerable number of noisy labels are introduced by inter- and intra-observer variability, thus hampering DNNs' performance. To address this problem, a robust noisy label correction method with the co-teaching learning paradigm is proposed. Methods: The proposed method aims to reduce the effect of noisy labels by correcting or removing them. It consists of two modules. An adaptive noise rate estimation module is employed to calculate the dataset's noise rate, which is helpful to detect noisy labels but is usually unavailable in clinical applications. A consistency-based noisy label correction module aims to detect noisy labels and correct them to reduce the disturbance from noisy labels and exploit useful information in data. Results: Experiments are conducted on the public skin lesion dataset ISIC-2017, ISIC-2019, and our constructed thyroid ultrasound image dataset. The results demonstrate that the proposed method outperforms other noisy label learning methods in medical image classification tasks. It is also evaluated on the natural image dataset CIFAR-10 to show its generalization. Conclusion: This paper proposes a noisy label correction method to handle noisy labels in medical image datasets. Experimental results show that it can self-adapt to different datasets and efficiently correct noisy labels, which is suitable for medical image classification. [ABSTRACT FROM AUTHOR]

Vehicular ad hoc networks (VANET) are an emergent technology with a promising future. VANETs are quite different from mobile ad hoc networks (MANETs) in terms of characteristics, challenges, system architecture, and application. One of the significant challenges of the VANET is the security from two essential points of view, the prediction and prevention attackers. In this article, we proposed an operating system that combined both software-defined network (SDN) and self-organizing map (SOM) for the 5G-based VANET system. The proposed system will be a new combination of the SDN- and a self-organizing map (SOM)-based network solution to enhance security in the two dimensions, detecting and preventing attacks. This article analysis the vulnerability of the network performance with taken into account the distributed denial of service (DDoS) attacks. Then, the security of the proposed system has been analyzed and checked with existing of the DDoS. The simulation results presented in this article show that, under general conditions of networks, the proposed system can enhance the networks performance compared to the existing work. [ABSTRACT FROM AUTHOR]

Objectives: To develop and evaluate an artificial intelligence (AI) system that can automatically calculate the glomerular filtration rate (GFR) from dynamic renal imaging without manually delineating the regions of interest (ROIs) of kidneys and the corresponding background. Methods: This study was a single-center retrospective analysis of the data of 14,634 patients who underwent 99mTc-DTPA dynamic renal imaging. Two systems based on convolutional neural networks (CNN) were developed and evaluated: sGFRa predicts the radioactive counts of ROIs and calculates GFR using the Gates equation and sGFRb directly predicts GFR from dynamic renal imaging without using other information. The root-mean-square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and R2 were used to evaluate the performance of our approach. Results: sGFRa achieved an RMSE of 5.05, MAE of 4.03, MAPE of 6.07%, and R2 of 0.93 for total GFR while sGFRb achieved an RMSE of 7.61, MAE of 5.92, MAPE of 8.92%, and R2 of 0.85 for total GFR. The accuracy of sGFRa and sGFRb in determining the stage of chronic kidney disease was 87.41% and 82.44%, respectively. Conclusions: The findings of sGFRa show that automatic GFR calculation based on CNN and using dynamic renal imaging is feasible and efficient and, additionally, can aid clinical diagnosis. Furthermore, the promising results of sGFRb demonstrate that CNN can predict GFR from dynamic renal imaging without additional information. Key Points: • Our CNN-based AI systems can automatically calculate GFR from dynamic renal imaging without manually delineating the ROIs of kidneys and the corresponding background. • sGFRaaccurately predicted the radioactive counts of ROIs and calculated GFR using the Gates method. • sGFRb-predicted GFR directly without any parameters related to the Gates equation. [ABSTRACT FROM AUTHOR]

Background: The ability of endoscopists to identify gastric lesions is uneven. Even experienced endoscopists may miss or misdiagnose lesions due to heavy workload or fatigue or subtle changes in lesions under white-light endoscopy (WLE). This study aimed to develop an artificial intelligence (AI) system that could diagnose six common gastric lesions under WLE and to explore its role in assisting endoscopists in diagnosis. Methods: Images of early gastric cancer, advanced gastric cancer, submucosal tumor, polyp, peptic ulcer, erosion, and lesion-free gastric mucosa were retrospectively collected to train and test the system. The performance of the system was compared with that of 12 endoscopists. The performance of endoscopists with or without referring to the system was also evaluated. Results: A total of 29,809 images from 8947 patients and 1579 images from 496 patients were used to train and test the system, respectively. For per-lesion analysis, the overall accuracy of the system was 85.7%, which was comparable to that of senior endoscopists (85.1%, P = 0.729) and significantly higher than that of junior endoscopists (78.8%, P < 0.001). With system assistance, the overall accuracies of senior and junior endoscopists increased to 89.3% (4.2%, P < 0.001) and 86.2% (7.4%, P < 0.001), respectively. Senior and junior endoscopists achieved varying degrees of improvement in the diagnostic performance of other types of lesions except for polyp. The diagnostic times of senior (3.8 vs 3.2 s per image, P = 0.500) and junior endoscopists (6.2 vs 4.6 s per image, P = 0.144) assisted by the system were both slightly shortened, despite no significant differences. Conclusions: The proposed AI system could be applied as an auxiliary tool to reduce the workload of endoscopists and improve the diagnostic accuracy of gastric lesions. [ABSTRACT FROM AUTHOR]

The vehicular ad hoc network, combined with the distribution of roadside units (RSUs), play a significant role in reducing accidents and traffic jams by sending warning and emergency messages to drivers in real-time. However, communication is not always available owing to the lack of RSUs, reliability of data access and control, to avoid such problems we have proposed a new strategy for the deployment of least possible number of RSUs by developing a support protocol in case of loss of communication that relies on smartphones. The simulation results show that the proposed distribution method clearly reduces the number of RSUs required (compared to existing methods), and contributes towards increased coverage, reduced costs, and less missing data. [ABSTRACT FROM AUTHOR]

Specificity protein (Sp) is a famous family of transcription factors including Sp1, Sp2 and Sp3. Sp1 is the first one of Sp family proteins to be characterized and cloned in mammalian. It has been proposed that Sp1 acts as a modulator of the expression of target gene through interacting with a series of proteins, especially with transcriptional factors, and thereby contributes to the regulation of diverse biological processes. Notably, growing evidence indicates that Sp1 is involved in the main events in the development of atherosclerosis (AS), such as inflammation, lipid metabolism, plaque stability, vascular smooth muscle cells (VSMCs) proliferation and endothelial dysfunction. This review is designed to provide useful clues to further understanding roles of Sp1 in the pathogenesis of AS, and may be helpful for the design of novel efficacious therapeutics agents targeting Sp1. [ABSTRACT FROM AUTHOR]

Face recognition plays the significant role in many human-computer interaction decvices and applications, whose access control systems are based on the verification of face biometrical features. Though great improvement in the recognition performances have been achieved, when under some specific conditions like faces with occlusions, the performance would suffer a severe drop. Occlusion is one of the most significant reasons for the performance degrade of the existing general face recognition systems. The biggest problem in occluded face recognition (OFR) lies in the lack of the occluded face data. To mitigate this problem, this paper has proposed one new OFR network DOMG-OFR (Dynamic Occlusion Mask Generator based Occluded Face Recognition), which keeps trying to generate the most informative occluded face training samples on feature level dynamically, in this way, the recognition model would always be fed with the most valuable training samples so as to save the labor in preparing the synthetic data while simultaneously improving the training efficiency. Besides, this paper also proposes one new module called Decision Module (DM) in an attempt to combine both the merits of the two mainstream methodologies in OFR which are face image reconstruction based methodologies and the face feature filtering based methodologies. Furthermore, to enable the existing face deocclusion methods that mostly target at near frontal faces to work well on faces under large poses, one head pose aware deocclusion pipeline based on the Condition Generative Adversarial Network (CGAN) is proposed. In the experimental parts, we have also investigated the effects of the occlusions upon face recognition performance, and the validity and the efficiency of our proposed Decision based OFR pipeline has been fully proved. Through comparing both the verification and the recognition performance upon both the real occluded face datasets and the synthetic occluded face datasets with other existing works, our proposed OFR architecture has demonstrated obvious advantages over other works. [ABSTRACT FROM AUTHOR]

Ultrasound imaging is one of the most frequently used diagnostic tools for detecting and analyzing abnormalities of the breast. Recently proposed methods for the automated analysis of breast ultrasound images have shown great success, especially in the classification of breast abnormalities into malignant or benign lesions. In this study, we explore the use of a deep convolutional neural network with a multi-scale module for the automated assessment of BI-RADS category on breast ultrasound images. We propose a multi-input region of interest extraction method to extract the breast region from ultrasound images, with this method avoiding the deformation of region of interest images and providing high classification performance. Moreover, we also propose an order-constrained loss function that fully considers the continuity between BI-RADS categories and shows higher performance than traditional loss functions. A large annotated dataset containing 8246 breast ultrasound images was collected to train and evaluate the proposed methods. Ablation experiments were performed to validate the effectiveness of the proposed methods. Experimental results indicate that our method can be used to mimic experienced radiologists in the assessment of the BI-RADS category of breast ultrasound images and that the automated interpretations could be acceptable in routine clinical breast ultrasound examination reports. [ABSTRACT FROM AUTHOR]

Metal artifacts in oral and maxillofacial computed tomography (CT) images affect the imaging quality, interfere with doctors' judgment of anatomical structure and may pose a threat to patients' lives. Although many metal artifact reduction (MAR) methods have been proposed, there is still no complete and ideal solution. We aimed to construct and validate a generative adversarial network (GAN) based model, which was called MAR-GAN, to reduce the artifacts and improve the image quality. We created a large CT images dataset containing 600 pairs of simulated images and 5203 clinical images, where five kinds of metal artifacts were involved. The 600 pairs of simulated images were divided into training and test groups (4:1) and tested using a 5-fold cross-validation schema. Each pair contained an original artifact-free image and a manually simulated artifact-affected image. The inputs were artifact-affected images, and the outputs were artifact-reduced images that were compared with the corresponding artifact-free images. The effectiveness of the proposed MAR-GAN was tested on the simulated dataset and the clinical dataset. The root mean square error and structural similarity values of MAR-GAN were 0.0170 ± 0.0049 and 0.9831 ± 0.0073, respectively, on the simulated dataset. MAR-GAN had significant advantages over two state-of-the-art methods on the clinical dataset, good performance was obtained and the results were all clinically acceptable. Experimental results demonstrated the superior capability of MAR-GAN, which provided improved performance and high reconstruction quality, including preservation of anatomical structures near metal implants and recovery of detailed structural information from low-quality images. [ABSTRACT FROM AUTHOR]

Aiming at the power supply problem of downhole measurement and control instruments, two kinds of vibration energy harvesters, axial vibration energy harvester and lateral vibration energy harvester, are designed in this paper to harvest the axial, lateral, and torsional vibration energy of the drillstring. The mathematical models of the two harvesters are established and the vibration coupling between the harvesters and the drillstring is realized by sharing the node with the drillstring. The performances of the two kinds of harvesters are numerically simulated. The results show that the output power of the axial harvester changes periodically with the change of its installation depth and the number of cycles increases with the increase of the drillstring length. The larger the natural frequency of the axial harvester, the smaller the amplitude of output power fluctuation with depth. When the rotating speed matches the natural frequency of the axial harvester, the maximum output power can reach 1.4 W. The outpower of the lateral VEH is maximum in the possible buckling section near the bit and the maximum output power is about 10 W. The results show that the output power of the designed vibration energy harvester can meet the power demand of most downhole measurement and control instruments. [ABSTRACT FROM AUTHOR]

Objectives: Coronary computed tomography angiography (CCTA) has rapidly developed in the coronary artery disease (CAD) field. However, manual coronary artery tree segmentation and reconstruction are time-consuming and tedious. Deep learning algorithms have been successfully developed for medical image analysis to process extensive data. Thus, we aimed to develop a deep learning tool for automatic coronary artery reconstruction and an automated CAD diagnosis model based on a large, single-centre retrospective CCTA cohort. Methods: Automatic CAD diagnosis consists of two subtasks. One is a segmentation task, which aims to extract the region of interest (ROI) from original images with U-Net. The second task is an identification task, which we implemented using 3DNet. The coronary artery tree images and clinical parameters were input into 3DNet, and the CAD diagnosis result was output. Results: We built a coronary artery segmentation model based on CCTA images with the corresponding labelling. The segmentation model had a mean Dice value of 0.771 ± 0.021. Based on this model, we built an automated diagnosis model (classification model) for CAD. The average accuracy and area under the receiver operating characteristic curve (AUC) were 0.750 ± 0.056 and 0.737, respectively. Conclusion: Herein, using a deep learning algorithm, we realized the rapid classification and diagnosis of CAD from CCTA images in two steps. Our deep learning model can automatically segment the coronary artery quickly and accurately and can deliver a diagnosis of ≥ 50% coronary artery stenosis. Artificial intelligence methods such as deep learning have the potential to elevate the efficiency in CCTA image analysis considerably. Key Points: • The deep learning model rapidly achieved a high Dice value (0.771 ± 0.0210) in the autosegmentation of coronary arteries using CCTA images. • Based on the segmentation model, we built a CAD autoclassifier with the 3DNet algorithm, which achieved a good diagnostic performance (AUC) of 0.737. • The deep neural network could be used in the image postprocessing of coronary computed tomography angiography to achieve a quick and accurate diagnosis of CAD. [ABSTRACT FROM AUTHOR]

In order to overcome the problems of long encryption process, low data security and poor anti-attack rate of traditional tobacco enterprise trade data encryption methods, a trade data encryption of tobacco enterprises based on adversarial neural network is proposed. This method optimizes the traditional neural network by generating countermeasure network, so as to form adversarial neural network. In the adversarial neural network, the encryption processing of tobacco enterprise trade data is completed through data feature classification, design of tobacco enterprise trade data encryption protocol and data encryption channel. The experimental results show that the encryption process of this method takes between 6 s and 20 s, and the encryption efficiency is high, and this method can effectively scramble the original arrangement of the data, so as to effectively hide the effective information in the tobacco enterprise trade data, improve the anti-attack rate of the data, and effectively improve the security of the data. [ABSTRACT FROM AUTHOR]

Faults in train mechanical parts pose a significant safety hazard to railway transportation. Although some image detection methods have replaced manual fault detection of train mechanical parts, the detection effect on small mechanical parts under low illumination conditions is not ideal. To improve the accuracy and efficiency of the detection of train faults under different environments, we propose a multi-mode aggregation feature enhanced network (MAFENet) based on a single-stage detector (SSD). This network uses the idea of a two-step adjustment structure from coarse to fine and uses the K-means algorithm to design anchors. The receptive field enhancement module (RFEM) is designed to obtain the fusion features of different receptive fields. The attention-guided detail feature enhancement module (ADEM) is designed to complement the detailed features of deep-level feature maps. Meanwhile, the complete intersection over union (CIoU) loss is used to obtain more accurate bounding boxes. The experimental results on the train mechanical parts fault (TMPF) dataset showed that the detection performance of MAFENet is better than those of other SSD models. MAFENet with an input size of 320 × 320 pixels can achieve a mean average precision (mAP) of 0.9787 and a detection speed of 33 frames per second (FPS), which indicates that it can realize real-time detection, has good robustness to images under different environmental conditions, and can be used to improve the efficiency of the detection of faulty train parts. [ABSTRACT FROM AUTHOR]

Computed tomography (CT) images of the brain aid the radiotherapy of glioma. The identification and contouring of the gross tumor volume (GTV) are important for radiotherapy. However, manual segmenting GTV is time-consuming, exhausting, and subjective, and automated methods are desired. To overcome these shortcomings, a novel neural network framework based on multi-view fusion is proposed to segment GTV in brain glioma automatically. The multi-view image that includes the previous image, current image, and following image is inputted in this framework to abstract extra spatial features and then aggregated to segment the GTV. Compared with the 2D segmentation framework, the proposed framework retains more spatial information due to the multi-view image. Meanwhile, compared with the 3D segmentation framework, the proposed framework considers fewer images, which means the model has fewer parameters and is easier to train while retaining much useful spatial information. Moreover, the GliomaCT dataset, a large CT dataset collected from West China Hospital, is used to train, validate, and test the proposed method. The performance of the proposed method and other state-of-the-art methods are compared on this dataset. The high dice similarity coefficient achieved in the experiments demonstrates the effectiveness of the proposed method for segmenting the GTV in brain glioma. [ABSTRACT FROM AUTHOR]

Relationship between samples is often ignored when training neural networks for classification tasks. If properly utilized, such information can bring many benefits for the trained models. On the one hand, neural networks trained ignoring similarities between samples may represent different samples closely even if they belong to different classes, which undermines discrimination abilities of the trained models. On the other hand, regularizing inter-class and intra-class similarities in the feature space during training can effectively disentangle the representation between classes and make the representation sparse. To achieve this, a new regularization method is proposed to penalize positive inter-class similarities and negative intra-class similarities in the feature space. Experimental results show that the proposed method can not only obtain sparse and disentangled representation but also improve the performance of the trained models on many datasets. [ABSTRACT FROM AUTHOR]

Polyimide (PI) possesses high heat resistance and low dielectric loss, but exhibits low dielectric constant (k) and energy storage density, which constrains its further application in the field of high-temperature energy storage dielectric. The compounding of high-k filler and PI can greatly improve the dielectric constant of polymer-based dielectric composites, but it is often accompanied by the increase of dielectric loss and deterioration of breakdown strength. This issue can be effectively solved by the fabrication of dielectric filler with core–shell structure and construction of a layered structure. Therefore, in this research, a new SiC@polydopamine (PDA)@Ag nanoparticles (AgNPs)/PI flexible composite film with a sandwich structure (SSP) was prepared by a step-by-step casting method, in which the insulating layer (pristine PI) was intercalated between two polarization layers (SiC@PDA@Ag/PI composites). Pristine PI in the middle layer effectively hinders the transmission of carriers in the middle layer of the composite multilayer film. The SSP shows the highest energy storage density (1.35 J cm−3 under 273.4 kV mm−1), and the tanδ is as low as 0.0057. Additionally, SSP also shows excellent thermal stability and moisture resistance. [ABSTRACT FROM AUTHOR]

The brief investigates the coexistence of multiple continuous attractors in a recurrent neural network, i.e., the symmetric saturated Satlins linear neural networks, based on a parameterized 2-D model. The saturated parts of the unliner activation function are the breakthrough for us to study the coexistence. The novel point of our research method is linearization in nonlinear networks. A continuous attractor is a set of connected stable equilibrium points. On the basis of the theorem that we proved on stability and existing findings on equilibria in mathematics, we propose the conditions for the coexistence of two or even multiple continuous attractors in a recurrent neural network. Simulations are also demonstrated to illustrate the theoretical results. [ABSTRACT FROM AUTHOR]

Purpose: Whole-body bone scintigraphy (WBS) is one of the common imaging methods in nuclear medicine. It is a time-consuming, tedious, and error-prone issue for physicians to determine the location of bone lesions which is important for the qualitative diagnosis of bone lesions. In this paper, an automatic fine-grained skeleton segmentation method for WBS is developed. Method: The proposed method contains four steps. In the first step, a novel denoising method is proposed to remove the noise from WBS which benefits the location of the skeleton. In the second step, a restoration method based on gray probability distribution is developed to repair the partial contamination caused by the high local density of radionuclide. Then, the standardization for WBS is performed by the exact histogram matching. Finally, the deformation field between the atlas and the input WBS is calculated by registration, and the segmentation mask of the input WBS is obtained by wrapping the segmentation mask of the atlas with the deformation field. Results: The experimental results show that the proposed method outperforms the traditional registration (Morphon): mean square error decreased from 11.14 × 10 - 3 to 2.10 × 10 - 3 , peak signal-to-noise ratio increased from 21.26 to 26.92, and mean structural similarity increased from 0.9986 to 0.9998. Conclusions: Our experiments show that the proposed method can achieve robust and fine-grained results which outperform the traditional registration method, indicating it could be helpful in clinical application. To the best of our knowledge, this is the first work that implements a fully automated fine-grained skeleton segmentation method for WBS. [ABSTRACT FROM AUTHOR]

The detection and characterization of lymph nodes through interpreting abdominal medical images are significant for diagnosing and treating colorectal cancer recurrence. However, interpreting abdominal medical images manually is labor-intensive and time-consuming. The related radiology education has many limitations as well. In this context, we seek to build an extensive collection of abdominal medical images with ground truth labels for lymph nodes recognition research and help junior doctors to train their interpretation skills. Therefore, we develop TeachMe, which is a web-based teaching system for annotating abdominal lymph nodes. The system has a three-level annotation-review workflow to construct an expert database of abdominal lymph nodes and a feedback mechanism helping junior doctors to learn the tricks of interpreting abdominal medical images. TeachMe's functionalities make itself stand out against other platforms. To validate these functionalities, we invite a medical team from Gastrointestinal Surgery Center, West China Hospital, to participate in the data collection workflow and experience the feedback mechanism. With the help of TeachMe, an expert dataset of abdominal lymph nodes has been created and an automated detection model for abdominal lymph nodes with incredible performances has been proposed. Moreover, through three rounds of practicing via TeachMe, our junior doctors' interpretation skills have been improved. [ABSTRACT FROM AUTHOR]

The article titled "Design and development of a hybrid (SDN + SOM) approach for enhancing security in VANET" has been retracted by the publisher, Applied Nanoscience. The retraction was made due to concerns regarding compromised editorial handling and peer review process, inappropriate or irrelevant references, and the article not being in scope of the journal or guest-edited issue. The authors have not responded to correspondence regarding this retraction. Springer Nature, the publisher, remains neutral in terms of jurisdictional claims and institutional affiliations. [Extracted from the article]

At 6pm on July 19, 2019, the Yahuokou earthflow (Zhouqu County, Gansu Province, China) was reactivated following a period of intense rainfall. A volume of approximately 3.9 × 106 m3 travelled approximately 1100 m in 27 days, blocking part of the Min River. Multi-temporal InSAR monitoring and interpretation of high-resolution optical images were used to map different phase of mass movements (including rock fall, topple, slide, and flowslide) to characterize the pre-failure deformation of the earthflow. Analysis of displacement patterns indicates that initiation of the earthflow is closely correlated with road construction, intense and long-lasting rainfall (443 mm from April 1, to July 19, 2019), and seismic activity. The potential to forecast the time of failure based on pre-failure deformation patterns observed by InSAR monitoring is demonstrated. Three types of pre-failure deformation are distinguished: steady creep, fluctuating creep, and accelerating creep. The monitoring process enabled identification and characterization of three key stages of the earthflow (initiation, transportation, and accumulation/deposition) providing important lessons learned for improvement of landslide early warning and hazard assessment in similar geological and geomorphological settings. [ABSTRACT FROM AUTHOR]

Poor eye health is a major public health problem, and the timely detection and diagnosis of fundus abnormalities is important for eye health protection. Traditional imaging models can hinder the comprehensive evaluation of fundus abnormalities due to the use of a narrow field of view. The emerging ultrawide-field (UWF) imaging model surpasses this limitation in a non invasive, wide-view manner and is suitable for fundus observation and screening. Nevertheless, manual screening is labour intensive and subjective, especially in the absence of an ophthalmologist. Therefore, a set of auxiliary screening methods for a fundus screening service using a combination of deep learning and UWF imaging technology, which is designated as DeepUWF-Plus, is proposed. This service includes a subsystem for the screening of fundus, a subsystem for the identification of abnormalities regarding four important fundus locations, and a subsystem for the diagnosis of four retinal diseases that threaten vision. The influence of two-stage and one-stage classification strategies on the prediction performance of the model is experimentally investigated to alleviate severe class imbalance and similarity between classes, and evaluate the effectiveness and reliability of the system. Our experimental results show that DeepUWF-Plus is effective when using the two-stage strategy, especially for identifying signs or symptoms of minor diseases. DeepUWF-Plus can improve the practicality of fundus screening and enable ophthalmologists to provide more comprehensive fundus assessments. [ABSTRACT FROM AUTHOR]

Purpose: The robust and automatic segmentation of the pulmonary lobe is vital to surgical planning and regional image analysis of pulmonary related diseases in real-time Computer Aided Diagnosis systems. While a number of studies have examined this issue, the segmentation of unclear borders of the five lobes of the lung remains challenging because of incomplete fissures, the diversity of anatomical pulmonary information, and obstructive lesions caused by pulmonary diseases. This study proposes a model called Regularized Pulmonary Lobe Segmentation Network to accurately predict the lobes as well as the borders. Methods: First, a 3D fully convolutional network is constructed to extract contextual features from computed tomography images. Second, multi-task learning is employed to learn the segmentations of the lobes and the borders between them to train the neural network to better predict the borders via shared representation. Third, a 3D depth-wise separable de-convolution block is proposed for deep supervision to efficiently train the network. We also propose a hybrid loss function by combining cross-entropy loss with focal loss using adaptive parameters to focus on the tissues and the borders of the lobes. Results: Experiments are conducted on a dataset annotated by experienced clinical radiologists. A 4-fold cross-validation result demonstrates that the proposed approach can achieve a mean dice coefficient of 0.9421 and average symmetric surface distance of 1.3546 mm, which is comparable to state of the art methods. The proposed approach has the capability to accurately segment voxels that are near the lung wall and fissure. Conclusion: In this paper, a 3D fully convolutional networks framework is proposed to segment pulmonary lobes in chest CT images accurately. Experimental results show the effectiveness of the proposed approach in segmenting the tissues as well as the borders of the lobes. [ABSTRACT FROM AUTHOR]

Purpose: Airway tree segmentation plays a pivotal role in chest computed tomography (CT) analysis tasks such as lesion localization, surgical planning, and intra-operative guidance. The remaining challenge is to identify small bronchi correctly, which facilitates further segmentation of the pulmonary anatomies. Methods: A three-dimensional (3D) multi-scale feature aggregation network (MFA-Net) is proposed against the scale difference of substructures in airway tree segmentation. In this model, the multi-scale feature aggregation (MFA) block is used to capture the multi-scale context information, which improves the sensitivity of the small bronchi segmentation and addresses the local discontinuities. Meanwhile, the concept of airway tree partition is introduced to evaluate the segmentation performance at a more granular level. Results: Experiments were conducted on a dataset of 250 CT scans, which were annotated by experienced clinical radiologists. Through the airway partition, we evaluated the segmentation results of the small bronchi compared with the state-of-the-art methods. Experiments show that MFA-Net achieves the best performance in the Dice similarity coefficient (DSC) in the intra-lobar airway and improves the true positive rate (TPR) by 7.59% on average. Besides, in the entire airway, the proposed method achieves the best results in DSC and TPR scores of 86.18% and 79.31%, respectively, with the consequence of higher false positives. Conclusion: The MFA-Net is competitive with the state-of-the-art methods. The experiment results indicate that the MFA block improves the performance of the network by utilizing multi-scale context information. More accurate segmentation results will be more helpful in further clinical analysis. [ABSTRACT FROM AUTHOR]

The scientific and reasonable layout of jack-up drilling platform can improve the efficiency and safety of drilling platform operation. Through the application of analytic hierarchy process (AHP) and the establishment of comprehensive evaluation and analysis model, this paper analyzes the commonly used drilling platform layout scheme and optimizes the scheme to meet operational requirements and to improve the ability of the drilling platform operation. [ABSTRACT FROM AUTHOR]

The aim of automated stem cell motility analysis is reliable processing and evaluation of cell behaviors such as translocation, mitosis, death, and so on. Cell tracking plays an important role in this research. In practice, tracking stem cells is difficult because they have frequent motion, deformation activities, and small resolution sizes in microscopy images. Previous tracking approaches designed to address this problem have been unable to generalize the rapid morphological deformation of cells in a complex living environment, especially for real-time tracking tasks. Herein, a deep learning framework with convolutional structure and multi-output layers is proposed for overcoming stem cell tracking problems. A convolutional structure is used to learn robust cell features through deep features learned on massive visual data by a transfer learning strategy. With multi-output layers, this framework tracks the cell's motion and simultaneously detects its mitosis as an assistant task. This improves the generalization ability of the model and facilitates practical applications for stem cell research. The proposed framework, tracking and detection neural networks, also contains a particle filter-based motion model, a specialized cell sampling strategy, and corresponding model update strategy. Its current application to a microscopy image dataset of human stem cells demonstrates increased tracking performance and robustness compared with other frequently used methods. Moreover, mitosis detection performance was verified against manually labeled mitotic events of the tracked cell. Experimental results demonstrate good performance of the proposed framework for addressing problems associated with stem cell tracking. [ABSTRACT FROM AUTHOR]

Lung cancer is one of the most common and fatal types of cancer, and pulmonary nodule detection plays a crucial role in the screening and diagnosis of this disease. A well-trained deep neural network model can help doctors to find nodules on computed tomography(CT) images while requiring lots of labeled data. However, currently available annotating systems are not suitable for annotating pulmonary nodules in CT images. We propose a web-based lung nodules annotating system named as DeepLNAnno. DeepLNAnno has a unique three-tier working process and loads of features like semi-automatic annotation, which not only make it much easier for doctors to annotate compared to some other annotating systems but also increase the accuracy of the labels. We invited a medical group from West China Hospital to annotate the CT images using our DeepLNAnno system, and collected a large number of labeled data. The results of our experiments demonstrated that a usable nodule-detection system is developed, and good benchmark scores on our evaluation data are obtained. [ABSTRACT FROM AUTHOR]

FeGa thin films have been deposited onto (100)-oriented GaAs and (001)-oriented Si substrates at different temperatures by pulsed laser deposition. It has been shown that at different substrate temperatures, magnetic properties of FeGa films greatly change, which is mainly dependent on different crystal textures. The presence of the structure consisting of D03-ordered grain phase and bcc α-Fe crystal structure (A2) has been revealed using X-ray diffraction analysis. The diffraction peak of FeGa (001) direction, which ‎corresponds to the D03-ordered granular phase, was observed when the temperatures of the GaAs substrate were 400 and 600°C. The diffraction peak that corresponds to (110) direction of the bcc A2 structure has appeared at 800°C. However, the FeGa film on the Si substrate did not demonstrate any obvious structural change. The changes in the magnetic properties can be mainly attributed to changes in the roughness and the lattice mismatch between FeGa and Si.‎ [ABSTRACT FROM AUTHOR]

In the film industry, the ability to predict a movie's box-office revenues before its theatrical release can decrease its financial risk. However, accurate predictions are not easily obtained. The complex relationship between movie-related data and movie box-office revenues, plus the increasing volume of data in online movie databases, pose challenges for their effective analysis. In this paper, a multimodal deep neural network, incorporating input about movie poster features learned in a data-driven fashion, is proposed for movie box-office revenues prediction. A convolutional neural network (CNN) is built to extract features from movie posters. By pre-training the CNN, features that are relevant to movie box-office revenues can be learned. To evaluate the performance of the proposed multimodal deep neural network, comparative studies with other prediction techniques were carried out on an Internet Movie Database dataset, and visualization of movie poster features was also performed. Experimental results demonstrate the superiority of the proposed multimodal deep neural network for movie box-office revenues prediction. [ABSTRACT FROM AUTHOR]

Audio classification is regarded as a great challenge in pattern recognition. Although audio classification tasks are always treated as independent tasks, tasks are essentially related to each other such as speakers' accent and speakers' identification. In this paper, we propose a Deep Neural Network (DNN)-based multi-task model that exploits such relationships and deals with multiple audio classification tasks simultaneously. We term our model as the gated Residual Networks (GResNets) model since it integrates Deep Residual Networks (ResNets) with a gate mechanism, which extract better representations between tasks compared with Convolutional Neural Networks (CNNs). Specifically, two multiplied convolutional layers are used to replace two feed-forward convolution layers in the ResNets. We tested our model on multiple audio classification tasks and found that our multi-task model achieves higher accuracy than task-specific models which train the models separately. [ABSTRACT FROM AUTHOR]

Considering the computational complexity and redundancy of traditional array signal arrival angle (DOA) estimation algorithms, the compressed sensing technology was used to improve the real-time and accurate performance of the DOA estimation algorithm, in which, the space sparse signals were reconstructed from the array data by means of array manifold matrix. Compared with the classical MUSIC algorithm, the compressed sensing DOA estimation method could effectively improve the direction finding accuracy and angle resolution with low SNR and snapshot deficiency. Moreover, the proposed algorithm could achieve the coherent signal estimation correctly, and the simulation results show that its performance was superior to that of traditional algorithm. [ABSTRACT FROM AUTHOR]

Carbon dioxide (CO2) enrichment is used to boost crop yield in greenhouse vegetable production that often exposes vegetables to the simultaneous occurrence of elevated CO2 and salinity due to frequent irrigation and fertilization in facility horticulture. The beneficial effects of exogenous spermidine (Spd, a kind of polyamine) on plant growth and development under salt stress have been widely reported; however, little information is available on the effects of Spd on the combined treatment of CO2 enrichment and iso-osmotic salt stress. In this work, the effects of exogenous Spd (0.25 mM) on plant growth, chlorophyll content, water status, osmotic adjustment, and the antioxidant system were investigated under CO2 enrichment (800 ppm) and iso-osmotic salt stress [150 mmol/L NaCl and 100 mmol/L Ca(NO3)2] in tomato (Solanum lycopersicum L.). The results showed that iso-osmotic salt stress significantly decreased fresh and dry weights of plants, relative water content, total root length, total root surface area, total root volume, and average root diameter in tomato plants. However, either elevated CO2 or exogenous Spd both attenuated iso-osmotic salt stress-induced reductions in growth parameters and the combined treatment of elevated CO2 and Spd showed a more profound effect, leading to the enhanced tolerance to salt stress in tomato plants. Elevated CO2 and/or exogenous Spd-induced alleviation of iso-osmotic salt stress was associated with efficient osmotic adjustment and antioxidant defense, which minimized the salt stress-induced oxidative stress as well. Therefore, Spd application can be advocated to mitigate secondary salinization in protected vegetable production where elevated CO2 is simultaneously used to boost crop yield. [ABSTRACT FROM AUTHOR]

Object tracking still remains challenging in computer vision because of the severe object variation, e.g., deformation, occlusion, and rotation. To handle the object variation and achieve robust object tracking performance, we propose a novel relationship-based tracking algorithm using neural networks in this paper. Compared with existing approaches in the literature, our method assumes the targeted object to be consisted of several parts and considers the evolution of the topology structure among these parts. After training a candidate neural network for predicting the probable areas each part may locate at in the successive frame, we then design a novel collaboration neural network to determine the precise area each part will locate at with account for the topology structure among these individual parts, which is learned from their historical physical locations during online tracking process. Experimental results show that the proposed method outperforms state-of-the-art trackers on a benchmark dataset, yielding the significant improvements in accuracy on high-distorted sequences. [ABSTRACT FROM AUTHOR]

Integrating together two dissimilar π-conjugated molecules into controlled complex topological configurations remains a largely unsolved problem owing to the diversity of organic species and their respective different assembly features. Here, we find that two structurally similar organic semiconductors, 9,10-bis(phenylethynyl)anthracene (BA) and 5,12-bis(phenylethynyl)naphthacene (BN), co-assemble into two-component helices by control of the growth kinetics as well as the molar ratio of BA/BN. The helical superstructures made of planar and twisted bis(phenylethynyl) derivatives can be regarded as (BA)x(BN)1-x alloys, which are formed due to compatible structural relationship between BA and BN. Moreover, epitaxial growth of (BA)x(BN)1-x alloy layer on the surface of BA tube to form BA@(BA)x(BN)1-x core-shell structure is also achieved via a solute exchange process. The precise control over composition and morphology towards organic alloy helices and core-shell microstructures opens a door for understanding the complex co-assembly features of two or more different material partners with similar structures. [ABSTRACT FROM AUTHOR]

The distribution of SiC particles in Al-SiC composite can greatly influence the mechanical performances of Al-SiC composite. To realize the homogeneous distribution of SiC particles in stir-casting Al-SiC composite, semisolid stir casting of Al-4.25vol.%SiC composite was conducted using a special electromagneticmechanical stirring equipment made by our team, in which there are three uniformly-distributed blades with a horizontal tilt angle of 25 ° to mechanically raise the SiC particles by creating an upward movement of slurry under electromagnetic stirring. The microstructure of the as-cast Al-SiC composites was observed by Scanning Electron Mcroscopy (SEM). The volume fraction of SiC particles was measured by image analysis using the Quantimet 520 Image Processing and Analysis System. The tensile strength of the Al-4.25vol.%SiC composites was measured by tensile testing. Results show that the Al-4.25vol.%SiC composites with the homogeneous distrbutin of SiC particles can be obtained by the electromagnetic-mechanical stirring casting with the speed of 300 and 600 r·min-1 at 620 °C. The differences between the volume fraction of SiC particles at the top of ingot and that at the bottom are both ~0.04vol.% with the stirring speed of 300 and 600 r·min-1, which are so small that the distribution of SiC particles can be seen as the homogeneous. The tensile strength of the Al matrix is enhanced by 51.2% due to the uniformly distributed SiC particles. The porosity of the composite mainly results from the solidification shrinkage of slurry and it is less than 0.04vol.%. [ABSTRACT FROM AUTHOR]

Oxidized phospholipids (oxPAPC) induce endothelial dysfunction and atherosclerosis. Here we show that oxPAPC induce a gene network regulating serine-glycine metabolism with the mitochondrial methylenetetrahydrofolate dehydrogenase/cyclohydrolase (MTHFD2) as a causal regulator using integrative network modeling and Bayesian network analysis in human aortic endothelial cells. The cluster is activated in human plaque material and by atherogenic lipoproteins isolated from plasma of patients with coronary artery disease (CAD). Single nucleotide polymorphisms (SNPs) within the MTHFD2-controlled cluster associate with CAD. The MTHFD2-controlled cluster redirects metabolism to glycine synthesis to replenish purine nucleotides. Since endothelial cells secrete purines in response to oxPAPC, the MTHFD2-controlled response maintains endothelial ATP. Accordingly, MTHFD2-dependent glycine synthesis is a prerequisite for angiogenesis. Thus, we propose that endothelial cells undergo MTHFD2-mediated reprogramming toward serine-glycine and mitochondrial one-carbon metabolism to compensate for the loss of ATP in response to oxPAPC during atherosclerosis. [ABSTRACT FROM AUTHOR]

Weyl fermions are a recently discovered ingredient for correlated states of electronic matter. A key difficulty has been that real materials also contain non-Weyl quasiparticles, and disentangling the experimental signatures has proven challenging. Here we use magnetic fields up to 95 T to drive the Weyl semimetal TaAs far into its quantum limit, where only the purely chiral 0th Landau levels of the Weyl fermions are occupied. We find the electrical resistivity to be nearly independent of magnetic field up to 50 T: unusual for conventional metals but consistent with the chiral anomaly for Weyl fermions. Above 50 T we observe a two-orderof- magnitude increase in resistivity, indicating that a gap opens in the chiral Landau levels. Above 80 T we observe strong ultrasonic attenuation below 2 K, suggesting a mesoscopically textured state of matter. These results point the way to inducing new correlated states of matter in the quantum limit of Weyl semimetals. [ABSTRACT FROM AUTHOR]