In the field of data-driven bearing fault diagnosis,convolutional neural network(CNN)has been widely researched and applied due to its superior feature extraction and classification ability.However,the convolutional o...In the field of data-driven bearing fault diagnosis,convolutional neural network(CNN)has been widely researched and applied due to its superior feature extraction and classification ability.However,the convolutional operation could only process a local neighborhood at a time and thus lack the ability of capturing long-range dependencies.Therefore,building an efficient learning method for long-range dependencies is crucial to comprehend and express signal features considering that the vibration signals obtained in a real industrial environment always have strong instability,periodicity,and temporal correlation.This paper introduces nonlocal mean to the CNN and presents a 1D nonlocal block(1D-NLB)to extract long-range dependencies.The 1D-NLB computes the response at a position as a weighted average value of the features at all positions.Based on it,we propose a nonlocal 1D convolutional neural network(NL-1DCNN)aiming at rolling bearing fault diagnosis.Furthermore,the 1D-NLB could be simply plugged into most existing deep learning architecture to improve their fault diagnosis ability.Under multiple noise conditions,the 1D-NLB improves the performance of the CNN on the wheelset bearing data set of high-speed train and the Case Western Reserve University bearing data set.The experiment results show that the NL-1DCNN exhibits superior results compared with six state-of-the-art fault diagnosis methods.展开更多
In the field of intelligent air combat,real-time and accurate recognition of within-visual-range(WVR)maneuver actions serves as the foundational cornerstone for constructing autonomous decision-making systems.However,...In the field of intelligent air combat,real-time and accurate recognition of within-visual-range(WVR)maneuver actions serves as the foundational cornerstone for constructing autonomous decision-making systems.However,existing methods face two major challenges:traditional feature engineering suffers from insufficient effective dimensionality in the feature space due to kinematic coupling,making it difficult to distinguish essential differences between maneuvers,while end-to-end deep learning models lack controllability in implicit feature learning and fail to model high-order long-range temporal dependencies.This paper proposes a trajectory feature pre-extraction method based on a Long-range Masked Autoencoder(LMAE),incorporating three key innovations:(1)Random Fragment High-ratio Masking(RFH-Mask),which enforces the model to learn long-range temporal correlations by masking 80%of trajectory data while retaining continuous fragments;(2)Kalman Filter-Guided Objective Function(KFG-OF),integrating trajectory continuity constraints to align the feature space with kinematic principles;and(3)Two-stage Decoupled Architecture,enabling efficient and controllable feature learning through unsupervised pre-training and frozen-feature transfer.Experimental results demonstrate that LMAE significantly improves the average recognition accuracy for 20-class maneuvers compared to traditional end-to-end models,while significantly accelerating convergence speed.The contributions of this work lie in:introducing high-masking-rate autoencoders into low-informationdensity trajectory analysis,proposing a feature engineering framework with enhanced controllability and efficiency,and providing a novel technical pathway for intelligent air combat decision-making systems.展开更多
The admission control scheme is investigated for a FIFO self-similar queuing system with Quality of Service (QoS) performance guarantees. Since the self-similar queuing system performance analysis is often carried out...The admission control scheme is investigated for a FIFO self-similar queuing system with Quality of Service (QoS) performance guarantees. Since the self-similar queuing system performance analysis is often carried out under the condition of infinite buffer, it is difficult to deduce the upper boundary of buffer overflow probability. To overcome this shortcoming, a simple overflow condition is proposed, which defines a buffer overflow occurrence whenever the arrival rate exceeds the service rate. The analytic formula for the buffer overflow probability upper boundary is easily obtained under this condition. The required bandwidth upper boundary with long-range dependence input and determined overflow probability is then derived from this formula. Based on the above analytic formulas, the upper boundaries of the admission control regions for homogeneous and heterogeneous long-range dependence traffic sources are separately obtained. Finally, an effective admission control scheme for long-range dependence input is proposed. Simulation studies with real traffic have confirmed the validity of these results.展开更多
Sluggish sulfur conversion kinetics pose an ongoing challenge in lithium-sulfur batteries(LSBs).Here,we present a solution through far-reaching long-range electronic regulation(LRER)on single-atom active sites.N-doped...Sluggish sulfur conversion kinetics pose an ongoing challenge in lithium-sulfur batteries(LSBs).Here,we present a solution through far-reaching long-range electronic regulation(LRER)on single-atom active sites.N-doped carbons(Co-NC)are implanted with densely-distributed Co single atoms,and supported on Ti_(3)C_(2)T_(x)MXene substrates to assemble 3D Co-NC/MXene catalyst.MXene effectively mediates interlayer charge transfer(~0.70|e|)contrasted with popular carbon materials(~0.06|e|)to produce LRER through surrounding carbon atoms.The synergy of LRER with near-range electronic regulation(NRER)tunes electronic structures,and enhances heterostructural stability,thus provoking desirous catalytic kinetics of Co single atoms in sulfur reduction.Thereby,the Co-NC/MXene/S cathodes exhibit impressive rate performance and excellent cycling stability(only 0.015%capacity decay per cycle over 600 cycles at 4 C)in LSBs,surpassing state-of-the-art sulfur cathodes.This work reveals the importance of LRER for improved catalysis,and provides new guidance to tailor heterostructures to achieve high-efficient catalysts in various process.展开更多
Catalysts with asymmetric coordination exhibit excellent electrocatalytic activity due to changes in the active sites,which affect the arrangement of reactants and catalytic activity/selectivity.Hence,the exploration ...Catalysts with asymmetric coordination exhibit excellent electrocatalytic activity due to changes in the active sites,which affect the arrangement of reactants and catalytic activity/selectivity.Hence,the exploration of the inherent characteristics of active sites within diverse coordination environments holds great significance for the experimental design of catalytic structures.Single-atom catalysts(SACs)characterized by high coordination with four carbons(26 candidates)and low coordination with dinitrogen(27candidates)are constructed using nitrogen-doped graphdiyne derivatives(NGDY)as the substrate.Additionally,5 species of dual-atom catalysts(DACs)with coexistence of both high and low coordination sites are also developed and their nitrogen reduction reaction(NRR)activities are systematically investigated by density functional theory.The results indicate that metals with low coordination exhibit superior catalytic performance,such as Mo^(L)-NGDY(U_(L)=-0.30 V)and Nb^(L)-NGDY(U_(L)=-0.32 V).Furthermore,machine learning(ML)methods have deeply analyzed and elucidated the primary intrinsic characteristics that influence catalytic performance.These results not only unveil the underlying mechanisms behind the exceptional catalytic performance exhibited by low-coordination metal atoms,but also provide relevant and significant descriptors.More importantly,based on an investigation of the catalytic activity of a series of DACs,the“buffer and low-coordination accumulate”asymmetric coordination mechanism is proposed to unveil the long-range interactions between low and high coordination atoms.Due to this remote communication,MoNb-NGDY(U_(L)=-0.09/-0.37 V)exhibits the best NRR activity.This mechanism provides valuable insights into the origin of long-range bipartite interactions and inspires the design and synthesis of NRR catalysts with different coordination environments.展开更多
A Hamiltonian mean-field model with long-range four-body interactions is proposed.The model describes a long-range mean-field system in which N unit-mass particles move on a unit circle.Each particleθi interacts with...A Hamiltonian mean-field model with long-range four-body interactions is proposed.The model describes a long-range mean-field system in which N unit-mass particles move on a unit circle.Each particleθi interacts with any three other particles through an infinite-range cosine potential with an attractive interaction(ε>0).By applying a method that remaps the average phase of global particle pairs onto a new unit circle,and using the saddle-point technique,the partition function is solved analytically after introducing four-body interactions,yielding expressions for the free energy f and the energy per particle U.These results were further validated through numerical simulations.The results show that the system undergoes a second-order phase transition at the critical energy Uc.Specifically,the critical energy corresponds to U_(c)=0.32 when the coupling constantε=5,and U_(c)=0.63 whenε=10.Finally,we calculated the system’s largest Lyapunov exponentλand kinetic energy fluctuationsΣthrough numerical simulations.It is found that the peak of the largest Lyapunov exponentλoccurs slightly below the critical energy Uc,which is consistent with the point of maximum kinetic energy fluctuationsΣ.And there is a scaling law ofΣ/N^(1/2)∝λbetween them.展开更多
We investigate the parity-time(PT)symmetry-breaking quantum phase transition in a one-dimensional(1D)bosonic lattice featuring cavity-mediated long-range interactions and spatially staggered dissipation.By mapping the...We investigate the parity-time(PT)symmetry-breaking quantum phase transition in a one-dimensional(1D)bosonic lattice featuring cavity-mediated long-range interactions and spatially staggered dissipation.By mapping the system to an effective spin chain under the constraints of hard-core bosons and integrating the mean-field decoupling approach with biorthogonal basis formalism,we derive a self-consistency equation.Numerical simulation results validate that the derived equation quantitatively captures thePT-symmetry order parameter’s phase diagram.Our findings reveal that coherent hopping maintainsPTsymmetry through quantum fluctuations.Conversely,cavity-engineered long-range interactions,in synergy with staggered dissipation,act in opposition to drive symmetry breaking.This competitive interplay can inspire further exploration of tunable quantum phase transitions in non-Hermitian systems.展开更多
Consecutive stresses,such as initial submergence during germination followed by water deficit during the seedling stage,pose significant challenges to direct-seeded rice cultivation.By Linkage disequilibrium analysis,...Consecutive stresses,such as initial submergence during germination followed by water deficit during the seedling stage,pose significant challenges to direct-seeded rice cultivation.By Linkage disequilibrium analysis,Sub1 and Dro1(Δbp:10 Mb),as well as Sub1 and TPP7(Δbp:6 Mb)were identified to exhibit long-range linkage disequilibrium(LRLD).Meta-QTL analysis further revealed that Sub1 and TPP7 co-segregated for tolerance to submergence at the germination and seedling stages.Based on this,we hypothesized that LRLD might influence plant responses to consecutive stresses.To test this hypothesis,we developed a structured recombinant inbred line population from a cross between Bhalum 2 and Nagina 22,with alleles(Sub1 and TPP7)in linkage equilibrium.Mendelian randomization analysis validated that the parental alleles,rather than the recombinant alleles of Sub1 and TPP7,significantly influenced 13 out of 41 traits under consecutive stress conditions.Additionally,16 minor additive effect QTLs were detected between the genomic regions,spanning Sub1 and TPP7 for various traits.A single allele difference between these genomic regions enhanced crown root number,root dry weight,and specific root area by 11.45%,15.69%,and 33.15%,respectively,under flooded germination conditions.Candidate gene analysis identified WAK79 and MRLK59 as regulators of stress responses during flooded germination,recovery,and subsequent water deficit conditions.These findings highlight the critical role of parental allele combinations and genomic regions between Sub1 and TPP7 in regulating the stress responses under consecutive stresses.Favourable haplotypes derived from these alleles can be utilized to improve stress resilience in direct-seeded rice.展开更多
The high-resolution photoassociation spectrum of the ultracold cesium molecular 0＋ state below the 6S1/2 ＋ 6PI/2 limit is presented in this paper. The saturation of the photoassociation scattering probability is obs...The high-resolution photoassociation spectrum of the ultracold cesium molecular 0＋ state below the 6S1/2 ＋ 6PI/2 limit is presented in this paper. The saturation of the photoassociation scattering probability is observed from the depen dence of the trap-loss probability on the photoassociation laser intensity. The corresponding resonant line width is also demonstrated to increase linearly with increasing photoassociation laser intensity. Our experimental data have good con sistency with the theoretical saturation model of Bohn and Julienne [Bohn J L and Julienne P S 1999 Phys. Rev. A 60 1].展开更多
Objective:Large-scale CRISPR screens have identified essential genes across cancer cell lines,but links between tumor functional properties and specific dependencies require investigation to reveal the mechanisms unde...Objective:Large-scale CRISPR screens have identified essential genes across cancer cell lines,but links between tumor functional properties and specific dependencies require investigation to reveal the mechanisms underlying dependencies and broaden understanding of targeted therapy.Methods:We selected 47 breast cancer cell lines from the Cancer Cell Line Encyclopedia(CCLE)with multi-omics data including gene dependency;somatic mutations;copy number alterations;and transcriptomic,proteomic,metabolomic,and methylation data.We established a dependency marker association(DMA)analytic pipeline by using linear regression modeling to assess associations between 3,874 representative gene dependencies and multi-omics markers.Additionally,we conducted non-negative matrix factorization clustering,to stratify breast cancer cell lines according to gene dependency features,and investigated cluster-specific DMAs.Results:We interpreted valuable DMAs according to two primary aspects.First,dependencies associated with gain-of-function alterations revealed addiction to lactate transporter SLC16A3,thus suggesting a promising therapeutic target.Second,dependencies associated with loss-of-function alterations included synthetic lethality(SL),collateral SL,and prioritized metabolic SL,encompassing paralog SL(e.g.,IMPDH1 and IMPDH2),single pathway SL(e.g.,GFPT1 and UAP1),and alternative pathway SL(e.g.,GPI and PGD).DMA analysis of the two clusters with divergent dependency signatures demonstrated that cluster1 cell lines exhibited extensive metabolism with mitochondrial protein dependencies,whereas cluster2 displays enhanced cell signaling,and reliance on DNA replication and membrane organelle regulators.Conclusions:We established a DMA analysis pipeline linking the gene dependencies of breast cancer cell lines to multi-omics characteristics,thus elucidating the underpinnings of tumor dependencies and offering a valuable resource for developing novel precision treatment strategies incorporating relevant markers.展开更多
In the realm of video understanding,the demand for accurate and contextually rich video captioning has surged with the increasing volume and complexity of multimedia content.This research introduces an innovative solu...In the realm of video understanding,the demand for accurate and contextually rich video captioning has surged with the increasing volume and complexity of multimedia content.This research introduces an innovative solution for video captioning by integrating a Convolutional BiLSTM Convolutional Bidirectional Long Short-Term Memory(BiLSTM)constructed Variational Sequence-to-Sequence(CBVSS)approach.The proposed framework is adept at capturing intricate temporal dependencies within video sequences,enabling a more nuanced and contextually relevant description of dynamic scenes.However,optimizing its parameters for improved performance remains a crucial challenge.In response,in this research Golden Eagle Optimization(GEO)a metaheuristic optimization technique is used to fine-tune the Convolutional BiLSTM variational sequence-to-sequence model parameters.The application of GEO aims to enhancing the CBVSS ability to produce more exact and contextually rich video captions.The proposed attains an overall higher Recall of 59.75%and Precision of 63.78%for both datasets.Additionally,the proposed CBVSS method demonstrated superior performance across both datasets,achieving the highest METEOR(25.67)and CIDER(39.87)scores on the ActivityNet dataset,and further outperforming all compared models on the YouCook2 dataset with METEOR(28.67)and CIDER(43.02),highlighting its effectiveness in generating semantically rich and contextually accurate video captions.展开更多
Forest biological disasters(FBD) seriously impact energy flow and material cycling in forest ecosystems,while the underlying causes of FBD are complex. These disasters involve large areas and cause tremendous losses. ...Forest biological disasters(FBD) seriously impact energy flow and material cycling in forest ecosystems,while the underlying causes of FBD are complex. These disasters involve large areas and cause tremendous losses. As a result,the occurrence of FBDs in China( CFBD) threatens the country's ability to realize its strategic target of increasing both forested area(40 million ha) and forest volume(1.3 billion m^3) from 2005 to 2020. Collectively,China has officially named this effort to increase forest area and volume the "Two Increases" as national goals related to forestry. Based on Hurst index analysis from fractal theory,we analyzed the time series of the occurrence area and related data of FBDs from 1950 to 2007 to quantitatively determine the patterns of the macro occurrence of FBDs in China. Results indicates that,the time series of( CFBD) areas is fractal( self-affinity fractal dimension D = 1. 3548),the fluctuation of( CFBD) areas is positively correlated( auto-correlation coefficient C = 0. 2170),and the occurrence of the time series of( CFBD) is long-range dependent( Hurst index H =0. 6416),showing considerably strong trend of increases in FBDC area. Three different methods were further carried out on the original time series,and its two surrogate series generated by function surrogate in library t series,and function Surrogate Data in library in Wavelet software R,so as to analyze the reliability of Hurst indexes. The results showed that the Hurst indices calculated using different estimation methods were greater than 0. 5,ranging from 0. 64 to 0. 97,which indicated that the change of occurrence area data of FBDs was positively autocorrelated.The long-range dependence in forest biological disasters in China is obvious,and the spatial extent of FBDs tended to increase during this study period indicating this trend should be expected to persistent and worsen in the future.展开更多
Background:This study aims to investigate the underlying mechanisms between parental marital conflict and adolescent short video dependence by constructing a chain mediation model,focusing on the mediating roles of ex...Background:This study aims to investigate the underlying mechanisms between parental marital conflict and adolescent short video dependence by constructing a chain mediation model,focusing on the mediating roles of experiential avoidance and emotional disturbance(anxiety,depression,and stress).Methods:Conducted in January 2025,the research recruited 4125 adolescents from multiple Chinese provinces through convenience sampling;after data cleaning,3957 valid participants(1959 males,1998 females)were included.Using a cross-sectional design,measures included parental marital conflict,experiential avoidance,anxiety,depression,stress,and short video dependence.Results:Pearson correlation analysis revealed significant positive correlations among all variables.Mediation analysis using the SPSS PROCESS macro showed that parental marital conflict directly predicted short video dependence(β=0.269,p<0.001),and also significantly predicted experiential avoidance(β=0.519,p<0.001),anxiety(β=0.072,p<0.001),depression(β=0.067,p<0.001),and stress(β=0.048,p<0.05).Experiential avoidance further predicted anxiety(β=0.521,p<0.001),depression(β=0.489,p<0.001),stress(β=0.408,p<0.001),and short video dependence(β=0.244,p<0.001).While both anxiety(β=0.050,p<0.05)and depression(β=0.116,p<0.001)positively predicted short video dependence,stress did not(β=0.019,p=0.257).Overall,experiential avoidance,anxiety,depression,and stress significantly mediated the relationship between parental marital conflict and short video dependence.Conclusion:These findings confirm that parental marital conflict not only directly influences adolescent short video dependence but also operates through a chain mediation pathway involving experiential avoidance and emotional disturbance,highlighting central psychological mechanisms and providing theoretical support for integrated mental health and behavioral interventions.展开更多
The coexistence of emerging containments,such as antibiotic resistant bacteria(ARB),antibiotic-resistant genes(ARGs)and antibiotics,potentially influence elimination efficiencies in UV light-emitting diode(UV-LED)alon...The coexistence of emerging containments,such as antibiotic resistant bacteria(ARB),antibiotic-resistant genes(ARGs)and antibiotics,potentially influence elimination efficiencies in UV light-emitting diode(UV-LED)alone and UV-LED/H_(2)O_(2) system as their complex interactions.Tetracycline(TC)degradation efficiency(kF)correlated closely with its UV molar absorbance(R^(2)=0.831)in UV-LED alone system and with·OH yield(R^(2)=0.999)in UV-LED/H_(2)O_(2) system across studied wavelengths(265,280 and 310 nm).The kF values for intracellular DNA(i-ARGs)also exhibited a high correlation with UV-LED wavelengths in both systems(R^(2)=0.997-0.999).The coexistence of TC and ARB/ARGs resulted in a mutual inhibition of their degradation efficiencies due to competition for photons and·OH,along with the consequent reduction in intracellular ROS within ARB,with their degradation efficiencies exhibiting marked dependence on wavelength in both systems.Notably,the UV-LED/H_(2)O_(2) system at 265 nm effectively achieved the simultaneous removal of TC,ARB and ARGs with minimal energy consumption,and successfully fragmented ARGs.The degradation pathway of TC was analyzed,and the biotoxicity of its degradation intermediates demonstrated the environmental friendliness and safety of UV-LED/H_(2)O_(2) technology.This study elucidated the competitive interactions between antibiotics and ARB/ARGs within UV-LED/H_(2)O_(2) system,providing a promising approach for their simultaneous removal while ensuring energy efficiency.展开更多
Distributions of nuclear magnetic resonance(NMR)relaxation times provide detailed information about the water in wood.This study documents the water dynamics analysis of T_(2)and T_(1)distributions for saturated delig...Distributions of nuclear magnetic resonance(NMR)relaxation times provide detailed information about the water in wood.This study documents the water dynamics analysis of T_(2)and T_(1)distributions for saturated delignified sapwood(DSW),delignified heartwood(DHW)and lignocellulose(LC)samples at different temperatures.Results indicate that below the freezing point of bulk water,free water freezes,causing its signal to disappear from the distribution.Then,the low temperature distributions of the unfrozen bound water contain more information about its components,with DSW,DHW and LC containing two distinct states of bound water(OH bound water(B-water)and more freely bound water(C-water)).Furthermore,it was observed that within the temperature range of−3°C to−60°C,B-water in DSW,DHW and LC maintained a higher unfrozen water content(UWC)value than C-water,and the T_(1)/T_(2)ratios for B-water were consistently higher than that for C-water,indicating that B-water has a greater antifreeze capacity.T_(2)and T_(1)distributions offer different kinds of information about water components,and all peaks within the distribution have been assigned.展开更多
Accurate navigation is important for long-range rocket projectile＇s precise striking. To obtain stable and high-per- formance navigation result, a ultra-tight global positioning system/inertial navigation system （GP...Accurate navigation is important for long-range rocket projectile＇s precise striking. To obtain stable and high-per- formance navigation result, a ultra-tight global positioning system/inertial navigation system （GPS/INS） integration based nav- igation approach is proposed. The accurate short-time output of INS is used by GPS receiver to assist in acquisition of signal, and output information of INS and GPS is fused based on federated filter. Meanwhile, the improved cubature Kalman filter with strong tracking ability is chosen to serve as the local filter, and then the federated filter is enhanced based on vector sharing theory. Finally, simulation results show that the navigation accuracy with the proposed method is higher than that with traditional methods. It provides reference for long-range rocket projectile navigation.展开更多
The effects of random long-range connections （shortcuts） on the transitions of neural firing patterns in coupled Hindmarsh-Rose neurons are investigated, where each neuron is subjected to an external current. It is ...The effects of random long-range connections （shortcuts） on the transitions of neural firing patterns in coupled Hindmarsh-Rose neurons are investigated, where each neuron is subjected to an external current. It is found that, on one hand, the system can achieve the transition of neural firing patterns from the fewer-period state to the multi-period one, when the number of the added shortcuts in the neural network is greater than a threshold value, indicating the occurrence of in-transition of neural firing patterns. On the other hand, for a stronger coupling strength, we can also find the similar but reverse results by adding some proper random connections. In addition, the influences of system size and coupling strength on such transition behavior, as well as the internality between the transition degree of firing patterns and its critical characteristics for different external stimulation current, are also discussed.展开更多
This study examines the time and regime dependencies of sensitive areas identified by the conditional nonlinear optiflml perturbation （CNOP） method for forecasts of two typhoons. Typhoon Meari （2004） was weakly no...This study examines the time and regime dependencies of sensitive areas identified by the conditional nonlinear optiflml perturbation （CNOP） method for forecasts of two typhoons. Typhoon Meari （2004） was weakly nonlinear and is herein referred to as the linear case, while Typhoon Matsa （2005） was strongly nonlinear and is herein referred to as the nonlinear case. In the linear case, the sensitive areas identified for special forecast times when the initial time was fixed resembled those identified for other forecast times. Targeted observations deployed to improve a special time forecast would thus also benefit forecasts at other times. In the nonlinear case, the similarities among the sensitive areas identified for different forecast times were more limited. The deployment of targeted observations in the nonlinear case would therefore need to be adapted to achieve large improvements for different targeted forecasts. For both cases, the closer the forecast time, the higher the similarities of the sensitive areas. When the forecast time was fixed, the sensitive areas in the linear case diverged continuously from the verification area as the forecast period lengthened, while those in the nonlinear case were always located around the initial cyclones. The deployment of targeted observations to improve a special forecast depends strongly on the time of deployment. An examination of the efficiency gained by reducing initial errors within the identified sensitive areas confirmed these results. In general, the greatest improvement in a special time forecast was obtained by identifying the sensitive areas for the corresponding forecast time period.展开更多
This study proposes a novel general image fusion framework based on cross-domain long-range learning and Swin Transformer,termed as SwinFusion.On the one hand,an attention-guided cross-domain module is devised to achi...This study proposes a novel general image fusion framework based on cross-domain long-range learning and Swin Transformer,termed as SwinFusion.On the one hand,an attention-guided cross-domain module is devised to achieve sufficient integration of complementary information and global interaction.More specifically,the proposed method involves an intra-domain fusion unit based on self-attention and an interdomain fusion unit based on cross-attention,which mine and integrate long dependencies within the same domain and across domains.Through long-range dependency modeling,the network is able to fully implement domain-specific information extraction and cross-domain complementary information integration as well as maintaining the appropriate apparent intensity from a global perspective.In particular,we introduce the shifted windows mechanism into the self-attention and cross-attention,which allows our model to receive images with arbitrary sizes.On the other hand,the multi-scene image fusion problems are generalized to a unified framework with structure maintenance,detail preservation,and proper intensity control.Moreover,an elaborate loss function,consisting of SSIM loss,texture loss,and intensity loss,drives the network to preserve abundant texture details and structural information,as well as presenting optimal apparent intensity.Extensive experiments on both multi-modal image fusion and digital photography image fusion demonstrate the superiority of our SwinFusion compared to the state-of-theart unified image fusion algorithms and task-specific alternatives.Implementation code and pre-trained weights can be accessed at https://github.com/Linfeng-Tang/SwinFusion.展开更多
基金supported by the State Key Laboratory of Traction Power,Southwest Jiaotong University (TPL2104)the National Natural Science Foundation of China (61833002).
文摘In the field of data-driven bearing fault diagnosis,convolutional neural network(CNN)has been widely researched and applied due to its superior feature extraction and classification ability.However,the convolutional operation could only process a local neighborhood at a time and thus lack the ability of capturing long-range dependencies.Therefore,building an efficient learning method for long-range dependencies is crucial to comprehend and express signal features considering that the vibration signals obtained in a real industrial environment always have strong instability,periodicity,and temporal correlation.This paper introduces nonlocal mean to the CNN and presents a 1D nonlocal block(1D-NLB)to extract long-range dependencies.The 1D-NLB computes the response at a position as a weighted average value of the features at all positions.Based on it,we propose a nonlocal 1D convolutional neural network(NL-1DCNN)aiming at rolling bearing fault diagnosis.Furthermore,the 1D-NLB could be simply plugged into most existing deep learning architecture to improve their fault diagnosis ability.Under multiple noise conditions,the 1D-NLB improves the performance of the CNN on the wheelset bearing data set of high-speed train and the Case Western Reserve University bearing data set.The experiment results show that the NL-1DCNN exhibits superior results compared with six state-of-the-art fault diagnosis methods.
文摘In the field of intelligent air combat,real-time and accurate recognition of within-visual-range(WVR)maneuver actions serves as the foundational cornerstone for constructing autonomous decision-making systems.However,existing methods face two major challenges:traditional feature engineering suffers from insufficient effective dimensionality in the feature space due to kinematic coupling,making it difficult to distinguish essential differences between maneuvers,while end-to-end deep learning models lack controllability in implicit feature learning and fail to model high-order long-range temporal dependencies.This paper proposes a trajectory feature pre-extraction method based on a Long-range Masked Autoencoder(LMAE),incorporating three key innovations:(1)Random Fragment High-ratio Masking(RFH-Mask),which enforces the model to learn long-range temporal correlations by masking 80%of trajectory data while retaining continuous fragments;(2)Kalman Filter-Guided Objective Function(KFG-OF),integrating trajectory continuity constraints to align the feature space with kinematic principles;and(3)Two-stage Decoupled Architecture,enabling efficient and controllable feature learning through unsupervised pre-training and frozen-feature transfer.Experimental results demonstrate that LMAE significantly improves the average recognition accuracy for 20-class maneuvers compared to traditional end-to-end models,while significantly accelerating convergence speed.The contributions of this work lie in:introducing high-masking-rate autoencoders into low-informationdensity trajectory analysis,proposing a feature engineering framework with enhanced controllability and efficiency,and providing a novel technical pathway for intelligent air combat decision-making systems.
文摘The admission control scheme is investigated for a FIFO self-similar queuing system with Quality of Service (QoS) performance guarantees. Since the self-similar queuing system performance analysis is often carried out under the condition of infinite buffer, it is difficult to deduce the upper boundary of buffer overflow probability. To overcome this shortcoming, a simple overflow condition is proposed, which defines a buffer overflow occurrence whenever the arrival rate exceeds the service rate. The analytic formula for the buffer overflow probability upper boundary is easily obtained under this condition. The required bandwidth upper boundary with long-range dependence input and determined overflow probability is then derived from this formula. Based on the above analytic formulas, the upper boundaries of the admission control regions for homogeneous and heterogeneous long-range dependence traffic sources are separately obtained. Finally, an effective admission control scheme for long-range dependence input is proposed. Simulation studies with real traffic have confirmed the validity of these results.
基金supported by the National Natural Science Foundation of China(Nos.21573059,12274118 and 22208088)Henan Center for Outstanding Overseas Scientists(No.GZS2023007)Special Project for Fundamental Research in University of Henan Province(No.22ZX013)。
文摘Sluggish sulfur conversion kinetics pose an ongoing challenge in lithium-sulfur batteries(LSBs).Here,we present a solution through far-reaching long-range electronic regulation(LRER)on single-atom active sites.N-doped carbons(Co-NC)are implanted with densely-distributed Co single atoms,and supported on Ti_(3)C_(2)T_(x)MXene substrates to assemble 3D Co-NC/MXene catalyst.MXene effectively mediates interlayer charge transfer(~0.70|e|)contrasted with popular carbon materials(~0.06|e|)to produce LRER through surrounding carbon atoms.The synergy of LRER with near-range electronic regulation(NRER)tunes electronic structures,and enhances heterostructural stability,thus provoking desirous catalytic kinetics of Co single atoms in sulfur reduction.Thereby,the Co-NC/MXene/S cathodes exhibit impressive rate performance and excellent cycling stability(only 0.015%capacity decay per cycle over 600 cycles at 4 C)in LSBs,surpassing state-of-the-art sulfur cathodes.This work reveals the importance of LRER for improved catalysis,and provides new guidance to tailor heterostructures to achieve high-efficient catalysts in various process.
基金supports by the National Natural Science Foundation of China(NSFC,Grant No.52271113)the Natural Science Foundation of Shaanxi Province,China(2020JM 218)the Fundamental Research Funds for the Central Universities(CHD300102311405)。
文摘Catalysts with asymmetric coordination exhibit excellent electrocatalytic activity due to changes in the active sites,which affect the arrangement of reactants and catalytic activity/selectivity.Hence,the exploration of the inherent characteristics of active sites within diverse coordination environments holds great significance for the experimental design of catalytic structures.Single-atom catalysts(SACs)characterized by high coordination with four carbons(26 candidates)and low coordination with dinitrogen(27candidates)are constructed using nitrogen-doped graphdiyne derivatives(NGDY)as the substrate.Additionally,5 species of dual-atom catalysts(DACs)with coexistence of both high and low coordination sites are also developed and their nitrogen reduction reaction(NRR)activities are systematically investigated by density functional theory.The results indicate that metals with low coordination exhibit superior catalytic performance,such as Mo^(L)-NGDY(U_(L)=-0.30 V)and Nb^(L)-NGDY(U_(L)=-0.32 V).Furthermore,machine learning(ML)methods have deeply analyzed and elucidated the primary intrinsic characteristics that influence catalytic performance.These results not only unveil the underlying mechanisms behind the exceptional catalytic performance exhibited by low-coordination metal atoms,but also provide relevant and significant descriptors.More importantly,based on an investigation of the catalytic activity of a series of DACs,the“buffer and low-coordination accumulate”asymmetric coordination mechanism is proposed to unveil the long-range interactions between low and high coordination atoms.Due to this remote communication,MoNb-NGDY(U_(L)=-0.09/-0.37 V)exhibits the best NRR activity.This mechanism provides valuable insights into the origin of long-range bipartite interactions and inspires the design and synthesis of NRR catalysts with different coordination environments.
基金supported by the National Natural Science Foundation of China(Grant No.11962002)the Innovation Project of the Guangxi Graduate Education(Grant Nos.YCBZ2021021 and YCSW2022070).
文摘A Hamiltonian mean-field model with long-range four-body interactions is proposed.The model describes a long-range mean-field system in which N unit-mass particles move on a unit circle.Each particleθi interacts with any three other particles through an infinite-range cosine potential with an attractive interaction(ε>0).By applying a method that remaps the average phase of global particle pairs onto a new unit circle,and using the saddle-point technique,the partition function is solved analytically after introducing four-body interactions,yielding expressions for the free energy f and the energy per particle U.These results were further validated through numerical simulations.The results show that the system undergoes a second-order phase transition at the critical energy Uc.Specifically,the critical energy corresponds to U_(c)=0.32 when the coupling constantε=5,and U_(c)=0.63 whenε=10.Finally,we calculated the system’s largest Lyapunov exponentλand kinetic energy fluctuationsΣthrough numerical simulations.It is found that the peak of the largest Lyapunov exponentλoccurs slightly below the critical energy Uc,which is consistent with the point of maximum kinetic energy fluctuationsΣ.And there is a scaling law ofΣ/N^(1/2)∝λbetween them.
基金supported by the National Natural Science Foundation of China(Grant No.12375025).
文摘We investigate the parity-time(PT)symmetry-breaking quantum phase transition in a one-dimensional(1D)bosonic lattice featuring cavity-mediated long-range interactions and spatially staggered dissipation.By mapping the system to an effective spin chain under the constraints of hard-core bosons and integrating the mean-field decoupling approach with biorthogonal basis formalism,we derive a self-consistency equation.Numerical simulation results validate that the derived equation quantitatively captures thePT-symmetry order parameter’s phase diagram.Our findings reveal that coherent hopping maintainsPTsymmetry through quantum fluctuations.Conversely,cavity-engineered long-range interactions,in synergy with staggered dissipation,act in opposition to drive symmetry breaking.This competitive interplay can inspire further exploration of tunable quantum phase transitions in non-Hermitian systems.
基金supported by the Director General,Indian Council of Agricultural Research(ICAR),New Delhithe Director,ICAR-National Rice Research Institute,Cuttack.
文摘Consecutive stresses,such as initial submergence during germination followed by water deficit during the seedling stage,pose significant challenges to direct-seeded rice cultivation.By Linkage disequilibrium analysis,Sub1 and Dro1(Δbp:10 Mb),as well as Sub1 and TPP7(Δbp:6 Mb)were identified to exhibit long-range linkage disequilibrium(LRLD).Meta-QTL analysis further revealed that Sub1 and TPP7 co-segregated for tolerance to submergence at the germination and seedling stages.Based on this,we hypothesized that LRLD might influence plant responses to consecutive stresses.To test this hypothesis,we developed a structured recombinant inbred line population from a cross between Bhalum 2 and Nagina 22,with alleles(Sub1 and TPP7)in linkage equilibrium.Mendelian randomization analysis validated that the parental alleles,rather than the recombinant alleles of Sub1 and TPP7,significantly influenced 13 out of 41 traits under consecutive stress conditions.Additionally,16 minor additive effect QTLs were detected between the genomic regions,spanning Sub1 and TPP7 for various traits.A single allele difference between these genomic regions enhanced crown root number,root dry weight,and specific root area by 11.45%,15.69%,and 33.15%,respectively,under flooded germination conditions.Candidate gene analysis identified WAK79 and MRLK59 as regulators of stress responses during flooded germination,recovery,and subsequent water deficit conditions.These findings highlight the critical role of parental allele combinations and genomic regions between Sub1 and TPP7 in regulating the stress responses under consecutive stresses.Favourable haplotypes derived from these alleles can be utilized to improve stress resilience in direct-seeded rice.
基金supported by the National Basic Research Program of China(Grant No.2012CB921603)the 863 Program(Grant No.2011AA010801)+4 种基金the National Natural Science Foundation of China(Grant Nos.61008012,10934004,60978001,60978018,and 11174187)the International Science&Technology Cooperation Program of China(Grant No.2011DFA12490)the National Natural Science Foundation of China for Excellent Research Team(Grant No.61121064)the Natural Science Foundation of Shanxi Province,China(Grant Nos.2011011004 and 2011081030)the New Teacher Fund of the Ministry of Education of China(Grant No.20101401120004)
文摘The high-resolution photoassociation spectrum of the ultracold cesium molecular 0＋ state below the 6S1/2 ＋ 6PI/2 limit is presented in this paper. The saturation of the photoassociation scattering probability is observed from the depen dence of the trap-loss probability on the photoassociation laser intensity. The corresponding resonant line width is also demonstrated to increase linearly with increasing photoassociation laser intensity. Our experimental data have good con sistency with the theoretical saturation model of Bohn and Julienne [Bohn J L and Julienne P S 1999 Phys. Rev. A 60 1].
基金supported by grants from the National Key Research and Development Project of China(Grant No.2020YFA0112304)the National Natural Science Foundation of China(Grant Nos.91959207 and 82202883).
文摘Objective:Large-scale CRISPR screens have identified essential genes across cancer cell lines,but links between tumor functional properties and specific dependencies require investigation to reveal the mechanisms underlying dependencies and broaden understanding of targeted therapy.Methods:We selected 47 breast cancer cell lines from the Cancer Cell Line Encyclopedia(CCLE)with multi-omics data including gene dependency;somatic mutations;copy number alterations;and transcriptomic,proteomic,metabolomic,and methylation data.We established a dependency marker association(DMA)analytic pipeline by using linear regression modeling to assess associations between 3,874 representative gene dependencies and multi-omics markers.Additionally,we conducted non-negative matrix factorization clustering,to stratify breast cancer cell lines according to gene dependency features,and investigated cluster-specific DMAs.Results:We interpreted valuable DMAs according to two primary aspects.First,dependencies associated with gain-of-function alterations revealed addiction to lactate transporter SLC16A3,thus suggesting a promising therapeutic target.Second,dependencies associated with loss-of-function alterations included synthetic lethality(SL),collateral SL,and prioritized metabolic SL,encompassing paralog SL(e.g.,IMPDH1 and IMPDH2),single pathway SL(e.g.,GFPT1 and UAP1),and alternative pathway SL(e.g.,GPI and PGD).DMA analysis of the two clusters with divergent dependency signatures demonstrated that cluster1 cell lines exhibited extensive metabolism with mitochondrial protein dependencies,whereas cluster2 displays enhanced cell signaling,and reliance on DNA replication and membrane organelle regulators.Conclusions:We established a DMA analysis pipeline linking the gene dependencies of breast cancer cell lines to multi-omics characteristics,thus elucidating the underpinnings of tumor dependencies and offering a valuable resource for developing novel precision treatment strategies incorporating relevant markers.
文摘In the realm of video understanding,the demand for accurate and contextually rich video captioning has surged with the increasing volume and complexity of multimedia content.This research introduces an innovative solution for video captioning by integrating a Convolutional BiLSTM Convolutional Bidirectional Long Short-Term Memory(BiLSTM)constructed Variational Sequence-to-Sequence(CBVSS)approach.The proposed framework is adept at capturing intricate temporal dependencies within video sequences,enabling a more nuanced and contextually relevant description of dynamic scenes.However,optimizing its parameters for improved performance remains a crucial challenge.In response,in this research Golden Eagle Optimization(GEO)a metaheuristic optimization technique is used to fine-tune the Convolutional BiLSTM variational sequence-to-sequence model parameters.The application of GEO aims to enhancing the CBVSS ability to produce more exact and contextually rich video captions.The proposed attains an overall higher Recall of 59.75%and Precision of 63.78%for both datasets.Additionally,the proposed CBVSS method demonstrated superior performance across both datasets,achieving the highest METEOR(25.67)and CIDER(39.87)scores on the ActivityNet dataset,and further outperforming all compared models on the YouCook2 dataset with METEOR(28.67)and CIDER(43.02),highlighting its effectiveness in generating semantically rich and contextually accurate video captions.
基金Supported by the Project "Researches of Southern China’s Forestry Strategy"(2013-R17) and "Improvement of the Forest Resources Monitoring System of China"(2011-R03) Funded by the State Forestry Administration of China
文摘Forest biological disasters(FBD) seriously impact energy flow and material cycling in forest ecosystems,while the underlying causes of FBD are complex. These disasters involve large areas and cause tremendous losses. As a result,the occurrence of FBDs in China( CFBD) threatens the country's ability to realize its strategic target of increasing both forested area(40 million ha) and forest volume(1.3 billion m^3) from 2005 to 2020. Collectively,China has officially named this effort to increase forest area and volume the "Two Increases" as national goals related to forestry. Based on Hurst index analysis from fractal theory,we analyzed the time series of the occurrence area and related data of FBDs from 1950 to 2007 to quantitatively determine the patterns of the macro occurrence of FBDs in China. Results indicates that,the time series of( CFBD) areas is fractal( self-affinity fractal dimension D = 1. 3548),the fluctuation of( CFBD) areas is positively correlated( auto-correlation coefficient C = 0. 2170),and the occurrence of the time series of( CFBD) is long-range dependent( Hurst index H =0. 6416),showing considerably strong trend of increases in FBDC area. Three different methods were further carried out on the original time series,and its two surrogate series generated by function surrogate in library t series,and function Surrogate Data in library in Wavelet software R,so as to analyze the reliability of Hurst indexes. The results showed that the Hurst indices calculated using different estimation methods were greater than 0. 5,ranging from 0. 64 to 0. 97,which indicated that the change of occurrence area data of FBDs was positively autocorrelated.The long-range dependence in forest biological disasters in China is obvious,and the spatial extent of FBDs tended to increase during this study period indicating this trend should be expected to persistent and worsen in the future.
文摘Background:This study aims to investigate the underlying mechanisms between parental marital conflict and adolescent short video dependence by constructing a chain mediation model,focusing on the mediating roles of experiential avoidance and emotional disturbance(anxiety,depression,and stress).Methods:Conducted in January 2025,the research recruited 4125 adolescents from multiple Chinese provinces through convenience sampling;after data cleaning,3957 valid participants(1959 males,1998 females)were included.Using a cross-sectional design,measures included parental marital conflict,experiential avoidance,anxiety,depression,stress,and short video dependence.Results:Pearson correlation analysis revealed significant positive correlations among all variables.Mediation analysis using the SPSS PROCESS macro showed that parental marital conflict directly predicted short video dependence(β=0.269,p<0.001),and also significantly predicted experiential avoidance(β=0.519,p<0.001),anxiety(β=0.072,p<0.001),depression(β=0.067,p<0.001),and stress(β=0.048,p<0.05).Experiential avoidance further predicted anxiety(β=0.521,p<0.001),depression(β=0.489,p<0.001),stress(β=0.408,p<0.001),and short video dependence(β=0.244,p<0.001).While both anxiety(β=0.050,p<0.05)and depression(β=0.116,p<0.001)positively predicted short video dependence,stress did not(β=0.019,p=0.257).Overall,experiential avoidance,anxiety,depression,and stress significantly mediated the relationship between parental marital conflict and short video dependence.Conclusion:These findings confirm that parental marital conflict not only directly influences adolescent short video dependence but also operates through a chain mediation pathway involving experiential avoidance and emotional disturbance,highlighting central psychological mechanisms and providing theoretical support for integrated mental health and behavioral interventions.
基金supported by Major Scientific and Technological Innovation Project of Shandong Province(No.2020CXGC011204)Qingdao Natural Science Foundation(No.23-2-1-234-zyyd-jch).
文摘The coexistence of emerging containments,such as antibiotic resistant bacteria(ARB),antibiotic-resistant genes(ARGs)and antibiotics,potentially influence elimination efficiencies in UV light-emitting diode(UV-LED)alone and UV-LED/H_(2)O_(2) system as their complex interactions.Tetracycline(TC)degradation efficiency(kF)correlated closely with its UV molar absorbance(R^(2)=0.831)in UV-LED alone system and with·OH yield(R^(2)=0.999)in UV-LED/H_(2)O_(2) system across studied wavelengths(265,280 and 310 nm).The kF values for intracellular DNA(i-ARGs)also exhibited a high correlation with UV-LED wavelengths in both systems(R^(2)=0.997-0.999).The coexistence of TC and ARB/ARGs resulted in a mutual inhibition of their degradation efficiencies due to competition for photons and·OH,along with the consequent reduction in intracellular ROS within ARB,with their degradation efficiencies exhibiting marked dependence on wavelength in both systems.Notably,the UV-LED/H_(2)O_(2) system at 265 nm effectively achieved the simultaneous removal of TC,ARB and ARGs with minimal energy consumption,and successfully fragmented ARGs.The degradation pathway of TC was analyzed,and the biotoxicity of its degradation intermediates demonstrated the environmental friendliness and safety of UV-LED/H_(2)O_(2) technology.This study elucidated the competitive interactions between antibiotics and ARB/ARGs within UV-LED/H_(2)O_(2) system,providing a promising approach for their simultaneous removal while ensuring energy efficiency.
基金supported by Natural Science Foundation of Inner Mongolia Autonomous Region of China (2023MS03027)the National Natural Science Foundation of China (31860185 and 31160141)
文摘Distributions of nuclear magnetic resonance(NMR)relaxation times provide detailed information about the water in wood.This study documents the water dynamics analysis of T_(2)and T_(1)distributions for saturated delignified sapwood(DSW),delignified heartwood(DHW)and lignocellulose(LC)samples at different temperatures.Results indicate that below the freezing point of bulk water,free water freezes,causing its signal to disappear from the distribution.Then,the low temperature distributions of the unfrozen bound water contain more information about its components,with DSW,DHW and LC containing two distinct states of bound water(OH bound water(B-water)and more freely bound water(C-water)).Furthermore,it was observed that within the temperature range of−3°C to−60°C,B-water in DSW,DHW and LC maintained a higher unfrozen water content(UWC)value than C-water,and the T_(1)/T_(2)ratios for B-water were consistently higher than that for C-water,indicating that B-water has a greater antifreeze capacity.T_(2)and T_(1)distributions offer different kinds of information about water components,and all peaks within the distribution have been assigned.
基金Project Funded by Chongqing Changjiang Electrical Appliances Industries Group Co.,Ltd
文摘Accurate navigation is important for long-range rocket projectile＇s precise striking. To obtain stable and high-per- formance navigation result, a ultra-tight global positioning system/inertial navigation system （GPS/INS） integration based nav- igation approach is proposed. The accurate short-time output of INS is used by GPS receiver to assist in acquisition of signal, and output information of INS and GPS is fused based on federated filter. Meanwhile, the improved cubature Kalman filter with strong tracking ability is chosen to serve as the local filter, and then the federated filter is enhanced based on vector sharing theory. Finally, simulation results show that the navigation accuracy with the proposed method is higher than that with traditional methods. It provides reference for long-range rocket projectile navigation.
文摘The effects of random long-range connections （shortcuts） on the transitions of neural firing patterns in coupled Hindmarsh-Rose neurons are investigated, where each neuron is subjected to an external current. It is found that, on one hand, the system can achieve the transition of neural firing patterns from the fewer-period state to the multi-period one, when the number of the added shortcuts in the neural network is greater than a threshold value, indicating the occurrence of in-transition of neural firing patterns. On the other hand, for a stronger coupling strength, we can also find the similar but reverse results by adding some proper random connections. In addition, the influences of system size and coupling strength on such transition behavior, as well as the internality between the transition degree of firing patterns and its critical characteristics for different external stimulation current, are also discussed.
基金supported by the National Natural Science Foundation of China(Grant Nos.41105038and40830955)the NationalKey Technology R&D Program(Grant No.2012BAC22B03)
文摘This study examines the time and regime dependencies of sensitive areas identified by the conditional nonlinear optiflml perturbation （CNOP） method for forecasts of two typhoons. Typhoon Meari （2004） was weakly nonlinear and is herein referred to as the linear case, while Typhoon Matsa （2005） was strongly nonlinear and is herein referred to as the nonlinear case. In the linear case, the sensitive areas identified for special forecast times when the initial time was fixed resembled those identified for other forecast times. Targeted observations deployed to improve a special time forecast would thus also benefit forecasts at other times. In the nonlinear case, the similarities among the sensitive areas identified for different forecast times were more limited. The deployment of targeted observations in the nonlinear case would therefore need to be adapted to achieve large improvements for different targeted forecasts. For both cases, the closer the forecast time, the higher the similarities of the sensitive areas. When the forecast time was fixed, the sensitive areas in the linear case diverged continuously from the verification area as the forecast period lengthened, while those in the nonlinear case were always located around the initial cyclones. The deployment of targeted observations to improve a special forecast depends strongly on the time of deployment. An examination of the efficiency gained by reducing initial errors within the identified sensitive areas confirmed these results. In general, the greatest improvement in a special time forecast was obtained by identifying the sensitive areas for the corresponding forecast time period.
基金This work was supported by the National Natural Science Foundation of China(62075169,62003247,62061160370)the Key Research and Development Program of Hubei Province(2020BAB113).
文摘This study proposes a novel general image fusion framework based on cross-domain long-range learning and Swin Transformer,termed as SwinFusion.On the one hand,an attention-guided cross-domain module is devised to achieve sufficient integration of complementary information and global interaction.More specifically,the proposed method involves an intra-domain fusion unit based on self-attention and an interdomain fusion unit based on cross-attention,which mine and integrate long dependencies within the same domain and across domains.Through long-range dependency modeling,the network is able to fully implement domain-specific information extraction and cross-domain complementary information integration as well as maintaining the appropriate apparent intensity from a global perspective.In particular,we introduce the shifted windows mechanism into the self-attention and cross-attention,which allows our model to receive images with arbitrary sizes.On the other hand,the multi-scene image fusion problems are generalized to a unified framework with structure maintenance,detail preservation,and proper intensity control.Moreover,an elaborate loss function,consisting of SSIM loss,texture loss,and intensity loss,drives the network to preserve abundant texture details and structural information,as well as presenting optimal apparent intensity.Extensive experiments on both multi-modal image fusion and digital photography image fusion demonstrate the superiority of our SwinFusion compared to the state-of-theart unified image fusion algorithms and task-specific alternatives.Implementation code and pre-trained weights can be accessed at https://github.com/Linfeng-Tang/SwinFusion.
