Lightweight Cross-Modal Multispectral Pedestrian Detection Based on Spatial Reweighted Attention Mechanism

Multispectral pedestrian detection technology leverages infrared images to provide reliable information for visible light images, demonstrating significant advantages in low-light conditions and background occlusion scenarios. However, while continuously improving cross-modal feature extraction and fusion, ensuring the model’s detection speed is also a challenging issue. We have devised a deep learning network model for cross-modal pedestrian detection based on Resnet50, aiming to focus on more reliable features and enhance the model’s detection efficiency. This model employs a spatial attention mechanism to reweight the input visible light and infrared image data, enhancing the model’s focus on different spatial positions and sharing the weighted feature data across different modalities, thereby reducing the interference of multi-modal features. Subsequently, lightweight modules with depthwise separable convolution are incorporated to reduce the model’s parameter count and computational load through channel-wise and point-wise convolutions. The network model algorithm proposed in this paper was experimentally validated on the publicly available KAIST dataset and compared with other existing methods. The experimental results demonstrate that our approach achieves favorable performance in various complex environments, affirming the effectiveness of the multispectral pedestrian detection technology proposed in this paper.

A Review of Deep Learning-Based Vulnerability Detection Tools for Ethernet Smart Contracts

In recent years,the number of smart contracts deployed on blockchain has exploded.However,the issue of vulnerability has caused incalculable losses.Due to the irreversible and immutability of smart contracts,vulnerability detection has become particularly important.With the popular use of neural network model,there has been a growing utilization of deep learning-based methods and tools for the identification of vulnerabilities within smart contracts.This paper commences by providing a succinct overview of prevalent categories of vulnerabilities found in smart contracts.Subsequently,it categorizes and presents an overview of contemporary deep learning-based tools developed for smart contract detection.These tools are categorized based on their open-source status,the data format and the type of feature extraction they employ.Then we conduct a comprehensive comparative analysis of these tools,selecting representative tools for experimental validation and comparing them with traditional tools in terms of detection coverage and accuracy.Finally,Based on the insights gained from the experimental results and the current state of research in the field of smart contract vulnerability detection tools,we suppose to provide a reference standard for developers of contract vulnerability detection tools.Meanwhile,forward-looking research directions are also proposed for deep learning-based smart contract vulnerability detection.

Effects of Al_(2)O_(3)-SiO_(2) Raw Material Types on Properties of Anorthite Based Insulation Refractories

To optimize their Al_(2)O_(3)-SiO_(2) raw materials,anorthite based insulation refractories were prepared by the in-situ sintering process combined with the foaming method after sintering at 1350℃for 3 h,using green and pollution-free kaolin,kyanite,andalusite and sillimanite as Al_(2)O_(3)-SiO_(2) raw materials,respectively,and industrial CaCO_(3) as the CaO source.Effects of Al_(2)O_(3)-SiO_(2) raw material types on the physical properties,phase composition and microstructure were investigated.The results are as follows.All samples prepared by different Al_(2)O_(3)-SiO_(2) raw materials have hexagonal flake anorthite and a small amount of mullite and corundum.Their bulk density and thermal conductivity decrease in the order of using kaolin,andalusite,kyanite and sillimanite as the Al_(2)O_(3)-SiO_(2) raw material,but their apparent porosity increases.Moreover,in the sample with kaolin,the bonding between anorthite crystals on the pore walls is closer than that of the other samples,which is conducive to increasing the cold crushing strength.The bonding between anorthite crystals on pore walls gradually decreases in the order of using kyanite,andalusite and sillimanite as the Al_(2)O_(3)-SiO_(2) raw material,thus their cold crushing strength decreases accordingly.In comprehensive consideration,the properties of the sample from kyanite are the optimal.Its apparent porosity,thermal conductivity and cold crushing strength are 84.6%,0.141 W·m^(-1)·K^(-1) and 1.89 MPa,respectively.

Review on mechanisms and structure-activity relationship of hypoglycemic effects of polysaccharides from natural resources

Diabetes mellitus(DM)is a common multifactorial disease,causing various complications,such as chronic metabolism.The current therapies for diabetes mellitus are commercial diabetic drugs that have different definite side effect.However,polysaccharides mainly extracted from natural resources,have advantages of safety,accessibility,and anti-diabetic potential.We have summarized recent research of natural polysaccharides with hypoglycemic activities,focusing on different pharmacological mechanisms in various cell and animal models.The relationships of structure-hypoglycemic effect are also discussed in detail.This review could provide a comprehensive perspective for better understanding on development and mechanism of natural polysaccharides against diabetes mellitus,which have been required by clinical studies yet.

Elucidation of the structure, antioxidant, and interfacial properties of flaxseed proteins tailored by microwave treatment

The microwave treatment is commonly applied to flaxseed to release nutrients, inactivate enzymes, remove cyanogens,and intensify flavors. The current study aimed to explore the influences of microwave exposure on the antioxidant and interfacial properties of flaxseed protein isolates(FPI), focusing on the altering composition and molecular structure.The results showed that after microwave exposure(700 W, 1–5 min), more compact assembly of storage proteins and subsequent permeation by membrane fragments of oil bodies occurred for cold-pressing flaxseed flours. Moreover, the particle sizes of FPI was progressively reduced with the decrement ranged from 37.84 to 60.66%(P<0.05), whereas the zeta potential values initially decreased and then substantially recovered during 1–5 min of microwave exposure. The conformation unfolding, chain cross-linking, and depolymerization were sequentially induced for FPI based on the analysis of fluorescence emission spectra, secondary structure, and protein subunit profiles, thereby affecting the dispersion or aggregation properties between albumin and globulin fractions in FPI. Microwave exposure retained specific phenolic acids and superior in vitro antioxidant activities of FPI. The inferior gas–water interface absorption and the loose/porous assembly structure were observed for the foams prepared by FPI, concurrent with obviously shrinking foaming properties upon microwave exposure. Improving oil–water interface activities of FPI produced the emulsion droplets with descending sizes and dense interface coating, which were then mildly destabilized due to the lipid leakage and weakened rheological behavior with microwave exposure extended to 5 min. Our findings elucidated that microwave treatment could tailor the application functionality of protein fractions in flaxseed based on their in situ structural remodeling.

In-situ incorporation of halloysite nanotubes with 2D zeolitic imidazolate framework-L based membrane for dye/salt separation

Layered assembled membranes of 2D leaf-like zeolitic imidazolate frameworks(ZIF-L)nanosheets have received great attention in the field of water treatment due to the porous structure and excellent antibacterial ability,but the dense accumulation on the membrane surface and the low permeate flux greatly hinder their application.Herein,we synthesized m HNTs(modified halloysite nanotubes)/ZIF-L nanocomposites on modified m HNTs by in situ growth method.Interestingly,due to the different size of m HNTs and ZIF-L,m HNTs were packed in ZIF-L nanosheets.The hollow m HNTs provided additional transport channels for water molecules,and the accumulation of the ZIF-L nanosheets was decreased after assembling m HNTs/ZIF-L nanocomposites into membrane by filtration.The prepared m HNTs/ZIF-L membrane presented high permeate flux(59.6 L·m^(-2)·h^(-1)),which is 2-4 times of the ZIF-L membranes(14.8 L·m^(-2)·h^(-1)).Moreover,m HNTs/ZIF-L membranes are intrinsically antimicrobial,which exhibit extremely high bacterial resistance.We provide a controllable strategy to improve 2D ZIF-L assembles,and develops novel membranes using 2D package structure as building units.

Breather and its interaction with rogue wave of the coupled modified nonlinear Schrodinger equation

We investigate the coupled modified nonlinear Schr?dinger equation.Breather solutions are constructed through the traditional Darboux transformation with nonzero plane-wave solutions.To obtain the higher-order localized wave solution,the N-fold generalized Darboux transformation is given.Under the condition that the characteristic equation admits a double-root,we present the expression of the first-order interactional solution.Then we graphically analyze the dynamics of the breather and rogue wave.Due to the simultaneous existence of nonlinear and self-steepening terms in the equation,different profiles in two components for the breathers are presented.

A Hybrid Heuristic Service Caching and Task Offloading Method for Mobile Edge Computing

Computing-intensive and latency-sensitive user requests pose significant challenges to traditional cloud computing.In response to these challenges,mobile edge computing(MEC)has emerged as a new paradigm that extends the computational,caching,and communication capabilities of cloud computing.By caching certain services on edge nodes,computational support can be provided for requests that are offloaded to the edges.However,previous studies on task offloading have generally not considered the impact of caching mechanisms and the cache space occupied by services.This oversight can lead to problems,such as high delays in task executions and invalidation of offloading decisions.To optimize task response time and ensure the availability of task offloading decisions,we investigate a task offloading method that considers caching mechanism.First,we incorporate the cache information of MEC into the model of task offloading and reduce the task offloading problem as a mixed integer nonlinear programming(MINLP)problem.Then,we propose an integer particle swarm optimization and improved genetic algorithm(IPSO_IGA)to solve the MINLP.IPSO_IGA exploits the evolutionary framework of particle swarm optimization.And it uses a crossover operator to update the positions of particles and an improved mutation operator to maintain the diversity of particles.Finally,extensive simulation experiments are conducted to evaluate the performance of the proposed algorithm.The experimental results demonstrate that IPSO_IGA can save 20%to 82%of the task completion time,compared with state-of-theart and classical algorithms.Moreover,IPSO_IGA is suitable for scenarios with complex network structures and computing-intensive tasks.

Photoluminescence Enhancement of Aluminum Ion Intercalated MoS_(2)Quantum Dots

Low photolumines-cence(PL)quantum yield of molybdenum disulfide(MoS_(2))quan-tum dots(QDs)has lim-ited practical applica-tion as potential fluores-cent materials.Here,we report the intercalation of aluminum ion(Al^(3+))to enhance the PL of MoS_(2)QDs and the un-derlying mechanism.With detailed characterization and exciton dynamics study,we suggest that additional surface states including new emission centers have been effectively introduced to MoS_(2)QDs by the Al^(3+)intercalation.The synergy of new radiative pathway for exciton re-combination and the passivation of non-radiative surface traps is responsible for the en-hanced fluorescence of MoS_(2)QDs.Our findings demonstrate an efficient strategy to improve the optical properties of MoS_(2)QDs and are important for understanding the regulation effect of surface states on the emission of two dimensional sulfide QDs.

A PSO Improved with Imbalanced Mutation and Task Rescheduling for Task Offloading in End-Edge-Cloud Computing

To serve various tasks requested by various end devices with different requirements,end-edge-cloud(E2C)has attracted more and more attention from specialists in both academia and industry,by combining both benefits of edge and cloud computing.But nowadays,E2C still suffers from low service quality and resource efficiency,due to the geographical distribution of edge resources and the high dynamic of network topology and user mobility.To address these issues,this paper focuses on task offloading,which makes decisions that which resources are allocated to tasks for their processing.This paper first formulates the problem into binary non-linear programming and then proposes a particle swarm optimization(PSO)-based algorithm to solve the problem.The proposed algorithm exploits an imbalance mutation operator and a task rescheduling approach to improve the performance of PSO.The proposed algorithm concerns the resource heterogeneity by correlating the probability that a computing node is decided to process a task with its capacity,by the imbalance mutation.The task rescheduling approach improves the acceptance ratio for a task offloading solution,by reassigning rejected tasks to computing nodes with available resources.Extensive simulated experiments are conducted.And the results show that the proposed offloading algorithm has an 8.93%–37.0%higher acceptance ratio than ten of the classical and up-to-date algorithms,and verify the effectiveness of the imbalanced mutation and the task rescheduling.

Numerical Simulation Analysis of the Transformer Fire Extinguishing Process with a High-Pressure Water Mist System under Different Conditions

To thoroughly study the extinguishing effect of a high-pressure water mist fire extinguishing system when a transformer fire occurs,a 3D experimental model of a transformer is established in this work by employing Fire Dynamics Simulator(FDS)software.More specifically,by setting different parameters,the process of the highpressure water mist fire extinguishing system with the presence of both diverse ambient temperatures and water mist sprinkler laying conditions is simulated.In addition,the fire extinguishing effect of the employed high-pressure water mist system with the implementation of different strategies is systematically analyzed.The extracted results show that a fire source farther away fromthe centerline leads to a lower local temperature distribution.In addition,as the ambient temperature increases,the temperature above the fire source decreases,while the temperature and the concentrationof theupperflue gas layer bothdecrease.Interestingly,after thehigh-pressurewatermist sprinkler begins to operate,both the temperature distribution above the fire source and the concentration of the flue gas decrease,which indicates that the high-pressure water mist system plays the role of cooling and dust removal.By comparing various sprinkler laying methods,it is found that the lower sprinkler height has a better effect on the temperature above the fire source,the temperature of the upper flue gas layer,and the concentration of the flue gas.Moreover,when the sprinkler is spread over thewhole transformer,the cooling effect on both the temperature above the fire source and the temperature of the upper flue gas layer is good,whereas the change in the concentration of the flue gas above the fire source is not obvious compared to the case where the sprinkler is not fully spread.

Efficient Certificateless Authenticated Key Agreement for Blockchain-Enabled Internet of Medical Things

Internet of Medical Things(IoMT)plays an essential role in collecting and managing personal medical data.In recent years,blockchain technology has put power in traditional IoMT systems for data sharing between different medical institutions and improved the utilization of medical data.However,some problems in the information transfer process between wireless medical devices and mobile medical apps,such as information leakage and privacy disclosure.This paper first designs a cross-device key agreement model for blockchain-enabled IoMT.This model can establish a key agreement mechanism for secure medical data sharing.Meanwhile,a certificateless authenticated key agreement(KA)protocol has been proposed to strengthen the information transfer security in the cross-device key agreement model.The proposed KA protocol only requires one exchange of messages between the two parties,which can improve the protocol execution efficiency.Then,any unauthorized tampering of the transmitted signed message sent by the sender can be detected by the receiver,so this can guarantee the success of the establishment of a session key between the strange entities.The blockchain ledger can ensure that the medical data cannot be tampered with,and the certificateless mechanism can weaken the key escrow problem.Moreover,the security proof and performance analysis are given,which show that the proposed model and KA protocol are more secure and efficient than other schemes in similar literature.

Physical and emulsifying properties of pea protein:influence of combined physical modification by flaxseed gum and ultrasonic treatment

This study characterized and compared the physical and emulsifying properties of pea protein(PP)and its modified proteins(ultrasound treated-PP(PPU),flaxseed gum(FG)treated PP(PPFG)and ultrasound treated-PPFG(PPFGU)).The results showed FG triggered the formation of loosely attached complex with PP via physical modification under gentle magnetic stirring at pH 7.0,while ultrasound played an important role in reducing protein size,increasing surface hydrophobicity and molecular fluidity onto oil-water interface.So ultrasound further enhanced the interaction of PP with FG,and produced the PPFGU complex with smaller droplet size,higherζ-potential and lower turbidity.Further,combination of FG and ultrasound improved the physical properties of PP with higher viscosity,stiffer gels(defined as higher elastic modulus),stronger hydrophobic properties,better thermal stability,and fast protein absorption rate.Therefore,the PPFGU coarse emulsion performed highest emulsifying activity index(EAI)and emulsion stability index(ESI)that the stabilized nanoemulsion obtained smallest droplet size,higherζ-potential,and longest storage stability.The combination of FG and ultrasonic treatment will be an effective approach to improving the emulsifying property and thermal stability of PP,which can be considered as a potential plant-based emulsifier applied in the food industry.

Research Progress on Structure and Bioactivity of Longan Polysaccharide

Longan originates from southern China and has high nutritional and health value.Recent phytochemistry and pharmacology studies have shown that polysaccharides are a main bioactive component of longan.Longan polysaccharides possess antioxidant,anti-aging,anti-tumor,immunomodulatory,and other bioactivities.Hot-water extraction,ethanol precipitation,and ultrasonic extraction are generally used to extract water-soluble longan polysaccharides.However,the relationship between the structure and bioactivity of longan polysaccharides remains unclear,requiring further investigation.The aim of this review is to evaluate the current literature focusing on the extraction,purification,structural characterization,and biological activity of longan polysaccharides.We believe that this review would provide a useful bibliography for further innovation and a basis for using longan polysaccharides in functional food.

Metaverse virtual social center for elderly communication in time of social distancing

Background The lack of social activities among the elderly due to physical limitations can result in loneliness and depression. The spread of COVID-19 has made it difficult for the elderly to conduct stable social activities, increasing feelings of loneliness. The metaverse is a virtual world that mirrors reality. This allows the elderly to overcome the constraints of reality and perform social activities stably and continuously, providing new ideas for alleviating loneliness. Methods By analyzing their needs, a virtual social center framework for the elderly was proposed in this study. In addition, a prototype system was designed according to this framework. The elderly can socialize in virtual reality with metaverse-related technologies and human–computer interaction tools. Additionally, a test was conducted jointly with the chief physician of the geriatric rehabilitation department of a tertiary hospital. Results The results demonstrated that the mental state of the elderly who had used the virtual social center was significantly better than that of those who had not used it. Conclusions Thus, virtual social centers can help relieve loneliness and depression among the elderly with increasing global epidemics and an aging society. Hence,they have promotional value.

Effects of Mobile-Assisted Language Learning on Young Learners’Linguistic Skills

Mobile-assisted language learning(MALL)has been regarded as an excellent tool in the field of language acquisition as modern technology develops.The popularity of using mobile devices in education makes it possible for people to learn languages through platforms like tablets and smartphones from anywhere and at any time.However,the application of MALL in a collaborative student-centered environment has received comparatively little attention.Since spoken fluency and vocabulary size are the two crucial components of language proficiency,this study aims to investigate if young learners can improve their native language level through learning online beyond the classroom.The quantitative data reveal that MALL does make a difference in students’linguistic skills.The results show that the incorporation of mobile applications into language learning could better help learners achieve the learning outcomes and improve their communication skills than simply using conventional methods.In addition,the data of the questionnaire exposed some issues that need to be continuously improved.The viable suggestions are also discussed to share ideas about building a more sustainable learning environment in a data-driven age.

Synthesis of CC/BiPO4/Bi2WO6 Composite Material and Its Photocatalytic Performance

Along with the popularity of environmental protection concepts, the environmental treatment of water pollution attracts widespread attention, among which, the research on Bi-based semiconductor photocatalytic degradation technology has made great progress. However, the development of such bismuth-based composites still remains a challenging task due to difficult recovery and low catalytic efficiency. Herein, a novel CC/BiPO4/Bi2WO6 composite was successfully synthesized through two-step hydrothermal method using activated flexible carbon cloth as a substrate. The results of the photocatalytic degradation experiments showed that the obtained CC/BiPO4/Bi2WO6 composites can degrade 92.1% RhB in 60 min under UV-visible light irradiation, which was much higher than that of unloaded BiPO4 (24.4%) and BiPO4/Bi2WO6 (52.9%), exhibiting a better adsorption-photocatalytic degradation performance than BiPO4 and BiPO4/Bi2WO6. Photoluminescence spectra indicated that the improved photocatalytic activity was due to the more effective inhibition of photogenerated carrier complexation. Furthermore, the radical capture experiments confirmed that h+, ·OH and O2- were the main active substances in the photocatalytic degradation process of RhB by the CC/BiPO4/Bi2WO6 composites. More importantly, the prepared CC/BiPO4/Bi2WO6 composite had a simple separation process and good recycling stability, and its photocatalytic degradation efficiency can still reach 53.3% after six cycles of RhB degradation. .

An Approach for Human Posture Recognition Based on the Fusion PSE-CNN-BiGRU Model

This study proposes a pose estimation-convolutional neural network-bidirectional gated recurrent unit(PSECNN-BiGRU)fusion model for human posture recognition to address low accuracy issues in abnormal posture recognition due to the loss of some feature information and the deterioration of comprehensive performance in model detection in complex home environments.Firstly,the deep convolutional network is integrated with the Mediapipe framework to extract high-precision,multi-dimensional information from the key points of the human skeleton,thereby obtaining a human posture feature set.Thereafter,a double-layer BiGRU algorithm is utilized to extract multi-layer,bidirectional temporal features from the human posture feature set,and a CNN network with an exponential linear unit(ELU)activation function is adopted to perform deep convolution of the feature map to extract the spatial feature of the human posture.Furthermore,a squeeze and excitation networks(SENet)module is introduced to adaptively learn the importance weights of each channel,enhancing the network’s focus on important features.Finally,comparative experiments are performed on available datasets,including the public human activity recognition using smartphone dataset(UCIHAR),the public human activity recognition 70 plus dataset(HAR70PLUS),and the independently developed home abnormal behavior recognition dataset(HABRD)created by the authors’team.The results show that the average accuracy of the proposed PSE-CNN-BiGRU fusion model for human posture recognition is 99.56%,89.42%,and 98.90%,respectively,which are 5.24%,5.83%,and 3.19%higher than the average accuracy of the five models proposed in the comparative literature,including CNN,GRU,and others.The F1-score for abnormal posture recognition reaches 98.84%(heartache),97.18%(fall),99.6%(bellyache),and 98.27%(climbing)on the self-builtHABRDdataset,thus verifying the effectiveness,generalization,and robustness of the proposed model in enhancing human posture recognition.

Research on Evacuation Path Planning Based on Improved Sparrow Search Algorithm

Reducing casualties and property losses through effective evacuation route planning has been a key focus for researchers in recent years.As part of this effort,an enhanced sparrow search algorithm(MSSA)was proposed.Firstly,the Golden Sine algorithm and a nonlinear weight factor optimization strategy were added in the discoverer position update stage of the SSA algorithm.Secondly,the Cauchy-Gaussian perturbation was applied to the optimal position of the SSA algorithm to improve its ability to jump out of local optima.Finally,the local search mechanism based on the mountain climbing method was incorporated into the local search stage of the SSA algorithm,improving its local search ability.To evaluate the effectiveness of the proposed algorithm,the Whale Algorithm,Gray Wolf Algorithm,Improved Gray Wolf Algorithm,Sparrow Search Algorithm,and MSSA Algorithm were employed to solve various test functions.The accuracy and convergence speed of each algorithm were then compared and analyzed.The results indicate that the MSSA algorithm has superior solving ability and stability compared to other algorithms.To further validate the enhanced algorithm’s capabilities for path planning,evacuation experiments were conducted using different maps featuring various obstacle types.Additionally,a multi-exit evacuation scenario was constructed according to the actual building environment of a teaching building.Both the sparrow search algorithm and MSSA algorithm were employed in the simulation experiment for multiexit evacuation path planning.The findings demonstrate that the MSSA algorithm outperforms the comparison algorithm,showcasing its greater advantages and higher application potential.

Multimodal Sentiment Analysis Based on a Cross-Modal Multihead Attention Mechanism

Multimodal sentiment analysis aims to understand people’s emotions and opinions from diverse data.Concate-nating or multiplying various modalities is a traditional multi-modal sentiment analysis fusion method.This fusion method does not utilize the correlation information between modalities.To solve this problem,this paper proposes amodel based on amulti-head attention mechanism.First,after preprocessing the original data.Then,the feature representation is converted into a sequence of word vectors and positional encoding is introduced to better understand the semantic and sequential information in the input sequence.Next,the input coding sequence is fed into the transformer model for further processing and learning.At the transformer layer,a cross-modal attention consisting of a pair of multi-head attention modules is employed to reflect the correlation between modalities.Finally,the processed results are input into the feedforward neural network to obtain the emotional output through the classification layer.Through the above processing flow,the model can capture semantic information and contextual relationships and achieve good results in various natural language processing tasks.Our model was tested on the CMU Multimodal Opinion Sentiment and Emotion Intensity(CMU-MOSEI)and Multimodal EmotionLines Dataset(MELD),achieving an accuracy of 82.04% and F1 parameters reached 80.59% on the former dataset.