YOLO-MFD:Remote Sensing Image Object Detection with Multi-Scale Fusion Dynamic Head

Remote sensing imagery,due to its high altitude,presents inherent challenges characterized by multiple scales,limited target areas,and intricate backgrounds.These inherent traits often lead to increased miss and false detection rates when applying object recognition algorithms tailored for remote sensing imagery.Additionally,these complexities contribute to inaccuracies in target localization and hinder precise target categorization.This paper addresses these challenges by proposing a solution:The YOLO-MFD model(YOLO-MFD:Remote Sensing Image Object Detection withMulti-scale Fusion Dynamic Head).Before presenting our method,we delve into the prevalent issues faced in remote sensing imagery analysis.Specifically,we emphasize the struggles of existing object recognition algorithms in comprehensively capturing critical image features amidst varying scales and complex backgrounds.To resolve these issues,we introduce a novel approach.First,we propose the implementation of a lightweight multi-scale module called CEF.This module significantly improves the model’s ability to comprehensively capture important image features by merging multi-scale feature information.It effectively addresses the issues of missed detection and mistaken alarms that are common in remote sensing imagery.Second,an additional layer of small target detection heads is added,and a residual link is established with the higher-level feature extraction module in the backbone section.This allows the model to incorporate shallower information,significantly improving the accuracy of target localization in remotely sensed images.Finally,a dynamic head attentionmechanism is introduced.This allows themodel to exhibit greater flexibility and accuracy in recognizing shapes and targets of different sizes.Consequently,the precision of object detection is significantly improved.The trial results show that the YOLO-MFD model shows improvements of 6.3%,3.5%,and 2.5%over the original YOLOv8 model in Precision,map@0.5 and map@0.5:0.95,separately.These results illustrate the clear advantages of the method.

Research of caged dynamics of clusters center atoms in Pd_(82)Si_(18) amorphous alloy

To date,there is still a lack of a comprehensive explanation for caged dynamics which is regarded as one of the intricate dynamic behaviors in amorphous alloys.This study focuses on Pd_(82)Si_(18)as the research object to further elucidate the underlying mechanism of caged dynamics from multiple perspectives,including the cage's lifetime,atomic local environment,and atomic potential energy.The results reveal that Si atoms exhibit a pronounced cage effect due to the hindrance of Pd atoms,resulting in an anomalous peak in the non-Gaussian parameters.An in-depth investigation was conducted on the caged dynamics differences between fast and slow Si atoms.In comparison to fast Si atoms,slow Si atoms were surrounded by more Pd atoms and occupied lower potential energy states,resulting in smaller diffusion displacements for the slow Si atoms.Concurrently,slow Si atoms tend to be in the centers of smaller clusters with coordination numbers of 9 and 10.During the isothermal relaxation process,clusters with coordination numbers 9 and 10 have longer lifetimes,suggesting that the escape of slow Si atoms from their cages is more challenging.The findings mentioned above hold significant implications for understanding the caged dynamics.

Simulating the Dynamics of Bimetallic Clusters Deposited onto a Surface Using Molecular Dynamics

This paper examines the interface development between a single crystalline Ag matrix and core-shell AgnCom nanoclusters that have been deposited with energies varying between 0.25 eV and 1.5 eV per atom using computer modeling techniques. Clusters undergo deformation as a result of the slowing down;they may also become epitaxial with the substrate and maintain their core-shell structure. A detailed analysis of the effects of the cluster-surface interaction is conducted over a realistic size and energy range, and a model is created to show how clusters accumulate. It is discovered that both the silver shells and the cobalt cluster cores exhibit limited epitaxy with the substrate, and that the contact produced is only a few atomic layers thick. The effect is higher for Ag shells than for Co cores, and it is not very energy dependent.

Dynamic response of multi-scale structure in flexible pavement to moving load

In order to study the dynamic responses in the microstructures of the pavement structure, the multi-scale modeling subjected to moving load is analyzed using the discrete element method （DEM）. The macro-scale discrete element model of the flexible pavement structure is established. The stress and strain at the bottom of the asphalt concrete layer under moving load are calculated. The DEM model is validated through comparison between DEM predictions and the results from the classical program. Based on the validated macro-scale DEM model, the distribution and the volumetric fraction of coarse aggregate, mastics and air voids at the bottom of the asphalt layer are modeled, and then the multi-scale model is constructed. The dynamic response in the microstructures of the multi-scale model are calculated and compared with the results from the macro model. The influence of mastic stiffness on the distribution of dynamic response in the microstructures is also analyzed. Results show that the average values and the variation coefficient of the tensile stress at the aggregate-mastic interface are far more than those within the mastics. The dynamic response including stress and strain distributes non-uniformly in both mastics and the interface. An increase in mastic stiffness tends to a uniform distribution of tensile stress in asphalt concrete.

Multi-scale physics-informed neural networks for solving high Reynolds number boundary layer flows based on matched asymptotic expansions

Multi-scale system remains a classical scientific problem in fluid dynamics,biology,etc.In the present study,a scheme of multi-scale Physics-informed neural networks is proposed to solve the boundary layer flow at high Reynolds numbers without any data.The flow is divided into several regions with different scales based on Prandtl's boundary theory.Different regions are solved with governing equations in different scales.The method of matched asymptotic expansions is used to make the flow field continuously.A flow on a semi infinite flat plate at a high Reynolds number is considered a multi-scale problem because the boundary layer scale is much smaller than the outer flow scale.The results are compared with the reference numerical solutions,which show that the msPINNs can solve the multi-scale problem of the boundary layer in high Reynolds number flows.This scheme can be developed for more multi-scale problems in the future.

UAV-assisted data collection for wireless sensor networks with dynamic working modes

Wireless Sensor Network(WSN)is a cornerstone of Internet of Things(IoT)and has rich application scenarios.In this work,we consider a heterogeneous WSN whose sensor nodes have a diversity in their Residual Energy(RE).In this work,to protect the sensor nodes with low RE,we investigate dynamic working modes for sensor nodes which are determined by their RE and an introduced energy threshold.Besides,we employ an Unmanned Aerial Vehicle(UAV)to collect the stored data from the heterogeneous WSN.We aim to jointly optimize the cluster head selection,energy threshold and sensor nodes’working mode to minimize the weighted sum of energy con-sumption from the WSN and UAV,subject to the data collection rate constraint.To this end,we propose an efficient search method to search for an optimal energy threshold,and develop a penalty-based successive convex approximation algorithm to select the cluster heads.Then we present a low-complexity iterative approach to solve the joint optimization problem and discuss the implementation procedure.Numerical results justify that our proposed approach is able to reduce the energy consumption of the sensor nodes with low RE significantly and also saves energy for the whole WSN.

Two Stages Segmentation Algorithm of Breast Tumor in DCE-MRI Based on Multi-Scale Feature and Boundary Attention Mechanism

Nuclearmagnetic resonance imaging of breasts often presents complex backgrounds.Breast tumors exhibit varying sizes,uneven intensity,and indistinct boundaries.These characteristics can lead to challenges such as low accuracy and incorrect segmentation during tumor segmentation.Thus,we propose a two-stage breast tumor segmentation method leveraging multi-scale features and boundary attention mechanisms.Initially,the breast region of interest is extracted to isolate the breast area from surrounding tissues and organs.Subsequently,we devise a fusion network incorporatingmulti-scale features and boundary attentionmechanisms for breast tumor segmentation.We incorporate multi-scale parallel dilated convolution modules into the network,enhancing its capability to segment tumors of various sizes through multi-scale convolution and novel fusion techniques.Additionally,attention and boundary detection modules are included to augment the network’s capacity to locate tumors by capturing nonlocal dependencies in both spatial and channel domains.Furthermore,a hybrid loss function with boundary weight is employed to address sample class imbalance issues and enhance the network’s boundary maintenance capability through additional loss.Themethod was evaluated using breast data from 207 patients at RuijinHospital,resulting in a 6.64%increase in Dice similarity coefficient compared to the benchmarkU-Net.Experimental results demonstrate the superiority of the method over other segmentation techniques,with fewer model parameters.

An algorithm for segmentation of lung ROI by mean-shift clustering combined with multi-scale HESSIAN matrix dot filtering

A new algorithm for segmentation of suspected lung ROI(regions of interest)by mean-shift clustering and multi-scale HESSIAN matrix dot filtering was proposed.Original image was firstly filtered by multi-scale HESSIAN matrix dot filters,round suspected nodular lesions in the image were enhanced,and linear shape regions of the trachea and vascular were suppressed.Then,three types of information,such as,shape filtering value of HESSIAN matrix,gray value,and spatial location,were introduced to feature space.The kernel function of mean-shift clustering was divided into product form of three kinds of kernel functions corresponding to the three feature information.Finally,bandwidths were calculated adaptively to determine the bandwidth of each suspected area,and they were used in mean-shift clustering segmentation.Experimental results show that by the introduction of HESSIAN matrix of dot filtering information to mean-shift clustering,nodular regions can be segmented from blood vessels,trachea,or cross regions connected to the nodule,non-nodular areas can be removed from ROIs properly,and ground glass object(GGO)nodular areas can also be segmented.For the experimental data set of 127 different forms of nodules,the average accuracy of the proposed algorithm is more than 90%.

Grey incidence clustering method based on multidimensional dynamic time warping distance

The traditional grey incidence degree is mainly based on the distance analysis methods, which is measured by the displacement difference between corresponding points between sequences. When some data of sequences are missing (inconsistency in the length of the sequences), the only way is to delete the longer sequences or to fill the shorter sequences. Therefore, some uncertainty is introduced. To solve this problem, by introducing three-dimensional grey incidence degree (3D-GID), a novel GID based on the multidimensional dynamic time warping distance (MDDTW distance-GID) is proposed. On the basis of it, the corresponding grey incidence clustering (MDDTW distance-GIC) method is constructed. It not only has the simpler computation process, but also can be applied to the incidence comparison between uncertain multidimensional sequences directly. The experiment shows that MDDTW distance-GIC is more accurate when dealing with the uncertain sequences. Compared with the traditional GIC method, the precision of the MDDTW distance-GIC method has increased nearly 30%.

Multi-scale calculation of settling speed of coarse particles by accelerated Stokesian dynamics without adjustable parameter

The calculation of settling speed of coarse particles is firstly addressed, with accelerated Stokesian dynamics without adjustable parameters, in which far field force acting on the particle instead of particle velocity is chosen as dependent variables to consider inter-particle hydrodynamic interactions. The sedimentation of a simple cubic array of spherical particles is simulated and compared to the results available to verify and validate the numerical code and computational scheme. The improved method keeps the same computational cost of the order O（NlogN） as usual accelerated Stokesian dynamics does. Then, more realistic random suspension sedimentation is investigated with the help of Mont Carlo method. The computational results agree well with experimental fitting. Finally, the sedimentation of finer cohesive particle, which is often observed in estuary environment, is presented as a further application in coastal engineering.

Research of Web Documents Clustering Based on Dynamic Concept

Conceptual clustering is mainly used for solving the deficiency and incompleteness of domain knowledge. Based on conceptual clustering technology and aiming at the institutional framework and characteristic of Web theme information, this paper proposes and implements dynamic conceptual clustering algorithm and merging algorithm for Web documents, and also analyses the super performance of the clustering algorithm in efficiency and clustering accuracy. Key words conceptual clustering - clustering center - dynamic conceptual clustering - theme - web documents clustering CLC number TP 311 Foundation item: Supported by the National “863” Program of China (2002AA111010, 2003AA001032)Biography: WANG Yun-hua(1979-), male, Master candidate, research direction: knowledge engineering and data mining.

Research on Dynamic Clustering Routing Considering Node Load for Wireless Sensor Networks

Aiming at the problem that node load is rarely considered in existing clustering routing algorithm for Wireless Sensor Networks (WSNs), a dynamic clustering routing algorithm for WSN is presented in this paper called DCRCL (Dynamic Clustering Routing Considering Load). This algorithm is comprised of three phases including cluster head (CH) selection, cluster setup and inter-cluster routing. First, the CHs are selected based on residual energy and node load. Then the non-CH nodes choose a cluster by comparing the cost function of its neighbor CHs. At last, each CH communicates with base station by using multi-hop communication. The simulation results show that comparing with the existing one, the techniques life cycle and date volume of the network are increased by 30.7 percent and 29.8 percent respectively by using the proposed algorithm DCRCL.

Clustering Algorithms to Analyze Molecular Dynamics Simulation Trajectories for Complex Chemical and Biological Systems

Molecular dynamics （MD） simulation has become a powerful tool to investigate the structure- function relationship of proteins and other biological macromolecules at atomic resolution and biologically relevant timescales. MD simulations often produce massive datasets con- taining millions of snapshots describing proteins in motion. Therefore, clustering algorithms have been in high demand to be developed and applied to classify these MD snapshots and gain biological insights. There mainly exist two categories of clustering algorithms that aim to group protein conformations into clusters based on the similarity of their shape （geometric clustering） and kinetics （kinetic clustering）. In this paper, we review a series of frequently used clustering algorithms applied in MD simulations, including divisive algorithms, ag- glomerative algorithms （single-linkage, complete-linkage, average-linkage, centroid-linkage and ward-linkage）, center-based algorithms （K-Means, K-Medoids, K-Centers, and APM）, density-based algorithms （neighbor-based, DBSCAN, density-peaks, and Robust-DB）, and spectral-based algorithms （PCCA and PCCA＋）. In particular, differences between geomet- ric and kinetic clustering metrics will be discussed along with the performances of diflhrent clustering algorithms. We note that there does not exist a one-size-fits-all algorithm in the classification of MD datasets. For a specific application, the right choice of clustering algo- rithm should be based on the purpose of clustering, and the intrinsic properties of the MD conformational ensembles. Therefore, a main focus of our review is to describe the merits and limitations of each clustering algorithm. We expect that this review would be helpful to guide researchers to choose appropriate clustering algorithms for their own MD datasets.

A QoS Mobicast-based dynamic clustering secure multicast scheme for large-scale tracking sensornets

Most of the existing security Mobicast routing protocols are not suitable for the monitoring applications with higher quality of service （QoS） requirement. A QoS dynamic clustering secure multicast scheme （QoS-DCSMS） based on Mobicast and multi-level IxTESLA protocol for large-scale tracking sensornets is presented in this paper. The multicast clusters are dynamically formed according to the real-time status of nodes, and the cluster-head node is responsible for status review and certificating management of cluster nodes to ensure the most optimized QoS and security of multicast in this scheme. Another contribution of this paper is the optimal QoS security authentication algorithm, which analyzes the relationship between the QoS and the level Mofmulti-level oTESLA. Based on the analysis and simulation results, it shows that the influence to the network survival cycle （＇NSC） and real-time communication caused by energy consumption and latency in authentication is acceptable when the optimal QoS security authentication algorithm is satisfied.

Molecular Dynamics Studies of Solid-Liquid Phase Transitions Ⅱ——Homogeneous Nucleation of (KI)_(108) Clusters

在一系列分子动力学计算中，（ＫＩ）１０８离子簇冷却时自发地由液相变为立方面心相．在３５０Ｋ和４００Ｋ的成核速率大于１０３６核／ｍ３·ｓ．由成核速率得到的固－液界面自由能约为５０ｍＪ／ｍ２．

Multi-scale simulation of diffusion behavior of deterrent in propellant

Concentration distribution of the deterrent in single-base propellant during the process of firing plays an important role in the ballistic properties of gun propellant in weapons. However, the diffusion coefficient calculated by molecular dynamics(MD) simulation is 6 orders of magnitude larger than the experimental values. Meanwhile, few simple and comprehensive theoretical models can explain the phenomenon and accurately predict the concentration distribution of the propellant. Herein, an onion model combining with MD simulation and finite element method of diffusion in propellants is introduced to bridge the gap between the experiments and simulations, and correctly predict the concentration distribution of deterrent. Furthermore, a new time scale is found to characterize the diffusion process. Finally, the time-and position-depended concentration distributions of dibutyl phthalate in nitrocellulose are measured by Raman spectroscopy to verify the correctness of the onion model. This work not only provides guidance for the design of the deterrent, but could be also extended to the diffusion of small molecules in polymer with different crystallinity.

SEMI-EMPIRICAL CALCULATION FOR ELECTRONIC STRUCTURE OF C60 CLUSTER BY MOLECULAR DYNAMICS AND MNDO APPROACH

The electronic structure for C 60 was semi empirically investigated by using MD (molecular dynamics) and MNDO (modified neglect of diatomic overlap) approach of quantum chemistry.Especially,taking both σ and π orbitals into account,one electron energy levels,those symmetries and π orbital occupancies as well as electron excitation energies for different select rules,cohesive energy,ionization energies and electronic affinity forces were calculated.The obtained molecular orbital ratio shows a wide separation of σ and π types,and near HOMO and LUMO levels there are π orbitals mainly.The calculated semi empirical calculation results are in good agreement with experimental and ab initio calculation data.

Projection Pursuit Dynamic Cluster Model and its Application to Water Resources Carrying Capacity Evaluation

The research shows that projection pursuit cluster (PPC) model is able to form a suitable index for overcom-ing the difficulties in comprehensive evaluation, which can be used to analyze complex multivariate prob-lems. The PPC model is widely used in multifactor cluster and evaluation analysis, but there are a few prob-lems needed to be solved in practice, such as cutoff radius parameter calibration. In this study, a new model-projection pursuit dynamic cluster (PPDC) model-based on projection pursuit principle is developed and used in water resources carrying capacity evaluation in China for the first time. In the PPDC model, there are two improvements compared with the PPC model, 1) a new projection index is constructed based on dynamic cluster principle, which avoids the problem of parameter calibration in the PPC model success-fully;2) the cluster results can be outputted directly according to the PPDC model, but the cluster results can be got based on the scatter points of projected characteristic values or the re-analysis for projected character-istic values in the PPC model. The results show that the PPDC model is a very effective and powerful tool in multifactor data exploratory analysis. It is a new method for water resources carrying capacity evaluation. The PPDC model and its application to water resources carrying capacity evaluation are introduced in detail in this paper.

Dynamic cluster member selection method for multi-target tracking in wireless sensor network

Multi-target tracking(MTT) is a research hotspot of wireless sensor networks at present.A self-organized dynamic cluster task allocation scheme is used to implement collaborative task allocation for MTT in WSN and a special cluster member(CM) node selection method is put forward in the scheme.An energy efficiency model was proposed under consideration of both energy consumption and remaining energy balance in the network.A tracking accuracy model based on area-sum principle was also presented through analyzing the localization accuracy of triangulation.Then,the two models mentioned above were combined to establish dynamic cluster member selection model for MTT where a comprehensive performance index function was designed to guide the CM node selection.This selection was fulfilled using genetic algorithm.Simulation results show that this method keeps both energy efficiency and tracking quality in optimal state,and also indicate the validity of genetic algorithm in implementing CM node selection.