Multi-label feature selection(MFS)is a crucial dimensionality reduction technique aimed at identifying informative features associated with multiple labels.However,traditional centralized methods face significant chal...Multi-label feature selection(MFS)is a crucial dimensionality reduction technique aimed at identifying informative features associated with multiple labels.However,traditional centralized methods face significant challenges in privacy-sensitive and distributed settings,often neglecting label dependencies and suffering from low computational efficiency.To address these issues,we introduce a novel framework,Fed-MFSDHBCPSO—federated MFS via dual-layer hybrid breeding cooperative particle swarm optimization algorithm with manifold and sparsity regularization(DHBCPSO-MSR).Leveraging the federated learning paradigm,Fed-MFSDHBCPSO allows clients to perform local feature selection(FS)using DHBCPSO-MSR.Locally selected feature subsets are encrypted with differential privacy(DP)and transmitted to a central server,where they are securely aggregated and refined through secure multi-party computation(SMPC)until global convergence is achieved.Within each client,DHBCPSO-MSR employs a dual-layer FS strategy.The inner layer constructs sample and label similarity graphs,generates Laplacian matrices to capture the manifold structure between samples and labels,and applies L2,1-norm regularization to sparsify the feature subset,yielding an optimized feature weight matrix.The outer layer uses a hybrid breeding cooperative particle swarm optimization algorithm to further refine the feature weight matrix and identify the optimal feature subset.The updated weight matrix is then fed back to the inner layer for further optimization.Comprehensive experiments on multiple real-world multi-label datasets demonstrate that Fed-MFSDHBCPSO consistently outperforms both centralized and federated baseline methods across several key evaluation metrics.展开更多
Underwater images frequently suffer from chromatic distortion,blurred details,and low contrast,posing significant challenges for enhancement.This paper introduces AquaTree,a novel underwater image enhancement(UIE)meth...Underwater images frequently suffer from chromatic distortion,blurred details,and low contrast,posing significant challenges for enhancement.This paper introduces AquaTree,a novel underwater image enhancement(UIE)method that reformulates the task as a Markov Decision Process(MDP)through the integration of Monte Carlo Tree Search(MCTS)and deep reinforcement learning(DRL).The framework employs an action space of 25 enhancement operators,strategically grouped for basic attribute adjustment,color component balance,correction,and deblurring.Exploration within MCTS is guided by a dual-branch convolutional network,enabling intelligent sequential operator selection.Our core contributions include:(1)a multimodal state representation combining CIELab color histograms with deep perceptual features,(2)a dual-objective reward mechanism optimizing chromatic fidelity and perceptual consistency,and(3)an alternating training strategy co-optimizing enhancement sequences and network parameters.We further propose two inference schemes:an MCTS-based approach prioritizing accuracy at higher computational cost,and an efficient network policy enabling real-time processing with minimal quality loss.Comprehensive evaluations on the UIEB Dataset and Color correction and haze removal comparisons on the U45 Dataset demonstrate AquaTree’s superiority,significantly outperforming nine state-of-the-art methods across five established underwater image quality metrics.展开更多
The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location re...The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location reidentification and correlation attacks.To address these challenges,privacy-preserving trajectory generation methods are critical for applications relying on sensitive location data.This paper introduces DPIL-Traj,an advanced framework designed to generate synthetic trajectories while achieving a superior balance between data utility and privacy preservation.Firstly,the framework incorporates Differential Privacy Clustering,which anonymizes trajectory data by applying differential privacy techniques that add noise,ensuring the protection of sensitive user information.Secondly,Imitation Learning is used to replicate decision-making behaviors observed in real-world trajectories.By learning from expert trajectories,this component generates synthetic data that closely mimics real-world decision-making processes while optimizing the quality of the generated trajectories.Finally,Markov-based Trajectory Generation is employed to capture and maintain the inherent temporal dynamics of movement patterns.Extensive experiments conducted on the GeoLife trajectory dataset show that DPIL-Traj improves utility performance by an average of 19.85%,and in terms of privacy performance by an average of 12.51%,compared to state-of-the-art approaches.Ablation studies further reveal that DP clustering effectively safeguards privacy,imitation learning enhances utility under noise,and the Markov module strengthens temporal coherence.展开更多
Power-over-fiber technology provides several advantages for harsh industrial application and can potentially be widely used in electric power system. In this paper, a reliable power-over-fiber system is presented, whi...Power-over-fiber technology provides several advantages for harsh industrial application and can potentially be widely used in electric power system. In this paper, a reliable power-over-fiber system is presented, which is able to adjust the optical power automatically corresponding to the variance of the output load. An optic-electric feedback energy management scheme is designed and experimentally verified. The static and dynamic responses of the system towards load variances are further tested. The results show that the working status of the laser can be dynamically changed according to the output load, rather than maintaining the laser at the status of high-power output, which will potentially prolong the life and enhance the reliability of the system.展开更多
The Multilayer Perceptron(MLP)is a fundamental neural network model widely applied in various domains,particularly for lightweight image classification,speech recognition,and natural language processing tasks.Despite ...The Multilayer Perceptron(MLP)is a fundamental neural network model widely applied in various domains,particularly for lightweight image classification,speech recognition,and natural language processing tasks.Despite its widespread success,training MLPs often encounter significant challenges,including susceptibility to local optima,slow convergence rates,and high sensitivity to initial weight configurations.To address these issues,this paper proposes a Latin Hypercube Opposition-based Elite Variation Artificial Protozoa Optimizer(LOEV-APO),which enhances both global exploration and local exploitation simultaneously.LOEV-APO introduces a hybrid initialization strategy that combines Latin Hypercube Sampling(LHS)with Opposition-Based Learning(OBL),thus improving the diversity and coverage of the initial population.Moreover,an Elite Protozoa Variation Strategy(EPVS)is incorporated,which applies differential mutation operations to elite candidates,accelerating convergence and strengthening local search capabilities around high-quality solutions.Extensive experiments are conducted on six classification tasks and four function approximation tasks,covering a wide range of problem complexities and demonstrating superior generalization performance.The results demonstrate that LOEV-APO consistently outperforms nine state-of-the-art metaheuristic algorithms and two gradient-based methods in terms of convergence speed,solution accuracy,and robustness.These findings suggest that LOEV-APO serves as a promising optimization tool for MLP training and provides a viable alternative to traditional gradient-based methods.展开更多
Deep learning-based semantic communication has achieved remarkable progress with CNNs and Transformers.However,CNNs exhibit constrained performance in high-resolution image transmission,while Transformers incur high c...Deep learning-based semantic communication has achieved remarkable progress with CNNs and Transformers.However,CNNs exhibit constrained performance in high-resolution image transmission,while Transformers incur high computational cost due to quadratic complexity.Recently,VMamba,a novel state space model with linear complexity and exceptional long-range dependency modeling capabilities,has shown great potential in computer vision tasks.Inspired by this,we propose MNTSCC,an efficient VMamba-based nonlinear joint source-channel coding(JSCC)model for wireless image transmission.Specifically,MNTSCC comprises a VMamba-based nonlinear transform module,an MCAM entropy model,and a JSCC module.In the encoding stage,the input image is first encoded into a latent representation via the nonlinear transformation module,which is then processed by the MCAM for source distribution modeling.The JSCC module then optimizes transmission efficiency by adaptively assigning transmission rate to the latent representation according to the estimated entropy values.The proposedMCAMenhances the channel-wise autoregressive entropy model with attention mechanisms,which enables the entropy model to effectively capture both global and local information within latent features,thereby enabling more accurate entropy estimation and improved rate-distortion performance.Additionally,to further enhance the robustness of the system under varying signal-to-noise ratio(SNR)conditions,we incorporate SNR adaptive net(SAnet)into the JSCCmodule,which dynamically adjusts the encoding strategy by integrating SNRinformationwith latent features,thereby improving SNR adaptability.Experimental results across diverse resolution datasets demonstrate that the proposed method achieves superior image transmission performance compared to existing CNN-and Transformer-based semantic communication models,while maintaining competitive computational efficiency.In particular,under an Additive White Gaussian Noise(AWGN)channel with SNR=10 dB and a channel bandwidth ratio(CBR)of 1/16,MNTSCC consistently outperforms NTSCC,achieving a 1.72 dB Peak Signal-to-Noise Ratio(PSNR)gain on the Kodak24 dataset,0.79 dB on CLIC2022,and 2.54 dB on CIFAR-10,while reducing computational cost by 32.23%.The code is available at https://github.com/WanChen10/MNTSCC(accessed on 09 July 2025).展开更多
Accurate wind speed measurements on maritime vessels are crucial for weather forecasting,sea state prediction,and safe navigation.However,vessel motion and challenging environmental conditions often affect measurement...Accurate wind speed measurements on maritime vessels are crucial for weather forecasting,sea state prediction,and safe navigation.However,vessel motion and challenging environmental conditions often affect measurement precision.To address this issue,this study proposes an innovative framework for correcting and predicting shipborne wind speed.By integrating a main network with a momentum updating network,the proposed framework effectively extracts features from the time and frequency domains,thereby allowing for precise adjustments and predictions of shipborne wind speed data.Validation using real sensor data collected at the Qingdao Oceanographic Institute demonstrates that the proposed method outperforms existing approaches in single-and multi-step predictions compared to existing methods,achieving higher accuracy in wind speed forecasting.The proposed innovative approach offers a promising direction for future validation in more realistic maritime onboard scenarios.展开更多
The Dynamical Density Functional Theory(DDFT)algorithm,derived by associating classical Density Functional Theory(DFT)with the fundamental Smoluchowski dynamical equation,describes the evolution of inhomo-geneous flui...The Dynamical Density Functional Theory(DDFT)algorithm,derived by associating classical Density Functional Theory(DFT)with the fundamental Smoluchowski dynamical equation,describes the evolution of inhomo-geneous fluid density distributions over time.It plays a significant role in studying the evolution of density distributions over time in inhomogeneous systems.The Sunway Bluelight II supercomputer,as a new generation of China’s developed supercomputer,possesses powerful computational capabilities.Porting and optimizing industrial software on this platform holds significant importance.For the optimization of the DDFT algorithm,based on the Sunway Bluelight II supercomputer and the unique hardware architecture of the SW39000 processor,this work proposes three acceleration strategies to enhance computational efficiency and performance,including direct parallel optimization,local-memory constrained optimization for CPEs,and multi-core groups collaboration and communication optimization.This method combines the characteristics of the program’s algorithm with the unique hardware architecture of the Sunway Bluelight II supercomputer,optimizing the storage and transmission structures to achieve a closer integration of software and hardware.For the first time,this paper presents Sunway-Dynamical Density Functional Theory(SW-DDFT).Experimental results show that SW-DDFT achieves a speedup of 6.67 times within a single-core group compared to the original DDFT implementation,with six core groups(a total of 384 CPEs),the maximum speedup can reach 28.64 times,and parallel efficiency can reach 71%,demonstrating excellent acceleration performance.展开更多
The purpose of infrared and visible image fusion is to create a single image containing the texture details and significant object information of the source images,particularly in challenging environments.However,exis...The purpose of infrared and visible image fusion is to create a single image containing the texture details and significant object information of the source images,particularly in challenging environments.However,existing image fusion algorithms are generally suitable for normal scenes.In the hazy scene,a lot of texture information in the visible image is hidden,the results of existing methods are filled with infrared information,resulting in the lack of texture details and poor visual effect.To address the aforementioned difficulties,we propose a haze-free infrared and visible fusion method,termed HaIVFusion,which can eliminate the influence of haze and obtain richer texture information in the fused image.Specifically,we first design a scene information restoration network(SIRNet)to mine the masked texture information in visible images.Then,a denoising fusion network(DFNet)is designed to integrate the features extracted from infrared and visible images and remove the influence of residual noise as much as possible.In addition,we use color consistency loss to reduce the color distortion resulting from haze.Furthermore,we publish a dataset of hazy scenes for infrared and visible image fusion to promote research in extreme scenes.Extensive experiments show that HaIVFusion produces fused images with increased texture details and higher contrast in hazy scenes,and achieves better quantitative results,when compared to state-ofthe-art image fusion methods,even combined with state-of-the-art dehazing methods.展开更多
1.Data security in smart manufacturing The global manufacturing sector is undergoing a digital transformation as traditional systems-reliant on physical assets such as raw materials and labor-struggle to meet demands ...1.Data security in smart manufacturing The global manufacturing sector is undergoing a digital transformation as traditional systems-reliant on physical assets such as raw materials and labor-struggle to meet demands for greater flexibility and efficiency.The integration of advanced information technology facilitates smart manufacturing(SM),which optimizes production,management,and supply chains[1].展开更多
The increasing adoption of unmanned aerial vehicles(UAVs)in urban low-altitude logistics systems,particularly for time-sensitive applications like parcel delivery and supply distribution,necessitates sophisticated coo...The increasing adoption of unmanned aerial vehicles(UAVs)in urban low-altitude logistics systems,particularly for time-sensitive applications like parcel delivery and supply distribution,necessitates sophisticated coordination mechanisms to optimize operational efficiency.However,the limited capability of UAVs to extract stateaction information in complex environments poses significant challenges to achieving effective cooperation in dynamic and uncertain scenarios.To address this,we presents an Improved Multi-Agent Hybrid Attention Critic(IMAHAC)framework that advances multi-agent deep reinforcement learning(MADRL)through two key innovations.Firstly,a Temporal Difference Error and Time-based Prioritized Experience Replay(TT-PER)mechanism that dynamically adjusts sample weights based on temporal relevance and prediction error magnitude,effectively reducing the interference from obsolete collaborative experiences while maintaining training stability.Secondly,a hybrid attention mechanism is developed,integrating a sensor fusion layer—which aggregates features from multi-sensor data to enhance decision-making—and a dissimilarity layer that evaluates the similarity between key-value pairs and query values.By combining this hybrid attention mechanism with theMulti-Actor Attention Critic(MAAC)framework,our approach strengthens UAVs’capability to extract critical state-action features in diverse environments.Comprehensive simulations in urban air mobility scenarios demonstrate IMAHAC’s superiority over conventional MADRL baselines and MAAC,achieving higher cumulative rewards,fewer collisions,and enhanced cooperative capabilities.This work provides both algorithmic advancements and empirical validation for developing robust autonomous aerial systems in smart city infrastructures.展开更多
Due to the development of cloud computing and machine learning,users can upload their data to the cloud for machine learning model training.However,dishonest clouds may infer user data,resulting in user data leakage.P...Due to the development of cloud computing and machine learning,users can upload their data to the cloud for machine learning model training.However,dishonest clouds may infer user data,resulting in user data leakage.Previous schemes have achieved secure outsourced computing,but they suffer from low computational accuracy,difficult-to-handle heterogeneous distribution of data from multiple sources,and high computational cost,which result in extremely poor user experience and expensive cloud computing costs.To address the above problems,we propose amulti-precision,multi-sourced,andmulti-key outsourcing neural network training scheme.Firstly,we design a multi-precision functional encryption computation based on Euclidean division.Second,we design the outsourcing model training algorithm based on a multi-precision functional encryption with multi-sourced heterogeneity.Finally,we conduct experiments on three datasets.The results indicate that our framework achieves an accuracy improvement of 6%to 30%.Additionally,it offers a memory space optimization of 1.0×2^(24) times compared to the previous best approach.展开更多
As the 5G architecture gains momentum,interest in 6G is growing.The proliferation of Internet of Things(IoT)devices,capable of capturing sensitive images,has increased the need for secure transmission and robust acces...As the 5G architecture gains momentum,interest in 6G is growing.The proliferation of Internet of Things(IoT)devices,capable of capturing sensitive images,has increased the need for secure transmission and robust access control mechanisms.The vast amount of data generated by low-computing devices poses a challenge to traditional centralized access control,which relies on trusted third parties and complex computations,resulting in intricate interactions,higher hardware costs,and processing delays.To address these issues,this paper introduces a novel distributed access control approach that integrates a decentralized and lightweight encryption mechanism with image transmission.This method enhances data security and resource efficiency without imposing heavy computational and network burdens.In comparison to the best existing approach,it achieves a 7%improvement in accuracy,effectively addressing existing gaps in lightweight encryption and recognition performance.展开更多
Variations in ocean mixed layer depth(MLD)show a significant impact on energy balance in the global climate systems and marine ecosystems.At present,the accuracy of modeling MLD,especially in the region with complex o...Variations in ocean mixed layer depth(MLD)show a significant impact on energy balance in the global climate systems and marine ecosystems.At present,the accuracy of modeling MLD,especially in the region with complex ocean dynamics,remains a challenge,thus calling for an emergency using artificial intelligence approach to improve the assessment of the MLD.A novel convolutional neural network model was developed based on a dual-attention module(DA-CNN)to estimate the MLD in the Bay of Bengal(BoB)by integrating multi-source remote sensing data and Argo gridded data.Compared with the original CNN model,the DA-CNN model exhibits superior performance with notable improvements in the annual average root mean square error(RMSE)and R2 values by 13.0%and 8.4%,respectively,while more accurately capturing the seasonal variations in MLD.Moreover,the results using the DA-CNN model show minimum RMSE and maximum R2 values,in comparison to the calculation by the random forest,artificial neural network model,and the hybrid coordinate ocean model.Accordingly,our findings suggest that the newly developed DA-CNN model provides an effective advantage in studying the MLD and the associated ocean processes.展开更多
To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities...To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities,this paper proposes a Multi-modal Pre-synergistic Entity Alignmentmodel based on Cross-modalMutual Information Strategy Optimization(MPSEA).The model first employs independent encoders to process multi-modal features,including text,images,and numerical values.Next,a multi-modal pre-synergistic fusion mechanism integrates graph structural and visual modal features into the textual modality as preparatory information.This pre-fusion strategy enables unified perception of heterogeneous modalities at the model’s initial stage,reducing discrepancies during the fusion process.Finally,using cross-modal deep perception reinforcement learning,the model achieves adaptive multilevel feature fusion between modalities,supporting learningmore effective alignment strategies.Extensive experiments on multiple public datasets show that the MPSEA method achieves gains of up to 7% in Hits@1 and 8.2% in MRR on the FBDB15K dataset,and up to 9.1% in Hits@1 and 7.7% in MRR on the FBYG15K dataset,compared to existing state-of-the-art methods.These results confirm the effectiveness of the proposed model.展开更多
As Internet ofThings(IoT)technologies continue to evolve at an unprecedented pace,intelligent big data control and information systems have become critical enablers for organizational digital transformation,facilitati...As Internet ofThings(IoT)technologies continue to evolve at an unprecedented pace,intelligent big data control and information systems have become critical enablers for organizational digital transformation,facilitating data-driven decision making,fostering innovation ecosystems,and maintaining operational stability.In this study,we propose an advanced deployment algorithm for Service Function Chaining(SFC)that leverages an enhanced Practical Byzantine Fault Tolerance(PBFT)mechanism.The main goal is to tackle the issues of security and resource efficiency in SFC implementation across diverse network settings.By integrating blockchain technology and Deep Reinforcement Learning(DRL),our algorithm not only optimizes resource utilization and quality of service but also ensures robust security during SFC deployment.Specifically,the enhanced PBFT consensus mechanism(VRPBFT)significantly reduces consensus latency and improves Byzantine node detection through the introduction of a Verifiable Random Function(VRF)and a node reputation grading model.Experimental results demonstrate that compared to traditional PBFT,the proposed VRPBFT algorithm reduces consensus latency by approximately 30%and decreases the proportion of Byzantine nodes by 40%after 100 rounds of consensus.Furthermore,the DRL-based SFC deployment algorithm(SDRL)exhibits rapid convergence during training,with improvements in long-term average revenue,request acceptance rate,and revenue/cost ratio of 17%,14.49%,and 20.35%,respectively,over existing algorithms.Additionally,the CPU resource utilization of the SDRL algorithmreaches up to 42%,which is 27.96%higher than other algorithms.These findings indicate that the proposed algorithm substantially enhances resource utilization efficiency,service quality,and security in SFC deployment.展开更多
Frequent typhoons can significantly change the temperature,nutrient availability,and phytoplankton biomass in marginal seas.The oceanic response to typhoons is usually influenced by the features of the typhoon,among w...Frequent typhoons can significantly change the temperature,nutrient availability,and phytoplankton biomass in marginal seas.The oceanic response to typhoons is usually influenced by the features of the typhoon,among which the translational speed is critically important.By using a high resolution coupled physical-biological model,we investigated the response of the Yellow and East China seas(YECS)to two typhoons at different translational speeds,Muifa in August 2011 and Bolaven in August 2012.The model well reproduced the spatial and temporal variations of temperature,chlorophyll-a concentration over the YECS.Results show that typhoons with slower translational speeds uplift more deep water,leading to a more significant oceanic response.Divergence and convergence caused nutrient fluxes in opposite directions in the surface and bottom layers.Moreover,the nutrient flux in the bottom layer was greater than that in the surface layer.These phenomena are closely related to the spatial distribution of nutrients.Further studies show that the degree of ocean response to typhoons is highly correlated with the initial conditions of physical and biological elements of the upper ocean before the typhoon,as well as with ocean structure.Pretyphoon initial conditions of oceanic physical and ecological elements,mixed layer depth,and potential energy anomalies can all alter the degree of typhoon-induced oceanic response.This study emphasizes the important roles of the translational speed of typhoons and the initial oceanic conditions in the oceanic response to typhoons.展开更多
Different phenomenological equations based on plasticity, primary creep （as a viscoplastic mechanism）, secondary creep （as another viscoplastic mechanism） and different combinations of these equations are presente...Different phenomenological equations based on plasticity, primary creep （as a viscoplastic mechanism）, secondary creep （as another viscoplastic mechanism） and different combinations of these equations are presented and used to describe the material inelastic deformation in uniaxial test. Agreement of the models with experimental results and with the theoretical concepts and physical realities is the criterion of choosing the most appropriate formulation for uniaxial test. A model is thus proposed in which plastic deformation, primary creep and secondary creep contribute to the inelastic deformation. However, it is believed that the hardening parameter is composed of plastic and primary creep parts. Accordingly, the axial plastic strain in a uniaxial test may no longer be considered as the hardening parameter. Therefore, a proportionality concept is proposed to calculate the plastic contribution of deformation.展开更多
In underground engineering,the detection of structural cracks on tunnel surfaces stands as a pivotal task in ensuring the health and reliability of tunnel structures.However,the dim and dusty environment inherent to u...In underground engineering,the detection of structural cracks on tunnel surfaces stands as a pivotal task in ensuring the health and reliability of tunnel structures.However,the dim and dusty environment inherent to under-ground engineering poses considerable challenges to crack segmentation.This paper proposes a crack segmentation algorithm termed as Focused Detection for Subsurface Cracks YOLOv8(FDSC-YOLOv8)specifically designed for underground engineering structural surfaces.Firstly,to improve the extraction of multi-layer convolutional features,the fixed convolutional module is replaced with a deformable convolutional module.Secondly,the model’s receptive field is enhanced by introducing a multi-branch convolutional module,improving the extraction of shallow features for small targets.Next,the Dynamic Snake Convolution module is incorporated to enhance the extraction capability for slender and weak cracks.Finally,the Convolutional Block Attention Module(CBAM)module is employed to achieve better target determination.The FDSC-YOLOv8s algorithm’s mAP50 and mAP50-95 reach 96.5%and 66.4%,according to the testing data.展开更多
文摘Multi-label feature selection(MFS)is a crucial dimensionality reduction technique aimed at identifying informative features associated with multiple labels.However,traditional centralized methods face significant challenges in privacy-sensitive and distributed settings,often neglecting label dependencies and suffering from low computational efficiency.To address these issues,we introduce a novel framework,Fed-MFSDHBCPSO—federated MFS via dual-layer hybrid breeding cooperative particle swarm optimization algorithm with manifold and sparsity regularization(DHBCPSO-MSR).Leveraging the federated learning paradigm,Fed-MFSDHBCPSO allows clients to perform local feature selection(FS)using DHBCPSO-MSR.Locally selected feature subsets are encrypted with differential privacy(DP)and transmitted to a central server,where they are securely aggregated and refined through secure multi-party computation(SMPC)until global convergence is achieved.Within each client,DHBCPSO-MSR employs a dual-layer FS strategy.The inner layer constructs sample and label similarity graphs,generates Laplacian matrices to capture the manifold structure between samples and labels,and applies L2,1-norm regularization to sparsify the feature subset,yielding an optimized feature weight matrix.The outer layer uses a hybrid breeding cooperative particle swarm optimization algorithm to further refine the feature weight matrix and identify the optimal feature subset.The updated weight matrix is then fed back to the inner layer for further optimization.Comprehensive experiments on multiple real-world multi-label datasets demonstrate that Fed-MFSDHBCPSO consistently outperforms both centralized and federated baseline methods across several key evaluation metrics.
基金supported by theHubei Provincial Technology Innovation Special Project and the Natural Science Foundation of Hubei Province under Grants 2023BEB024,2024AFC066,respectively.
文摘Underwater images frequently suffer from chromatic distortion,blurred details,and low contrast,posing significant challenges for enhancement.This paper introduces AquaTree,a novel underwater image enhancement(UIE)method that reformulates the task as a Markov Decision Process(MDP)through the integration of Monte Carlo Tree Search(MCTS)and deep reinforcement learning(DRL).The framework employs an action space of 25 enhancement operators,strategically grouped for basic attribute adjustment,color component balance,correction,and deblurring.Exploration within MCTS is guided by a dual-branch convolutional network,enabling intelligent sequential operator selection.Our core contributions include:(1)a multimodal state representation combining CIELab color histograms with deep perceptual features,(2)a dual-objective reward mechanism optimizing chromatic fidelity and perceptual consistency,and(3)an alternating training strategy co-optimizing enhancement sequences and network parameters.We further propose two inference schemes:an MCTS-based approach prioritizing accuracy at higher computational cost,and an efficient network policy enabling real-time processing with minimal quality loss.Comprehensive evaluations on the UIEB Dataset and Color correction and haze removal comparisons on the U45 Dataset demonstrate AquaTree’s superiority,significantly outperforming nine state-of-the-art methods across five established underwater image quality metrics.
基金supported by the Natural Science Foundation of Fujian Province of China(2025J01380)National Natural Science Foundation of China(No.62471139)+3 种基金the Major Health Research Project of Fujian Province(2021ZD01001)Fujian Provincial Units Special Funds for Education and Research(2022639)Fujian University of Technology Research Start-up Fund(GY-S24002)Fujian Research and Training Grants for Young and Middle-aged Leaders in Healthcare(GY-H-24179).
文摘The generation of synthetic trajectories has become essential in various fields for analyzing complex movement patterns.However,the use of real-world trajectory data poses significant privacy risks,such as location reidentification and correlation attacks.To address these challenges,privacy-preserving trajectory generation methods are critical for applications relying on sensitive location data.This paper introduces DPIL-Traj,an advanced framework designed to generate synthetic trajectories while achieving a superior balance between data utility and privacy preservation.Firstly,the framework incorporates Differential Privacy Clustering,which anonymizes trajectory data by applying differential privacy techniques that add noise,ensuring the protection of sensitive user information.Secondly,Imitation Learning is used to replicate decision-making behaviors observed in real-world trajectories.By learning from expert trajectories,this component generates synthetic data that closely mimics real-world decision-making processes while optimizing the quality of the generated trajectories.Finally,Markov-based Trajectory Generation is employed to capture and maintain the inherent temporal dynamics of movement patterns.Extensive experiments conducted on the GeoLife trajectory dataset show that DPIL-Traj improves utility performance by an average of 19.85%,and in terms of privacy performance by an average of 12.51%,compared to state-of-the-art approaches.Ablation studies further reveal that DP clustering effectively safeguards privacy,imitation learning enhances utility under noise,and the Markov module strengthens temporal coherence.
基金supported by the National Natural Science Foundation of China(61203129,61174038,61473151,51507080)the Fundamental Research Funds for the Central Universities(30915011104,30920130121010,30920140112005)
基金supported by the project of The State Grid Corporation of China “Power by Laser Key Technology Research for Remote Power Supply of Sensing Devices in High Voltage Network with High Electrical Potential”,SGRIXTKJ[2017](No.840)
文摘Power-over-fiber technology provides several advantages for harsh industrial application and can potentially be widely used in electric power system. In this paper, a reliable power-over-fiber system is presented, which is able to adjust the optical power automatically corresponding to the variance of the output load. An optic-electric feedback energy management scheme is designed and experimentally verified. The static and dynamic responses of the system towards load variances are further tested. The results show that the working status of the laser can be dynamically changed according to the output load, rather than maintaining the laser at the status of high-power output, which will potentially prolong the life and enhance the reliability of the system.
基金supported by the National Natural Science Foundation of China(Grant Nos.62376089,62302153,62302154)the Key Research and Development Program of Hubei Province,China(Grant No.2023BEB024)+1 种基金the Young and Middle-Aged Scientific and Technological Innovation Team Plan in Higher Education Institutions in Hubei Province,China(Grant No.T2023007)the National Natural Science Foundation of China(Grant No.U23A20318).
文摘The Multilayer Perceptron(MLP)is a fundamental neural network model widely applied in various domains,particularly for lightweight image classification,speech recognition,and natural language processing tasks.Despite its widespread success,training MLPs often encounter significant challenges,including susceptibility to local optima,slow convergence rates,and high sensitivity to initial weight configurations.To address these issues,this paper proposes a Latin Hypercube Opposition-based Elite Variation Artificial Protozoa Optimizer(LOEV-APO),which enhances both global exploration and local exploitation simultaneously.LOEV-APO introduces a hybrid initialization strategy that combines Latin Hypercube Sampling(LHS)with Opposition-Based Learning(OBL),thus improving the diversity and coverage of the initial population.Moreover,an Elite Protozoa Variation Strategy(EPVS)is incorporated,which applies differential mutation operations to elite candidates,accelerating convergence and strengthening local search capabilities around high-quality solutions.Extensive experiments are conducted on six classification tasks and four function approximation tasks,covering a wide range of problem complexities and demonstrating superior generalization performance.The results demonstrate that LOEV-APO consistently outperforms nine state-of-the-art metaheuristic algorithms and two gradient-based methods in terms of convergence speed,solution accuracy,and robustness.These findings suggest that LOEV-APO serves as a promising optimization tool for MLP training and provides a viable alternative to traditional gradient-based methods.
文摘Deep learning-based semantic communication has achieved remarkable progress with CNNs and Transformers.However,CNNs exhibit constrained performance in high-resolution image transmission,while Transformers incur high computational cost due to quadratic complexity.Recently,VMamba,a novel state space model with linear complexity and exceptional long-range dependency modeling capabilities,has shown great potential in computer vision tasks.Inspired by this,we propose MNTSCC,an efficient VMamba-based nonlinear joint source-channel coding(JSCC)model for wireless image transmission.Specifically,MNTSCC comprises a VMamba-based nonlinear transform module,an MCAM entropy model,and a JSCC module.In the encoding stage,the input image is first encoded into a latent representation via the nonlinear transformation module,which is then processed by the MCAM for source distribution modeling.The JSCC module then optimizes transmission efficiency by adaptively assigning transmission rate to the latent representation according to the estimated entropy values.The proposedMCAMenhances the channel-wise autoregressive entropy model with attention mechanisms,which enables the entropy model to effectively capture both global and local information within latent features,thereby enabling more accurate entropy estimation and improved rate-distortion performance.Additionally,to further enhance the robustness of the system under varying signal-to-noise ratio(SNR)conditions,we incorporate SNR adaptive net(SAnet)into the JSCCmodule,which dynamically adjusts the encoding strategy by integrating SNRinformationwith latent features,thereby improving SNR adaptability.Experimental results across diverse resolution datasets demonstrate that the proposed method achieves superior image transmission performance compared to existing CNN-and Transformer-based semantic communication models,while maintaining competitive computational efficiency.In particular,under an Additive White Gaussian Noise(AWGN)channel with SNR=10 dB and a channel bandwidth ratio(CBR)of 1/16,MNTSCC consistently outperforms NTSCC,achieving a 1.72 dB Peak Signal-to-Noise Ratio(PSNR)gain on the Kodak24 dataset,0.79 dB on CLIC2022,and 2.54 dB on CIFAR-10,while reducing computational cost by 32.23%.The code is available at https://github.com/WanChen10/MNTSCC(accessed on 09 July 2025).
基金supported by the Major Innovation Project for the Integration of Science,Education,and Industry of Qilu University of Technology(Shandong Academy of Sciences)(Nos.2023HYZX01,2023JBZ02)the Open Project of Key Laboratory of Computing Power Network and Information Security,Ministry of Education,Qilu University of Technology(Shandong Academy of Sciences)(No.2023ZD007)+2 种基金the Talent Research Projects of Qilu University of Technology(Shandong Academy of Sciences)(No.2023RCKY136)the Technology and Innovation Major Project of the Ministry of Science and Technology of China(No.2022ZD0118600)the Jinan‘20 New Colleges and Universities’Funded Project(No.202333043)。
文摘Accurate wind speed measurements on maritime vessels are crucial for weather forecasting,sea state prediction,and safe navigation.However,vessel motion and challenging environmental conditions often affect measurement precision.To address this issue,this study proposes an innovative framework for correcting and predicting shipborne wind speed.By integrating a main network with a momentum updating network,the proposed framework effectively extracts features from the time and frequency domains,thereby allowing for precise adjustments and predictions of shipborne wind speed data.Validation using real sensor data collected at the Qingdao Oceanographic Institute demonstrates that the proposed method outperforms existing approaches in single-and multi-step predictions compared to existing methods,achieving higher accuracy in wind speed forecasting.The proposed innovative approach offers a promising direction for future validation in more realistic maritime onboard scenarios.
基金supported by National Key Research and Development Program of China under Grant 2024YFE0210800National Natural Science Foundation of China under Grant 62495062Beijing Natural Science Foundation under Grant L242017.
文摘The Dynamical Density Functional Theory(DDFT)algorithm,derived by associating classical Density Functional Theory(DFT)with the fundamental Smoluchowski dynamical equation,describes the evolution of inhomo-geneous fluid density distributions over time.It plays a significant role in studying the evolution of density distributions over time in inhomogeneous systems.The Sunway Bluelight II supercomputer,as a new generation of China’s developed supercomputer,possesses powerful computational capabilities.Porting and optimizing industrial software on this platform holds significant importance.For the optimization of the DDFT algorithm,based on the Sunway Bluelight II supercomputer and the unique hardware architecture of the SW39000 processor,this work proposes three acceleration strategies to enhance computational efficiency and performance,including direct parallel optimization,local-memory constrained optimization for CPEs,and multi-core groups collaboration and communication optimization.This method combines the characteristics of the program’s algorithm with the unique hardware architecture of the Sunway Bluelight II supercomputer,optimizing the storage and transmission structures to achieve a closer integration of software and hardware.For the first time,this paper presents Sunway-Dynamical Density Functional Theory(SW-DDFT).Experimental results show that SW-DDFT achieves a speedup of 6.67 times within a single-core group compared to the original DDFT implementation,with six core groups(a total of 384 CPEs),the maximum speedup can reach 28.64 times,and parallel efficiency can reach 71%,demonstrating excellent acceleration performance.
基金supported by the Natural Science Foundation of Shandong Province,China(ZR2022MF237)the National Natural Science Foundation of China Youth Fund(62406155)the Major Innovation Project(2023JBZ02)of Qilu University of Technology(Shandong Academy of Sciences).
文摘The purpose of infrared and visible image fusion is to create a single image containing the texture details and significant object information of the source images,particularly in challenging environments.However,existing image fusion algorithms are generally suitable for normal scenes.In the hazy scene,a lot of texture information in the visible image is hidden,the results of existing methods are filled with infrared information,resulting in the lack of texture details and poor visual effect.To address the aforementioned difficulties,we propose a haze-free infrared and visible fusion method,termed HaIVFusion,which can eliminate the influence of haze and obtain richer texture information in the fused image.Specifically,we first design a scene information restoration network(SIRNet)to mine the masked texture information in visible images.Then,a denoising fusion network(DFNet)is designed to integrate the features extracted from infrared and visible images and remove the influence of residual noise as much as possible.In addition,we use color consistency loss to reduce the color distortion resulting from haze.Furthermore,we publish a dataset of hazy scenes for infrared and visible image fusion to promote research in extreme scenes.Extensive experiments show that HaIVFusion produces fused images with increased texture details and higher contrast in hazy scenes,and achieves better quantitative results,when compared to state-ofthe-art image fusion methods,even combined with state-of-the-art dehazing methods.
基金supported in part by the National Natural Science Foundation of China(62293511 and 62402256)in part by the Shandong Provincial Natural Science Foundation of China(ZR2024MF100)+1 种基金in part by the Taishan Scholars Program(tsqn202408239)in part by the Open Research Project of the State Key Laboratory of Industrial Control Technology,Zhejiang University,China(ICT2025B13).
文摘1.Data security in smart manufacturing The global manufacturing sector is undergoing a digital transformation as traditional systems-reliant on physical assets such as raw materials and labor-struggle to meet demands for greater flexibility and efficiency.The integration of advanced information technology facilitates smart manufacturing(SM),which optimizes production,management,and supply chains[1].
基金supported by theHubei Provincial Technology Innovation Special Project and the Natural Science Foundation of Hubei Province under Grants 2023BEB024,2024AFC066,respectively.
文摘The increasing adoption of unmanned aerial vehicles(UAVs)in urban low-altitude logistics systems,particularly for time-sensitive applications like parcel delivery and supply distribution,necessitates sophisticated coordination mechanisms to optimize operational efficiency.However,the limited capability of UAVs to extract stateaction information in complex environments poses significant challenges to achieving effective cooperation in dynamic and uncertain scenarios.To address this,we presents an Improved Multi-Agent Hybrid Attention Critic(IMAHAC)framework that advances multi-agent deep reinforcement learning(MADRL)through two key innovations.Firstly,a Temporal Difference Error and Time-based Prioritized Experience Replay(TT-PER)mechanism that dynamically adjusts sample weights based on temporal relevance and prediction error magnitude,effectively reducing the interference from obsolete collaborative experiences while maintaining training stability.Secondly,a hybrid attention mechanism is developed,integrating a sensor fusion layer—which aggregates features from multi-sensor data to enhance decision-making—and a dissimilarity layer that evaluates the similarity between key-value pairs and query values.By combining this hybrid attention mechanism with theMulti-Actor Attention Critic(MAAC)framework,our approach strengthens UAVs’capability to extract critical state-action features in diverse environments.Comprehensive simulations in urban air mobility scenarios demonstrate IMAHAC’s superiority over conventional MADRL baselines and MAAC,achieving higher cumulative rewards,fewer collisions,and enhanced cooperative capabilities.This work provides both algorithmic advancements and empirical validation for developing robust autonomous aerial systems in smart city infrastructures.
基金supported by Natural Science Foundation of China(Nos.62303126,62362008,author Z.Z,https://www.nsfc.gov.cn/,accessed on 20 December 2024)Major Scientific and Technological Special Project of Guizhou Province([2024]014)+2 种基金Guizhou Provincial Science and Technology Projects(No.ZK[2022]General149) ,author Z.Z,https://kjt.guizhou.gov.cn/,accessed on 20 December 2024)The Open Project of the Key Laboratory of Computing Power Network and Information Security,Ministry of Education under Grant 2023ZD037,author Z.Z,https://www.gzu.edu.cn/,accessed on 20 December 2024)Open Research Project of the State Key Laboratory of Industrial Control Technology,Zhejiang University,China(No.ICT2024B25),author Z.Z,https://www.gzu.edu.cn/,accessed on 20 December 2024).
文摘Due to the development of cloud computing and machine learning,users can upload their data to the cloud for machine learning model training.However,dishonest clouds may infer user data,resulting in user data leakage.Previous schemes have achieved secure outsourced computing,but they suffer from low computational accuracy,difficult-to-handle heterogeneous distribution of data from multiple sources,and high computational cost,which result in extremely poor user experience and expensive cloud computing costs.To address the above problems,we propose amulti-precision,multi-sourced,andmulti-key outsourcing neural network training scheme.Firstly,we design a multi-precision functional encryption computation based on Euclidean division.Second,we design the outsourcing model training algorithm based on a multi-precision functional encryption with multi-sourced heterogeneity.Finally,we conduct experiments on three datasets.The results indicate that our framework achieves an accuracy improvement of 6%to 30%.Additionally,it offers a memory space optimization of 1.0×2^(24) times compared to the previous best approach.
基金supported in part by the National Natural Science Foundation of China under Grants(62250410365,62071084)the Youth Program of Humanities and Social Sciences of the MoE(23YJCZH291)+1 种基金the Key Laboratory of Computing Power Network and Information Security,Ministry of Education(2023ZD02)Deanship of Research and Graduate Studies at King Khalid University for funding this work through Large Research Project under grant number RGP2/15/46.
文摘As the 5G architecture gains momentum,interest in 6G is growing.The proliferation of Internet of Things(IoT)devices,capable of capturing sensitive images,has increased the need for secure transmission and robust access control mechanisms.The vast amount of data generated by low-computing devices poses a challenge to traditional centralized access control,which relies on trusted third parties and complex computations,resulting in intricate interactions,higher hardware costs,and processing delays.To address these issues,this paper introduces a novel distributed access control approach that integrates a decentralized and lightweight encryption mechanism with image transmission.This method enhances data security and resource efficiency without imposing heavy computational and network burdens.In comparison to the best existing approach,it achieves a 7%improvement in accuracy,effectively addressing existing gaps in lightweight encryption and recognition performance.
基金Supported by the Ministry of Science and Technology of the People’s Republic of China(No.2019 YFE 0125000)the National Natural Science Foundation of China(No.42376032)。
文摘Variations in ocean mixed layer depth(MLD)show a significant impact on energy balance in the global climate systems and marine ecosystems.At present,the accuracy of modeling MLD,especially in the region with complex ocean dynamics,remains a challenge,thus calling for an emergency using artificial intelligence approach to improve the assessment of the MLD.A novel convolutional neural network model was developed based on a dual-attention module(DA-CNN)to estimate the MLD in the Bay of Bengal(BoB)by integrating multi-source remote sensing data and Argo gridded data.Compared with the original CNN model,the DA-CNN model exhibits superior performance with notable improvements in the annual average root mean square error(RMSE)and R2 values by 13.0%and 8.4%,respectively,while more accurately capturing the seasonal variations in MLD.Moreover,the results using the DA-CNN model show minimum RMSE and maximum R2 values,in comparison to the calculation by the random forest,artificial neural network model,and the hybrid coordinate ocean model.Accordingly,our findings suggest that the newly developed DA-CNN model provides an effective advantage in studying the MLD and the associated ocean processes.
基金partially supported by the National Natural Science Foundation of China under Grants 62471493 and 62402257(for conceptualization and investigation)partially supported by the Natural Science Foundation of Shandong Province,China under Grants ZR2023LZH017,ZR2024MF066,and 2023QF025(for formal analysis and validation)+1 种基金partially supported by the Open Foundation of Key Laboratory of Computing Power Network and Information Security,Ministry of Education,Qilu University of Technology(Shandong Academy of Sciences)under Grant 2023ZD010(for methodology and model design)partially supported by the Russian Science Foundation(RSF)Project under Grant 22-71-10095-P(for validation and results verification).
文摘To address the challenge of missing modal information in entity alignment and to mitigate information loss or bias arising frommodal heterogeneity during fusion,while also capturing shared information acrossmodalities,this paper proposes a Multi-modal Pre-synergistic Entity Alignmentmodel based on Cross-modalMutual Information Strategy Optimization(MPSEA).The model first employs independent encoders to process multi-modal features,including text,images,and numerical values.Next,a multi-modal pre-synergistic fusion mechanism integrates graph structural and visual modal features into the textual modality as preparatory information.This pre-fusion strategy enables unified perception of heterogeneous modalities at the model’s initial stage,reducing discrepancies during the fusion process.Finally,using cross-modal deep perception reinforcement learning,the model achieves adaptive multilevel feature fusion between modalities,supporting learningmore effective alignment strategies.Extensive experiments on multiple public datasets show that the MPSEA method achieves gains of up to 7% in Hits@1 and 8.2% in MRR on the FBDB15K dataset,and up to 9.1% in Hits@1 and 7.7% in MRR on the FBYG15K dataset,compared to existing state-of-the-art methods.These results confirm the effectiveness of the proposed model.
基金supported by the National Natural Science Foundation of China under Grant 62471493 and 62402257partially supported by the Natural Science Foundation of Shandong Province under Grant ZR2023LZH017,ZR2024MF066 and 2023QF025+2 种基金partially supported by the Open Research Subject of State Key Laboratory of Intelligent Game(No.ZBKF-24-12)partially supported by the Foundation of Key Laboratory of Education Informatization for Nationalities(Yunnan Normal University),the Ministry of Education(No.EIN2024C006)partially supported by the Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance of MOE(No.202306).
文摘As Internet ofThings(IoT)technologies continue to evolve at an unprecedented pace,intelligent big data control and information systems have become critical enablers for organizational digital transformation,facilitating data-driven decision making,fostering innovation ecosystems,and maintaining operational stability.In this study,we propose an advanced deployment algorithm for Service Function Chaining(SFC)that leverages an enhanced Practical Byzantine Fault Tolerance(PBFT)mechanism.The main goal is to tackle the issues of security and resource efficiency in SFC implementation across diverse network settings.By integrating blockchain technology and Deep Reinforcement Learning(DRL),our algorithm not only optimizes resource utilization and quality of service but also ensures robust security during SFC deployment.Specifically,the enhanced PBFT consensus mechanism(VRPBFT)significantly reduces consensus latency and improves Byzantine node detection through the introduction of a Verifiable Random Function(VRF)and a node reputation grading model.Experimental results demonstrate that compared to traditional PBFT,the proposed VRPBFT algorithm reduces consensus latency by approximately 30%and decreases the proportion of Byzantine nodes by 40%after 100 rounds of consensus.Furthermore,the DRL-based SFC deployment algorithm(SDRL)exhibits rapid convergence during training,with improvements in long-term average revenue,request acceptance rate,and revenue/cost ratio of 17%,14.49%,and 20.35%,respectively,over existing algorithms.Additionally,the CPU resource utilization of the SDRL algorithmreaches up to 42%,which is 27.96%higher than other algorithms.These findings indicate that the proposed algorithm substantially enhances resource utilization efficiency,service quality,and security in SFC deployment.
基金Supported by the National Natural Science Foundation of China(Nos.42006018,42276009,42376199)the Open Fund Project of the Key Laboratory of Ocean Observation and Information of Hainan Province(No.HKLOOI-OF-2023-03)the Tianjin Natural Science Foundation(Nos.21JCYBJC00500,21JCQNJC00590)。
文摘Frequent typhoons can significantly change the temperature,nutrient availability,and phytoplankton biomass in marginal seas.The oceanic response to typhoons is usually influenced by the features of the typhoon,among which the translational speed is critically important.By using a high resolution coupled physical-biological model,we investigated the response of the Yellow and East China seas(YECS)to two typhoons at different translational speeds,Muifa in August 2011 and Bolaven in August 2012.The model well reproduced the spatial and temporal variations of temperature,chlorophyll-a concentration over the YECS.Results show that typhoons with slower translational speeds uplift more deep water,leading to a more significant oceanic response.Divergence and convergence caused nutrient fluxes in opposite directions in the surface and bottom layers.Moreover,the nutrient flux in the bottom layer was greater than that in the surface layer.These phenomena are closely related to the spatial distribution of nutrients.Further studies show that the degree of ocean response to typhoons is highly correlated with the initial conditions of physical and biological elements of the upper ocean before the typhoon,as well as with ocean structure.Pretyphoon initial conditions of oceanic physical and ecological elements,mixed layer depth,and potential energy anomalies can all alter the degree of typhoon-induced oceanic response.This study emphasizes the important roles of the translational speed of typhoons and the initial oceanic conditions in the oceanic response to typhoons.
文摘Different phenomenological equations based on plasticity, primary creep （as a viscoplastic mechanism）, secondary creep （as another viscoplastic mechanism） and different combinations of these equations are presented and used to describe the material inelastic deformation in uniaxial test. Agreement of the models with experimental results and with the theoretical concepts and physical realities is the criterion of choosing the most appropriate formulation for uniaxial test. A model is thus proposed in which plastic deformation, primary creep and secondary creep contribute to the inelastic deformation. However, it is believed that the hardening parameter is composed of plastic and primary creep parts. Accordingly, the axial plastic strain in a uniaxial test may no longer be considered as the hardening parameter. Therefore, a proportionality concept is proposed to calculate the plastic contribution of deformation.
基金This research was funded by the National Key R&D Program of China(Project:Key Technologies and Equipment for Multi-View Stereoscopic Disaster Detection and Emergency Response to Derived Disasters in Underground Spaces,2022YFC3005600)the National Natural Science Foundation of China(52378402)+2 种基金Shandong Provincial Natural Science Foundation Youth Project(ZR2022QE021 and ZR202211100077)Shandong Province Higher Education Young Innovative Team Project(2022KJ037)State Key Laboratory of Precision Blasting and Hubei Key Laboratory of Blasting Engineering,Jianghan University(PBSKL2022C03),funding from Shandong Railway Investment Holding Group Co.,Ltd.(“Key Technologies for Rapid and Intelligent Construction of Large Section High-Speed Railway Tunnels in Low Mountain and Hilly Areas”and“Intelligent Construction Trolley Equipment and Key Technologies for the Lining of Ultra-Long Open Tunnel Sections”).
文摘In underground engineering,the detection of structural cracks on tunnel surfaces stands as a pivotal task in ensuring the health and reliability of tunnel structures.However,the dim and dusty environment inherent to under-ground engineering poses considerable challenges to crack segmentation.This paper proposes a crack segmentation algorithm termed as Focused Detection for Subsurface Cracks YOLOv8(FDSC-YOLOv8)specifically designed for underground engineering structural surfaces.Firstly,to improve the extraction of multi-layer convolutional features,the fixed convolutional module is replaced with a deformable convolutional module.Secondly,the model’s receptive field is enhanced by introducing a multi-branch convolutional module,improving the extraction of shallow features for small targets.Next,the Dynamic Snake Convolution module is incorporated to enhance the extraction capability for slender and weak cracks.Finally,the Convolutional Block Attention Module(CBAM)module is employed to achieve better target determination.The FDSC-YOLOv8s algorithm’s mAP50 and mAP50-95 reach 96.5%and 66.4%,according to the testing data.
